Getting started

clean up
general custom functions
necessary packages

Download data

Import data

tie_maintenance.RDa

wave 1 –> wave 2 (maintained vs dropped)
wave 2 created
wave 2 –> wave 3

data_nested.Rda

wave 1 –> wave 2
wave 2 –> wave 3

The following scripts can be used to replicate the data-set of Franken, Bekhuis, and Tolsma. It may also be obtained by downloading: Download data_nested.RDa

Getting started

To copy the code, click the button in the upper right corner of the code-chunks.

clean up

rm(list = ls())
gc()

general custom functions

fpackage.check: Check if packages are installed (and install if not) in R
fsave: Function to save data with time stamp in correct directory
fload: Load R-objects under new names
fshowdf: Print objects (tibble / data.frame) nicely on screen in .Rmd.

fpackage.check <- function(packages) {
    lapply(packages, FUN = function(x) {
        if (!require(x, character.only = TRUE)) {
            install.packages(x, dependencies = TRUE)
            library(x, character.only = TRUE)
        }
    })
}

fsave <- function(x, file, location = "./data/processed/", ...) {
    if (!dir.exists(location))
        dir.create(location)
    datename <- substr(gsub("[:-]", "", Sys.time()), 1, 8)
    totalname <- paste(location, datename, file, sep = "")
    print(paste("SAVED: ", totalname, sep = ""))
    save(x, file = totalname)
}


fload <- function(fileName) {
    load(fileName)
    get(ls()[ls() != "fileName"])
}

fshowdf <- function(x, digits = 2, ...) {
    knitr::kable(x, digits = digits, "html", ...) %>%
        kableExtra::kable_styling(bootstrap_options = c("striped", "hover")) %>%
        kableExtra::scroll_box(width = "100%", height = "300px")
}

necessary packages

dplyr: package for data wrangling
lubridate: parse and manipulate dates
stringr: string manipulation

packages = c("dplyr", "lubridate", "stringr")
fpackage.check(packages)
rm(packages)

Download data

Anonymized data-sets of the ‘Sports and Friendships’ (Franken, Bekhuis, and Tolsma 2023) study are deposited in DANS EASY. For this study we use waves 3-5 (Cohort II):

Download the data-files, and put them in the ./data/ folder. But first, make a ./data/ folder:

ifelse(!dir.exists("data"), dir.create("data"), FALSE)

To see the code used to anonymize the raw data files, click the button below.

load in raw data

data1 <- read.csv("./rawdata/results-survey938243.csv")  #wave1
data2 <- read.csv("./rawdata/results-survey868275.csv")  #wave2
data3 <- read.csv("./rawdata/results-survey493144.csv")  #wave3
data4a <- read.csv("./rawdata/results-survey952354a.csv")  #wave4; the data was exported in the collection period of this wave, to restore a bug in limesurvey.
data4b <- read.csv("./rawdata/results-survey952354b.csv")
data5 <- read.csv("./rawdata/results-survey356557.csv")  #wave 5

make public data-sets to deposit

The following scripts were used to delete sensitive information from the respective survey rounds (e.g., information that could identify the respondent, names of alters, etc.) and to make public data-sets for each wave.

#copy dataframe
df <- data1

# first, delete fake responses
fakeid <- c(337, 575, 576) #id of fake responses

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=170

#make new id variable
df$respnr <- 1:nrow(df)

#add 10000, sygnifying the cohort
df$respnr <- df$respnr + 10000

#remove first 10 columns (e.g,. token, submitdate, refurl, etc.)
#i dont need these
head(colnames(df),25)
df <- df[,-c(1:10)]
#now, column 1 is the first question of the survey

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
study <- 6 # study program
occupation <- 37
residency <- c(153,154,157,158) #residence (municipality name; 4digit postal code)
exclude_colnumber <- c(study,occupation,residency)
#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]

#i also need to deal with the name-generator data;
#i replace all listed names with a unique, anonymous alter-id,

#first, get all column numbers corresponding to names
net1 <- c(10:14)
net2 <- c(50:54)
net3 <- c(166:170)
net4 <- c(440:444)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, we delete columns that are irrelevant
df<-df[,-c(1:4)]

#and the last columns (excluding the very last, which denotes `respnr`)
#but first, we subset the respnr-email combination, so that we can merge with wave2
merge <- df[,c("respnr", "email")]
excl <- c(869:length(df))
excl <- excl[-length(excl)]
df <- df[, -excl]

#fix(df)#check if everything went alright
row.names(df) <- 1:nrow(df)

#save the dataframe to the processed data-folder
fsave(df, "wave1_public.RDa")

#copy dataframe
df <- data2

# first, delete fake response
fakeid <- c(61) #id of fake response

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=93

#attach respnr
df <- merge(df,merge,by="email")

#reorder by respr
df <- df[order(df$respnr),]
rownames(df) <- 1:nrow(df)

#delete first columns
df <- df[,-c(1:12)]

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
#in wave 2, only residency variables
exclude_colnumber <- c(130,131)
#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]

#again, replace named persons with unique alter_id p_i
net1 <- c(1:5)
net2 <- c(26:30)
net3 <- c(142:146)
net4 <- c(336:340)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, remove all irrelevant variables,
#from "interviewtime" on.
#excluding the last column, as this is respnr
#which( colnames(df)=="interviewtime" )
excl <- c( which( colnames(df)=="interviewtime" ):length(df))
excl <- excl[-length(excl)]
df <- df[, -excl]

fsave(df, "wave2_public.RDa")

#copy dataframe
df <- data3

# first, delete fake respons;
fakeid <- which(df$token  %in% "test")

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=1135

#make new id variable
#add 20000, so we have a signifier for the 2nd cohort
df$respnr <- 20000 + 1:nrow(df)

#remove first columns (e.g,. token, refurl, etc.)
#i dont need these
#but do keep submit and startdate
head(colnames(df),25)
df <- df[,-c(1,3:6,8:10)]

#now, column 3 is the first question of the survey

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
study <- 6 # study program
occupation <- 37 #other category 
residency <- c(153,154,157,158) #residence (municipality name; 4digit postal code)
exclude_colnumber <- c(study,occupation,residency)

#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]
#i also need to deal with the name-generator data;
#i replace all listed names with a unique, anonymous alter-id,

#first, get all column numbers corresponding to names
net1 <- c(10:14)
net2 <- c(50:54)
net3 <- c(166:170)
net4 <- c(441:445)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, we delete columns that are irrelevant
df<-df[,-c(3,4)]

#we also delete a lot of columns (excluding the very last, which denotes `respnr`; and variables indicating group times)

#but first, we subset the respnr-email combination, so that we can merge with wave2
merge <- df[,c("respnr", "email")]

fix(merge)

#fsave(merge, file = "sleutelbestand_respnr_email.RDa")

excl <- c(995:length(df))
excl <- excl[-length(excl)] #minus last columns (respnr)

#also keep group times
excl <- excl[ ! excl %in% which(grepl("groupTime", names(df)))]

df <- df[, -excl]

#fix(df)#check if everything went alright

#save the dataframe to the processed data-folder
fsave(df, "wave3_public.RDa")

While conducting this survey round, I encountered an issue with a matching matrix not loading (matching1N1.SQ001_SQ001 did not load; see the codebook for more info). As a result, the survey was shut down temporarily to fix the bug, and then reactivated, thus resulting in two distinct data-files.

Before making the public data-set for this wave, I (a) delete participants that had to be presented matching1N1.SQ001_SQ001 (i.e., those who named at least one CDN and 1 study partner), and (b) merge the data-files.

# participants that named one or more CDN alters, AND 1 study partner, were not presented with a
# matching matrix.
length(which(data4a$egonet1.SQ001. != "" & data4a$egonet2.SQ001. != "" & data4a$egonet2.SQ002. == ""))  # 88 particpants. 

# i delete those...
data4a. <- data4a[-which(data4a$egonet1.SQ001. != "" & data4a$egonet2.SQ001. != "" & data4a$egonet2.SQ002. ==
    ""), ]
data4a.$matching1N1.SQ001_SQ001. <- NA

# in order to merge the files, make sure the have the same columns
data4a.. <- data4a.[, which(names(data4a.) %in% names(data4b))]

# rowbind the separate data-files
data4 <- rbind(data4a.., data4b)
nrow(data4)  #n=520

#copy dataframe
df <- data4

# first, delete fake/test response
fakeid <- c(3,4) #id of fake response

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=518

#attach respnr
df <- merge(df,merge,by="email")

#reorder by respr
df <- df[order(df$respnr),]
rownames(df) <- 1:nrow(df)

#delete first columns
#but keep start/submit date
#head(names(df),12)

df <- df[,-c(1,2,4:7,9:12)]

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
#study: 129
#residency variables: 138, 139

exclude_colnumber <- c(129, 138,139)
#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]

#again, replace named persons with unique alter_id p_i
net1 <- c(3:7)
net2 <- c(28:32)
net3 <- c(151:155)
net4 <- c(396:400)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, remove all irrelevant variables,
#excluding the last column, as this is respnr
excl <- 1356:length(df)
excl <- excl[-length(excl)] #minus last columns (respnr)

#and time records for question groups
excl <- excl[ ! excl %in% which(grepl("groupTime", names(df)))]

df <- df[, -excl]

fsave(df, "wave4_public.RDa")

#copy dataframe
df <- data5

#attach respnr
df <- merge(df,merge,by="email")

#reorder by respr
df <- df[order(df$respnr),]
rownames(df) <- 1:nrow(df)

#delete first columns
#but keep start/submit date
#head(names(df),12)
df <- df[,-c(1,2,4:7,9:12)]

#again, replace named persons with unique alter_id p_i
net1 <- c(3:7)
net2 <- c(8:12)
net3 <- c(98:102)
net4 <- c(253:257)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#also, anonymize the 'other' reasons given for not renaming an alter.
open <- c(1166:1185,1206:1225)
df[,open][df[,open]==""] <- NA
df[,open][!is.na(df[,open])] <- 1 #just put "1" instead...

#last, remove all irrelevant variables,
#excluding the last column, as this is respnr

#tail(names(df),308)


excl <- 1226:length(df)
excl <- excl[-length(excl)] #minus last columns (respnr)

#and time records for question groups
excl <- excl[ ! excl %in% which(grepl("groupTime", names(df)))]

df <- df[, -excl]
#fix(df)

fsave(df, "wave5_public.RDa")

Import data

Load the downloaded data. First, we clean our environment (but we keep our functions; we need them later on).

# clean environment, but keep functions
rm(list = setdiff(ls(), lsf.str()))

# load public data

# Cohort II
data3 <- fload("./data/wave3_public.RDa")
data4 <- fload("./data/wave4_public.RDa")
data5 <- fload("./data/wave5_public.RDa")

tie_maintenance.RDa

In the following script, I construct a long data-frame (based on Cohort II, waves 1-2-3), with alters nested in egos, to investigate which alters are maintained in the personal networks of ego over the academic year. To each ego, relevant individual-level, contextual-level, and network-level attributes are assigned; to each alter, relevant alter and dyadic features are assigned. The data-set is saved.

wave 1 –> wave 2 (maintained vs dropped)

# subset respondents that filled out both surveys w1 and w2;
# by matching on `respnr`
data <- data3[which(data3$respnr %in% unique(data4$respnr)),]
row.names(data) <- 1:nrow(data)
nrow(data) #N=516

# make a dataframe of w1-alters, with potential duplicates...
# recall that all alter names were replaced with a unique anonymized alter-id
df_names <- data.frame(
  p1 = data$egonet1.SQ001.,
  p2 = data$egonet1.SQ002.,
  p3 = data$egonet1.SQ003.,
  p4 = data$egonet1.SQ004.,
  p5 = data$egonet1.SQ005.,
  p6 = data$egonet2.SQ001.,
  p7 = data$egonet2.SQ002.,
  p8 = data$egonet2.SQ003.,
  p9 = data$egonet2.SQ004.,
  p10= data$egonet2.SQ005.,
  p11= data$egonet3.SQ001.,
  p12= data$egonet3.SQ002.,
  p13= data$egonet3.SQ003.,
  p14= data$egonet3.SQ004.,
  p15= data$egonet3.SQ005.,
  p16= data$egonet4.SQ001.,
  p17= data$egonet4.SQ002.,
  p18= data$egonet4.SQ003.,
  p19= data$egonet4.SQ004.,
  p20= data$egonet4.SQ005.)

#"pre-allocate" empty list of length equals number of egos
alterL <- vector("list", nrow(df_names))

# loop over all egos
for ( i in 1:length(alterL)) {
    alterL[[i]] <- data.frame(
      ego_gender = NA, ego_age = NA, ego_educ = NA,
      alterid = 1:20, name1 = NA, name2 = NA, name3 = NA, name4 = NA,
      alter_gender = NA, alter_age = NA, alter_educ=NA,
      same_gender = NA, dif_age = NA, sim_educ = NA,
      cdn_embed.t1 = NA, study_embed.t1 = NA, bff_embed.t1 = NA, csn_embed.t1 = NA)
}

# fill the names based on the names data-frame and replace empty strings with <NA>
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=FALSE)
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
}

# a matching matrix allowed ego to match names that referred to the same person.
# in limesurvey, i could not let the number of columns condition
# on the number of alters named in the latest egonet, so i made 5 separate matrices,
# conditional on netsize.
# i calculate netsize for each net
{
  net1 <- cbind(data$egonet1.SQ001.,data$egonet1.SQ002., data$egonet1.SQ003.,data$egonet1.SQ004., data$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data$egonet2.SQ001.,data$egonet2.SQ002., data$egonet2.SQ003.,data$egonet2.SQ004., data$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data$egonet3.SQ001.,data$egonet3.SQ002., data$egonet3.SQ003.,data$egonet3.SQ004., data$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data$egonet4.SQ001.,data$egonet4.SQ002., data$egonet4.SQ003.,data$egonet4.SQ004., data$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}

# i construct the matching matrices, 1 - 5 (conditional on the ns of the latest net)
# put these in a list and list these again, for each respondent (so a list of lists...)
matchingList <- list()
for (i in 1:length(alterL)) { # for ego i
  matchingL <- list()
  # make 5 seperate matching matrices. naturally, only 1 is relevant... but i will select that later.
  matchingL[[1]] <- cbind(data$matching1N1.SQ001_SQ001.[i],data$matching1N1.SQ002_SQ001.[i],data$matching1N1.SQ003_SQ001.[i],data$matching1N1.SQ004_SQ001.[i],data$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data$matching1N2.SQ001_SQ001.[i], data$matching1N2.SQ002_SQ001.[i], data$matching1N2.SQ003_SQ001.[i], data$matching1N2.SQ004_SQ001.[i], data$matching1N2.SQ005_SQ001.[i]),
    cbind(data$matching1N2.SQ001_SQ002.[i], data$matching1N2.SQ002_SQ002.[i], data$matching1N2.SQ003_SQ002.[i], data$matching1N2.SQ004_SQ002.[i], data$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data$matching1N3.SQ001_SQ001.[i], data$matching1N3.SQ002_SQ001.[i], data$matching1N3.SQ003_SQ001.[i], data$matching1N3.SQ004_SQ001.[i], data$matching1N3.SQ005_SQ001.[i]),
    cbind(data$matching1N3.SQ001_SQ002.[i], data$matching1N3.SQ002_SQ002.[i], data$matching1N3.SQ003_SQ002.[i], data$matching1N3.SQ004_SQ002.[i], data$matching1N3.SQ005_SQ002.[i]),
    cbind(data$matching1N3.SQ001_SQ003.[i], data$matching1N3.SQ002_SQ003.[i], data$matching1N3.SQ003_SQ003.[i], data$matching1N3.SQ004_SQ003.[i], data$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data$matching1N4.SQ001_SQ001.[i], data$matching1N4.SQ002_SQ001.[i], data$matching1N4.SQ003_SQ001.[i], data$matching1N4.SQ004_SQ001.[i], data$matching1N4.SQ005_SQ001.[i]),
    cbind(data$matching1N4.SQ001_SQ002.[i], data$matching1N4.SQ002_SQ002.[i], data$matching1N4.SQ003_SQ002.[i], data$matching1N4.SQ004_SQ002.[i], data$matching1N4.SQ005_SQ002.[i]),
    cbind(data$matching1N4.SQ001_SQ003.[i], data$matching1N4.SQ002_SQ003.[i], data$matching1N4.SQ003_SQ003.[i], data$matching1N4.SQ004_SQ003.[i], data$matching1N4.SQ005_SQ003.[i]),
    cbind(data$matching1N4.SQ001_SQ004.[i], data$matching1N4.SQ002_SQ004.[i], data$matching1N4.SQ003_SQ004.[i], data$matching1N4.SQ004_SQ004.[i], data$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data$matching1N5.SQ001_SQ001.[i], data$matching1N5.SQ002_SQ001.[i], data$matching1N5.SQ003_SQ001.[i], data$matching1N5.SQ004_SQ001.[i], data$matching1N5.SQ005_SQ001.[i]),
    cbind(data$matching1N5.SQ001_SQ002.[i], data$matching1N5.SQ002_SQ002.[i], data$matching1N5.SQ003_SQ002.[i], data$matching1N5.SQ004_SQ002.[i], data$matching1N5.SQ005_SQ002.[i]),
    cbind(data$matching1N5.SQ001_SQ003.[i], data$matching1N5.SQ002_SQ003.[i], data$matching1N5.SQ003_SQ003.[i], data$matching1N5.SQ004_SQ003.[i], data$matching1N5.SQ005_SQ003.[i]),
    cbind(data$matching1N5.SQ001_SQ004.[i], data$matching1N5.SQ002_SQ004.[i], data$matching1N5.SQ003_SQ004.[i], data$matching1N5.SQ004_SQ004.[i], data$matching1N5.SQ005_SQ004.[i]),
    cbind(data$matching1N5.SQ001_SQ005.[i], data$matching1N5.SQ002_SQ005.[i], data$matching1N5.SQ003_SQ005.[i], data$matching1N5.SQ004_SQ005.[i], data$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
} # so... matchingL[[1]][[5]] is matchingmatrix 5 (i.e., 5 alters named in egonet2) for ego 1

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
} #ignore warning

#repeat for the second matching matrices
# (i.e., matching egonet3 alters to prev. alters);
matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data$matching2N1.SQ001_SQ001.[i],data$matching2N1.SQ002_SQ001.[i],data$matching2N1.SQ003_SQ001.[i],data$matching2N1.SQ004_SQ001.[i],data$matching2N1.SQ005_SQ001.[i],data$matching2N1.SQ006_SQ001.[i],data$matching2N1.SQ007_SQ001.[i],data$matching2N1.SQ008_SQ001.[i],data$matching2N1.SQ009_SQ001.[i],data$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data$matching2N2.SQ001_SQ001.[i],data$matching2N2.SQ002_SQ001.[i],data$matching2N2.SQ003_SQ001.[i],data$matching2N2.SQ004_SQ001.[i],data$matching2N2.SQ005_SQ001.[i],data$matching2N2.SQ006_SQ001.[i],data$matching2N2.SQ007_SQ001.[i],data$matching2N2.SQ008_SQ001.[i],data$matching2N2.SQ009_SQ001.[i],data$matching2N2.SQ010_SQ001.[i]),
    cbind(data$matching2N2.SQ001_SQ002.[i],data$matching2N2.SQ002_SQ002.[i],data$matching2N2.SQ003_SQ002.[i],data$matching2N2.SQ004_SQ002.[i],data$matching2N2.SQ005_SQ002.[i],data$matching2N2.SQ006_SQ002.[i],data$matching2N2.SQ007_SQ002.[i],data$matching2N2.SQ008_SQ002.[i],data$matching2N2.SQ009_SQ002.[i],data$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data$matching2N3.SQ001_SQ001.[i],data$matching2N3.SQ002_SQ001.[i],data$matching2N3.SQ003_SQ001.[i],data$matching2N3.SQ004_SQ001.[i],data$matching2N3.SQ005_SQ001.[i],data$matching2N3.SQ006_SQ001.[i],data$matching2N3.SQ007_SQ001.[i],data$matching2N3.SQ008_SQ001.[i],data$matching2N3.SQ009_SQ001.[i],data$matching2N3.SQ010_SQ001.[i]),
    cbind(data$matching2N3.SQ001_SQ002.[i],data$matching2N3.SQ002_SQ002.[i],data$matching2N3.SQ003_SQ002.[i],data$matching2N3.SQ004_SQ002.[i],data$matching2N3.SQ005_SQ002.[i],data$matching2N3.SQ006_SQ002.[i],data$matching2N3.SQ007_SQ002.[i],data$matching2N3.SQ008_SQ002.[i],data$matching2N3.SQ009_SQ002.[i],data$matching2N3.SQ010_SQ002.[i]),
    cbind(data$matching2N3.SQ001_SQ003.[i],data$matching2N3.SQ002_SQ003.[i],data$matching2N3.SQ003_SQ003.[i],data$matching2N3.SQ004_SQ003.[i],data$matching2N3.SQ005_SQ003.[i],data$matching2N3.SQ006_SQ003.[i],data$matching2N3.SQ007_SQ003.[i],data$matching2N3.SQ008_SQ003.[i],data$matching2N3.SQ009_SQ003.[i],data$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data$matching2N4.SQ001_SQ001.[i],data$matching2N4.SQ002_SQ001.[i],data$matching2N4.SQ003_SQ001.[i],data$matching2N4.SQ004_SQ001.[i],data$matching2N4.SQ005_SQ001.[i],data$matching2N4.SQ006_SQ001.[i],data$matching2N4.SQ007_SQ001.[i],data$matching2N4.SQ008_SQ001.[i],data$matching2N4.SQ009_SQ001.[i],data$matching2N4.SQ010_SQ001.[i]),
    cbind(data$matching2N4.SQ001_SQ002.[i],data$matching2N4.SQ002_SQ002.[i],data$matching2N4.SQ003_SQ002.[i],data$matching2N4.SQ004_SQ002.[i],data$matching2N4.SQ005_SQ002.[i],data$matching2N4.SQ006_SQ002.[i],data$matching2N4.SQ007_SQ002.[i],data$matching2N4.SQ008_SQ002.[i],data$matching2N4.SQ009_SQ002.[i],data$matching2N4.SQ010_SQ002.[i]),
    cbind(data$matching2N4.SQ001_SQ003.[i],data$matching2N4.SQ002_SQ003.[i],data$matching2N4.SQ003_SQ003.[i],data$matching2N4.SQ004_SQ003.[i],data$matching2N4.SQ005_SQ003.[i],data$matching2N4.SQ006_SQ003.[i],data$matching2N4.SQ007_SQ003.[i],data$matching2N4.SQ008_SQ003.[i],data$matching2N4.SQ009_SQ003.[i],data$matching2N4.SQ010_SQ003.[i]),
    cbind(data$matching2N4.SQ001_SQ004.[i],data$matching2N4.SQ002_SQ004.[i],data$matching2N4.SQ003_SQ004.[i],data$matching2N4.SQ004_SQ004.[i],data$matching2N4.SQ005_SQ004.[i],data$matching2N4.SQ006_SQ004.[i],data$matching2N4.SQ007_SQ004.[i],data$matching2N4.SQ008_SQ004.[i],data$matching2N4.SQ009_SQ004.[i],data$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data$matching2N5.SQ001_SQ001.[i],data$matching2N5.SQ002_SQ001.[i],data$matching2N5.SQ003_SQ001.[i],data$matching2N5.SQ004_SQ001.[i],data$matching2N5.SQ005_SQ001.[i],data$matching2N5.SQ006_SQ001.[i],data$matching2N5.SQ007_SQ001.[i],data$matching2N5.SQ008_SQ001.[i],data$matching2N5.SQ009_SQ001.[i],data$matching2N5.SQ010_SQ001.[i]),
    cbind(data$matching2N5.SQ001_SQ002.[i],data$matching2N5.SQ002_SQ002.[i],data$matching2N5.SQ003_SQ002.[i],data$matching2N5.SQ004_SQ002.[i],data$matching2N5.SQ005_SQ002.[i],data$matching2N5.SQ006_SQ002.[i],data$matching2N5.SQ007_SQ002.[i],data$matching2N5.SQ008_SQ002.[i],data$matching2N5.SQ009_SQ002.[i],data$matching2N5.SQ010_SQ002.[i]),
    cbind(data$matching2N5.SQ001_SQ003.[i],data$matching2N5.SQ002_SQ003.[i],data$matching2N5.SQ003_SQ003.[i],data$matching2N5.SQ004_SQ003.[i],data$matching2N5.SQ005_SQ003.[i],data$matching2N5.SQ006_SQ003.[i],data$matching2N5.SQ007_SQ003.[i],data$matching2N5.SQ008_SQ003.[i],data$matching2N5.SQ009_SQ003.[i],data$matching2N5.SQ010_SQ003.[i]),
    cbind(data$matching2N5.SQ001_SQ004.[i],data$matching2N5.SQ002_SQ004.[i],data$matching2N5.SQ003_SQ004.[i],data$matching2N5.SQ004_SQ004.[i],data$matching2N5.SQ005_SQ004.[i],data$matching2N5.SQ006_SQ004.[i],data$matching2N5.SQ007_SQ004.[i],data$matching2N5.SQ008_SQ004.[i],data$matching2N5.SQ009_SQ004.[i],data$matching2N5.SQ010_SQ004.[i]),
    cbind(data$matching2N5.SQ001_SQ005.[i],data$matching2N5.SQ002_SQ005.[i],data$matching2N5.SQ003_SQ005.[i],data$matching2N5.SQ004_SQ005.[i],data$matching2N5.SQ005_SQ005.[i],data$matching2N5.SQ006_SQ005.[i],data$matching2N5.SQ007_SQ005.[i],data$matching2N5.SQ008_SQ005.[i],data$matching2N5.SQ009_SQ005.[i],data$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

#and for matching matrix 3 (i.e., matching egonet4 with egonets 1-3)
matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data$matching3N1.SQ001_SQ001.[i],data$matching3N1.SQ002_SQ001.[i],data$matching3N1.SQ003_SQ001.[i],data$matching3N1.SQ004_SQ001.[i],data$matching3N1.SQ005_SQ001.[i],data$matching3N1.SQ006_SQ001.[i],data$matching3N1.SQ007_SQ001.[i],data$matching3N1.SQ008_SQ001.[i],data$matching3N1.SQ009_SQ001.[i],data$matching3N1.SQ010_SQ001.[i], data$matching3N1.SQ011_SQ001.[i], data$matching3N1.SQ012_SQ001.[i], data$matching3N1.SQ013_SQ001.[i], data$matching3N1.SQ014_SQ001.[i], data$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data$matching3N2.SQ001_SQ001.[i],data$matching3N2.SQ002_SQ001.[i],data$matching3N2.SQ003_SQ001.[i],data$matching3N2.SQ004_SQ001.[i],data$matching3N2.SQ005_SQ001.[i],data$matching3N2.SQ006_SQ001.[i],data$matching3N2.SQ007_SQ001.[i],data$matching3N2.SQ008_SQ001.[i],data$matching3N2.SQ009_SQ001.[i],data$matching3N2.SQ010_SQ001.[i], data$matching3N2.SQ011_SQ001.[i], data$matching3N2.SQ012_SQ001.[i], data$matching3N2.SQ013_SQ001.[i], data$matching3N2.SQ014_SQ001.[i], data$matching3N2.SQ015_SQ001.[i]),
    cbind(data$matching3N2.SQ001_SQ002.[i],data$matching3N2.SQ002_SQ002.[i],data$matching3N2.SQ003_SQ002.[i],data$matching3N2.SQ004_SQ002.[i],data$matching3N2.SQ005_SQ002.[i],data$matching3N2.SQ006_SQ002.[i],data$matching3N2.SQ007_SQ002.[i],data$matching3N2.SQ008_SQ002.[i],data$matching3N2.SQ009_SQ002.[i],data$matching3N2.SQ010_SQ002.[i], data$matching3N2.SQ011_SQ002.[i], data$matching3N2.SQ012_SQ002.[i], data$matching3N2.SQ013_SQ002.[i], data$matching3N2.SQ014_SQ002.[i], data$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data$matching3N3.SQ001_SQ001.[i],data$matching3N3.SQ002_SQ001.[i],data$matching3N3.SQ003_SQ001.[i],data$matching3N3.SQ004_SQ001.[i],data$matching3N3.SQ005_SQ001.[i],data$matching3N3.SQ006_SQ001.[i],data$matching3N3.SQ007_SQ001.[i],data$matching3N3.SQ008_SQ001.[i],data$matching3N3.SQ009_SQ001.[i],data$matching3N3.SQ010_SQ001.[i], data$matching3N3.SQ011_SQ001.[i], data$matching3N3.SQ012_SQ001.[i], data$matching3N3.SQ013_SQ001.[i], data$matching3N3.SQ014_SQ001.[i], data$matching3N3.SQ015_SQ001.[i]),
    cbind(data$matching3N3.SQ001_SQ002.[i],data$matching3N3.SQ002_SQ002.[i],data$matching3N3.SQ003_SQ002.[i],data$matching3N3.SQ004_SQ002.[i],data$matching3N3.SQ005_SQ002.[i],data$matching3N3.SQ006_SQ002.[i],data$matching3N3.SQ007_SQ002.[i],data$matching3N3.SQ008_SQ002.[i],data$matching3N3.SQ009_SQ002.[i],data$matching3N3.SQ010_SQ002.[i], data$matching3N3.SQ011_SQ002.[i], data$matching3N3.SQ012_SQ002.[i], data$matching3N3.SQ013_SQ002.[i], data$matching3N3.SQ014_SQ002.[i], data$matching3N3.SQ015_SQ002.[i]),
    cbind(data$matching3N3.SQ001_SQ003.[i],data$matching3N3.SQ002_SQ003.[i],data$matching3N3.SQ003_SQ003.[i],data$matching3N3.SQ004_SQ003.[i],data$matching3N3.SQ005_SQ003.[i],data$matching3N3.SQ006_SQ003.[i],data$matching3N3.SQ007_SQ003.[i],data$matching3N3.SQ008_SQ003.[i],data$matching3N3.SQ009_SQ003.[i],data$matching3N3.SQ010_SQ003.[i], data$matching3N3.SQ011_SQ003.[i], data$matching3N3.SQ012_SQ003.[i], data$matching3N3.SQ013_SQ003.[i], data$matching3N3.SQ014_SQ003.[i], data$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data$matching3N4.SQ001_SQ001.[i],data$matching3N4.SQ002_SQ001.[i],data$matching3N4.SQ003_SQ001.[i],data$matching3N4.SQ004_SQ001.[i],data$matching3N4.SQ005_SQ001.[i],data$matching3N4.SQ006_SQ001.[i],data$matching3N4.SQ007_SQ001.[i],data$matching3N4.SQ008_SQ001.[i],data$matching3N4.SQ009_SQ001.[i],data$matching3N4.SQ010_SQ001.[i], data$matching3N4.SQ011_SQ001.[i], data$matching3N4.SQ012_SQ001.[i], data$matching3N4.SQ013_SQ001.[i], data$matching3N4.SQ014_SQ001.[i], data$matching3N4.SQ015_SQ001.[i]),
    cbind(data$matching3N4.SQ001_SQ002.[i],data$matching3N4.SQ002_SQ002.[i],data$matching3N4.SQ003_SQ002.[i],data$matching3N4.SQ004_SQ002.[i],data$matching3N4.SQ005_SQ002.[i],data$matching3N4.SQ006_SQ002.[i],data$matching3N4.SQ007_SQ002.[i],data$matching3N4.SQ008_SQ002.[i],data$matching3N4.SQ009_SQ002.[i],data$matching3N4.SQ010_SQ002.[i], data$matching3N4.SQ011_SQ002.[i], data$matching3N4.SQ012_SQ002.[i], data$matching3N4.SQ013_SQ002.[i], data$matching3N4.SQ014_SQ002.[i], data$matching3N4.SQ015_SQ002.[i]),
    cbind(data$matching3N4.SQ001_SQ003.[i],data$matching3N4.SQ002_SQ003.[i],data$matching3N4.SQ003_SQ003.[i],data$matching3N4.SQ004_SQ003.[i],data$matching3N4.SQ005_SQ003.[i],data$matching3N4.SQ006_SQ003.[i],data$matching3N4.SQ007_SQ003.[i],data$matching3N4.SQ008_SQ003.[i],data$matching3N4.SQ009_SQ003.[i],data$matching3N4.SQ010_SQ003.[i], data$matching3N4.SQ011_SQ003.[i], data$matching3N4.SQ012_SQ003.[i], data$matching3N4.SQ013_SQ003.[i], data$matching3N4.SQ014_SQ003.[i], data$matching3N4.SQ015_SQ003.[i]),
    cbind(data$matching3N4.SQ001_SQ003.[i],data$matching3N4.SQ002_SQ003.[i],data$matching3N4.SQ003_SQ003.[i],data$matching3N4.SQ004_SQ003.[i],data$matching3N4.SQ005_SQ003.[i],data$matching3N4.SQ006_SQ003.[i],data$matching3N4.SQ007_SQ003.[i],data$matching3N4.SQ008_SQ003.[i],data$matching3N4.SQ009_SQ003.[i],data$matching3N4.SQ010_SQ003.[i], data$matching3N4.SQ011_SQ003.[i], data$matching3N4.SQ012_SQ003.[i], data$matching3N4.SQ013_SQ003.[i], data$matching3N4.SQ014_SQ003.[i], data$matching3N4.SQ015_SQ003.[i]),
    cbind(data$matching3N4.SQ001_SQ004.[i],data$matching3N4.SQ002_SQ004.[i],data$matching3N4.SQ003_SQ004.[i],data$matching3N4.SQ004_SQ004.[i],data$matching3N4.SQ005_SQ004.[i],data$matching3N4.SQ006_SQ004.[i],data$matching3N4.SQ007_SQ004.[i],data$matching3N4.SQ008_SQ004.[i],data$matching3N4.SQ009_SQ004.[i],data$matching3N4.SQ010_SQ004.[i], data$matching3N4.SQ011_SQ004.[i], data$matching3N4.SQ012_SQ004.[i], data$matching3N4.SQ013_SQ004.[i], data$matching3N4.SQ014_SQ004.[i], data$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data$matching3N5.SQ001_SQ001.[i],data$matching3N5.SQ002_SQ001.[i],data$matching3N5.SQ003_SQ001.[i],data$matching3N5.SQ004_SQ001.[i],data$matching3N5.SQ005_SQ001.[i],data$matching3N5.SQ006_SQ001.[i],data$matching3N5.SQ007_SQ001.[i],data$matching3N5.SQ008_SQ001.[i],data$matching3N5.SQ009_SQ001.[i],data$matching3N5.SQ010_SQ001.[i], data$matching3N5.SQ011_SQ001.[i], data$matching3N5.SQ012_SQ001.[i], data$matching3N5.SQ013_SQ001.[i], data$matching3N5.SQ014_SQ001.[i], data$matching3N5.SQ015_SQ001.[i]),
    cbind(data$matching3N5.SQ001_SQ002.[i],data$matching3N5.SQ002_SQ002.[i],data$matching3N5.SQ003_SQ002.[i],data$matching3N5.SQ004_SQ002.[i],data$matching3N5.SQ005_SQ002.[i],data$matching3N5.SQ006_SQ002.[i],data$matching3N5.SQ007_SQ002.[i],data$matching3N5.SQ008_SQ002.[i],data$matching3N5.SQ009_SQ002.[i],data$matching3N5.SQ010_SQ002.[i], data$matching3N5.SQ011_SQ002.[i], data$matching3N5.SQ012_SQ002.[i], data$matching3N5.SQ013_SQ002.[i], data$matching3N5.SQ014_SQ002.[i], data$matching3N5.SQ015_SQ002.[i]),
    cbind(data$matching3N5.SQ001_SQ003.[i],data$matching3N5.SQ002_SQ003.[i],data$matching3N5.SQ003_SQ003.[i],data$matching3N5.SQ004_SQ003.[i],data$matching3N5.SQ005_SQ003.[i],data$matching3N5.SQ006_SQ003.[i],data$matching3N5.SQ007_SQ003.[i],data$matching3N5.SQ008_SQ003.[i],data$matching3N5.SQ009_SQ003.[i],data$matching3N5.SQ010_SQ003.[i], data$matching3N5.SQ011_SQ003.[i], data$matching3N5.SQ012_SQ003.[i], data$matching3N5.SQ013_SQ003.[i], data$matching3N5.SQ014_SQ003.[i], data$matching3N5.SQ015_SQ003.[i]),
    cbind(data$matching3N5.SQ001_SQ003.[i],data$matching3N5.SQ002_SQ003.[i],data$matching3N5.SQ003_SQ003.[i],data$matching3N5.SQ004_SQ003.[i],data$matching3N5.SQ005_SQ003.[i],data$matching3N5.SQ006_SQ003.[i],data$matching3N5.SQ007_SQ003.[i],data$matching3N5.SQ008_SQ003.[i],data$matching3N5.SQ009_SQ003.[i],data$matching3N5.SQ010_SQ003.[i], data$matching3N5.SQ011_SQ003.[i], data$matching3N5.SQ012_SQ003.[i], data$matching3N5.SQ013_SQ003.[i], data$matching3N5.SQ014_SQ003.[i], data$matching3N5.SQ015_SQ003.[i]),
    cbind(data$matching3N5.SQ001_SQ005.[i],data$matching3N5.SQ002_SQ005.[i],data$matching3N5.SQ003_SQ005.[i],data$matching3N5.SQ004_SQ005.[i],data$matching3N5.SQ005_SQ005.[i],data$matching3N5.SQ006_SQ005.[i],data$matching3N5.SQ007_SQ005.[i],data$matching3N5.SQ008_SQ005.[i],data$matching3N5.SQ009_SQ005.[i],data$matching3N5.SQ010_SQ005.[i], data$matching3N5.SQ011_SQ005.[i], data$matching3N5.SQ012_SQ005.[i], data$matching3N5.SQ013_SQ005.[i], data$matching3N5.SQ014_SQ005.[i], data$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

#i also calculate structural embeddedness (alter-level) per ego-net.

#first i fix some irregularities in the data
#1. a mistake was made in labeling the adjacency matrix variables for the friends network, so i first relabel them, so the same script below for constructing density scores can be used.
#(I started with N1 instead of N2 (N denotes the number of named alters; which should start at 2, because only at netsizes > 1, I asked for alter-ties) )
data <- data %>%
  rename (adj3N2a.SQ001. = adj3N1a.SQ001., adj3N3a.SQ001. = adj3N2a.SQ001.,adj3N3a.SQ002. = adj3N2a.SQ002.,adj3N3b.SQ001. = adj3N2b.SQ001.,adj3N4a.SQ001. = adj3N3a.SQ001.,adj3N4a.SQ002. = adj3N3a.SQ002.,adj3N4a.SQ003. = adj3N3a.SQ003.,adj3N4b.SQ001. = adj3N3b.SQ001.,adj3N4b.SQ002. = adj3N3b.SQ002.,   adj3N4c.SQ001. = adj3N3c.SQ001.,adj3N5a.SQ001. = adj3N4a.SQ001.,adj3N5a.SQ002. = adj3N4a.SQ002.,adj3N5a.SQ003. = adj3N4a.SQ003.,adj3N5a.SQ004. = adj3N4a.SQ004.,adj3N5b.SQ001. = adj3N4b.SQ001.,adj3N5b.SQ002. = adj3N4b.SQ002.,adj3N5b.SQ003. = adj3N4b.SQ003.,adj3N5c.SQ001. = adj3N4c.SQ001.,adj3N5c.SQ002. = adj3N4c.SQ002.,adj3N5d.SQ001. = adj3N4d.SQ001.)

#2. a labeling mistake: for the adjacency matrix for sports ties, with ns=3, the question about relation between alter 2 and 3 was coded wrongly. 
data$adj4N3b.SQ001. <- data$adj4N3b.SQ002.
##

for (i in 1:length(alterL)) { # for ego i

  #get number of names listed in the egonets
  nnames_cdn <- length(which(data[i,c("egonet1.SQ001.","egonet1.SQ002.","egonet1.SQ003.","egonet1.SQ004.","egonet1.SQ005.")]!=""))
  nnames_study <- length(which(data[i,c("egonet2.SQ001.","egonet2.SQ002.","egonet2.SQ003.","egonet2.SQ004.","egonet2.SQ005.")]!=""))
  nnames_bff <- length(which(data[i,c("egonet3.SQ001.","egonet3.SQ002.","egonet3.SQ003.","egonet3.SQ004.","egonet3.SQ005.")]!=""))
  nnames_csn <- length(which(data[i,c("egonet4.SQ001.","egonet4.SQ002.","egonet4.SQ003.","egonet4.SQ004.","egonet4.SQ005.")]!=""))
  
  #CDN
  
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj1N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj1N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj1N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj1N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj1N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj1N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj1N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj1N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj1N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj1N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj1N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj1N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj1N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj1N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj1N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj1N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj1N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj1N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj1N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj1N5d.SQ001.[i]
  }

  if(nnames_cdn>1) { #we only know alter-alter relations for nnames>1
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_cdn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
  
  #CDN: 1:nnames_CDN
  alterL[[i]]$cdn_embed.t1[1:nnames_cdn] <- embed 
  }
  
  #STUDY

  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj2N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj2N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj2N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj2N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj2N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj2N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj2N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj2N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj2N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj2N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj2N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj2N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj2N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj2N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj2N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj2N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj2N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj2N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj2N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj2N5d.SQ001.[i]
  }

  if(nnames_study>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_study]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$study_embed.t1[6:(5+nnames_study)] <- embed 
  }
  
  # BEST FRIENDS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj3N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj3N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj3N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj3N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj3N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj3N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj3N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj3N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj3N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj3N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj3N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj3N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj3N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj3N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj3N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj3N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj3N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj3N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj3N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj3N5d.SQ001.[i]
  }

  if(nnames_bff>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_bff]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$bff_embed.t1[11:(10+nnames_bff)] <- embed 
  }
  
  # SPORTS PARTNERS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj4N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj4N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj4N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj4N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj4N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj4N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj4N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj4N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj4N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj4N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj4N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj4N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj4N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj4N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj4N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj4N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj4N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj4N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj4N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj4N5d.SQ001.[i]
  }

  if(nnames_csn>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_csn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$csn_embed.t1[16:(15+nnames_csn)] <- embed 
  }
}

#alters could be named in multiple name generators. each element of alterL contains rows corresponding to alters that may be 'duplicates'.
#i match the structural embeddedness measures

for (i in 1:length(alterL)) { #for ego i 
  for (j in 1:nrow(alterL[[i]])) { #for alter j
        
    #if name2 is not empty, alter j was matched to the study network; and more precisely, to the study partner with name1: alterL[[i]]$name2[j]
    alterL[[i]]$study_embed.t1[j] <- ifelse(!is.na(alterL[[i]]$name2[j]), 
                                            alterL[[i]]$study_embed.t1[which(alterL[[i]]$name1==alterL[[i]]$name2[j] & !is.na(alterL[[i]]$study_embed.t1))],
                                            alterL[[i]]$study_embed.t1[j])
    
    #if name3 is not empty, alter j was matched to the friends network; and more precisely, to the friend with name1: alterL[[i]]$name3[j]
    alterL[[i]]$bff_embed.t1[j] <- ifelse(!is.na(alterL[[i]]$name3[j]), 
                                          alterL[[i]]$bff_embed.t1[which(alterL[[i]]$name1==alterL[[i]]$name3[j] & !is.na(alterL[[i]]$bff_embed.t1))],
                                          alterL[[i]]$bff_embed.t1[j])
    
    #if name4 is not empty, alter j was matched to the sports network; and more precisely, to the sports partner with name1: alterL[[i]]$name4[j]
    alterL[[i]]$csn_embed.t1[j] <- ifelse(!is.na(alterL[[i]]$name4[j]), 
                                          alterL[[i]]$csn_embed.t1[which(alterL[[i]]$name1==alterL[[i]]$name4[j] & !is.na(alterL[[i]]$csn_embed.t1))],     
                                          alterL[[i]]$csn_embed.t1[j] )
  }
}

# i use the name interpreter data to add (stable) alter characteristics (e.g., gender)
# in name-interpreter questions, only unique (i.e., non-matched) alters were listed... thus, alter ids without a gender assigned to them are non-uniques, and can later be filtered out.
df_gender <- data.frame(p1 = data$gender.SQ001.,p2 = data$gender.SQ002.,p3 = data$gender.SQ003.,p4 = data$gender.SQ004.,p5 = data$gender.SQ005.,p6 = data$gender.SQ006.,p7 = data$gender.SQ007.,p8 = data$gender.SQ008.,p9 = data$gender.SQ009.,p10= data$gender.SQ010.,p11= data$gender.SQ011.,p12= data$gender.SQ012.,p13= data$gender.SQ013.,p14= data$gender.SQ014.,p15= data$gender.SQ015.,p16= data$gender.SQ016.,p17= data$gender.SQ017.,p18= data$gender.SQ018.,p19= data$gender.SQ019.,p20= data$gender.SQ020.)

# same for age...
df_age <- data.frame(p1 = data$age.SQ001.,p2 = data$age.SQ002.,p3 = data$age.SQ003.,p4 = data$age.SQ004.,p5 = data$age.SQ005.,p6 = data$age.SQ006.,p7 = data$age.SQ007.,p8 = data$age.SQ008.,p9 = data$age.SQ009.,p10= data$age.SQ010.,p11= data$age.SQ011.,p12= data$age.SQ012.,p13= data$age.SQ013.,p14= data$age.SQ014.,p15= data$age.SQ015.,p16= data$age.SQ016.,p17= data$age.SQ017.,p18= data$age.SQ018.,p19= data$age.SQ019.,p20= data$age.SQ020.)

# kin
df_kin <- data.frame( p1 = data$kin.SQ001.,p2 = data$kin.SQ002.,p3 = data$kin.SQ003.,p4 = data$kin.SQ004.,p5 = data$kin.SQ005.,p6 = data$kin.SQ006.,p7 = data$kin.SQ007.,p8 = data$kin.SQ008.,p9 = data$kin.SQ009.,p10= data$kin.SQ010.,p11= data$kin.SQ011.,p12= data$kin.SQ012.,p13= data$kin.SQ013.,p14= data$kin.SQ014.,p15= data$kin.SQ015.,p16= data$kin.SQ016.,p17= data$kin.SQ017.,p18= data$kin.SQ018.,p19= data$kin.SQ019.,p20= data$kin.SQ020.)

#education
df_educ <- data.frame(p1 = data$educ.SQ001.,p2 = data$educ.SQ002.,p3 = data$educ.SQ003.,p4 = data$educ.SQ004.,p5 = data$educ.SQ005.,p6 = data$educ.SQ006.,p7 = data$educ.SQ007.,p8 = data$educ.SQ008., p9 = data$educ.SQ009.,p10= data$educ.SQ010., p11= data$educ.SQ011., p12= data$educ.SQ012., p13= data$educ.SQ013., p14= data$educ.SQ014., p15= data$educ.SQ015., p16= data$educ.SQ016., p17= data$educ.SQ017., p18= data$educ.SQ018., p19= data$educ.SQ019., p20= data$educ.SQ020.)

#also dynamic characteristics
#communication frequency
df_freq <- data.frame(p1 = data$freq.SQ001.,p2 = data$freq.SQ002.,p3 = data$freq.SQ003.,p4 = data$freq.SQ004.,p5 = data$freq.SQ005.,p6 = data$freq.SQ006.,p7 = data$freq.SQ007.,p8 = data$freq.SQ008., p9 = data$freq.SQ009.,p10= data$freq.SQ010., p11= data$freq.SQ011., p12= data$freq.SQ012., p13= data$freq.SQ013., p14= data$freq.SQ014., p15= data$freq.SQ015., p16= data$freq.SQ016., p17= data$freq.SQ017., p18= data$freq.SQ018., p19= data$freq.SQ019., p20= data$freq.SQ020.)

#closeness
df_close <- data.frame(p1 = data$close.SQ001.,p2 = data$close.SQ002.,p3 = data$close.SQ003.,p4 = data$close.SQ004.,p5 = data$close.SQ005.,p6 = data$close.SQ006.,p7 = data$close.SQ007.,p8 = data$close.SQ008., p9 = data$close.SQ009.,p10= data$close.SQ010., p11= data$close.SQ011., p12= data$close.SQ012., p13= data$close.SQ013., p14= data$close.SQ014., p15= data$close.SQ015., p16= data$close.SQ016., p17= data$close.SQ017., p18= data$close.SQ018., p19= data$close.SQ019., p20= data$close.SQ020.)

#and other dyadic featuers
# duration;
df_duration <- data.frame(p1 = data$duur.SQ001.,p2 = data$duur.SQ002.,p3 = data$duur.SQ003.,p4 = data$duur.SQ004.,p5 = data$duur.SQ005.,p6 = data$duur.SQ006.,p7 = data$duur.SQ007.,p8 = data$duur.SQ008., p9 = data$duur.SQ009.,p10= data$duur.SQ010., p11= data$duur.SQ011., p12= data$duur.SQ012., p13= data$duur.SQ013., p14= data$duur.SQ014., p15= data$duur.SQ015., p16= data$duur.SQ016., p17= data$duur.SQ017., p18= data$duur.SQ018., p19= data$duur.SQ019., p20= data$duur.SQ020.)

#geographical proximity
df_proximity <- data.frame(p1 = data$prox.SQ001.,p2 = data$prox.SQ002.,p3 = data$prox.SQ003.,p4 = data$prox.SQ004.,p5 = data$prox.SQ005.,p6 = data$prox.SQ006.,p7 = data$prox.SQ007.,p8 = data$prox.SQ008., p9 = data$prox.SQ009.,p10= data$prox.SQ010., p11= data$prox.SQ011., p12= data$prox.SQ012., p13= data$prox.SQ013., p14= data$prox.SQ014., p15= data$prox.SQ015., p16= data$prox.SQ016., p17= data$prox.SQ017., p18= data$prox.SQ018., p19= data$prox.SQ019., p20= data$prox.SQ020.)

for (i in 1:length(alterL)) {
  
  #gender
  alterL[[i]]$alter_gender <- 
    ifelse(unlist(df_gender[i,], use.names=F)=="Man", 0,  # male = ref.
           ifelse(unlist(df_gender[i,], use.names=F)=="Vrouw", 1,
                  ifelse(unlist(df_gender[i,], use.names=F)=="Anders", 2, "")))
  #kin  
  alterL[[i]]$kin <- 
    ifelse(unlist(df_kin[i,], use.names=F)=="Ja", 1,
           ifelse(unlist(df_kin[i,], use.names=F)=="Nee", 0, ""))
  #age  
  alterL[[i]]$alter_age <- 
    ifelse(unlist(df_age[i,], use.names=F)=="Jonger dan 18 jaar", 16, #???
           ifelse(unlist(df_age[i,], use.names=F)=="18 tot 21 jaar", 20,
                  ifelse(unlist(df_age[i,], use.names=F)=="22 tot 25 jaar", 23,
                         ifelse(unlist(df_age[i,], use.names=F)=="26 tot 30 jaar", 28,
                                ifelse(unlist(df_age[i,], use.names=F)=="31 tot 40 jaar", 35,
                                       ifelse(unlist(df_age[i,], use.names=F)=="Ouder dan 40 jaar", 45, #???
                                              ifelse(unlist(df_age[i,], use.names=F)=="Weet ik niet", NA, # i don't know = missing
                                                     unlist(df_age[i,], use.names=F))))))))
  #education
  alterL[[i]]$alter_educ <- ifelse(unlist(df_educ[i,],use.names=F)=="lagere school", 1,
                               ifelse(unlist(df_educ[i,],use.names=F)=="vmbo, mavo", 2,
                                      ifelse(unlist(df_educ[i,],use.names=F)=="mbo", 3,
                                         ifelse(unlist(df_educ[i,],use.names=F)=="havo", 4,
                                            ifelse(unlist(df_educ[i,],use.names=F)=="vwo / gymnasium", 5,
                                                 ifelse(unlist(df_educ[i,],use.names=F)=="hbo", 6,
                                                     ifelse(unlist(df_educ[i,],use.names=F)=="universiteit", 7, NA))))))) 
  #frequency
  alterL[[i]]$frequency.t1 <- ifelse(unlist(df_freq[i,],use.names=F)=="(Bijna) elke dag", 7,
                               ifelse(unlist(df_freq[i,],use.names=F)=="1-2 keer per week",6,
                                      ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per maand",5, 
                                         ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per maand",4, 
                                            ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per jaar",3,
                                                 ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per jaar",2,
                                                     ifelse(unlist(df_freq[i,],use.names=F)=="Nooit",1, NA )))))))
  #closeness
  alterL[[i]]$closeness.t1 <- ifelse(unlist(df_close[i,],use.names=F)=="Niet hecht", 1, 
                               ifelse(unlist(df_close[i,],use.names=F)=="Enigszins hecht", 2,
                                      ifelse(unlist(df_close[i,],use.names=F)=="Hecht", 3,
                                             ifelse(unlist(df_close[i,],use.names=F)=="Heel erg hecht", 4, NA ))))
  #proximity
  alterL[[i]]$proximity <- ifelse(unlist(df_proximity[i,], use.names=FALSE) == "In hetzelfde huis", "roommate",
                                  ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In dezelfde buurt" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde straat" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde gemeente", "close",
                                    ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In hetzelfde land" |
                                             unlist(df_proximity[i,], use.names = FALSE) == "In een ander land","far", NA)))

  
   #duration:  take midpoint on scale.
   alterL[[i]]$duration <- ifelse(unlist(df_duration[i,],use.names=F)=="Minder dan 1 jaar", 0, 
                               ifelse(unlist(df_duration[i,],use.names=F)=="1 tot 3 jaar", 2,
                                      ifelse(unlist(df_duration[i,],use.names=F)=="4 tot 8 jaar", 6,
                                             ifelse(unlist(df_duration[i,],use.names=F)=="9 tot 15 jaar", 12,
                                                    ifelse(unlist(df_duration[i,],use.names=F)=="Meer dan 15 jaar", 15,NA )))))
}

# before we can construct similarity indices, we must add ego-attributes
for (i in 1:length(alterL)) {
  #  gender
  alterL[[i]]$ego_gender <- ifelse(data$A1[i] == "Man", 0,ifelse(data$A1[i] == "Vrouw", 1,ifelse(data$A1[i] == "Overige", 2,NA)))
  
  # education 
  # these score should correspond to scores given to alter educ...
  alterL[[i]]$ego_educ <- ifelse(data$T1[i] == "aan de Hogeschool van Arnhem en Nijmegen (HAN)", "6", "7" )
 
  # age
  # calculated as the difference between submission date and birthdate
  # in years.
  birth_date <- as.Date(data$A2[i])
  #x_date <- as.Date(data$submitdate) #submission date was not deposited; so, take 1-jan 2023
  x_date <- as.Date("2023-01-01")
  alterL[[i]]$ego_age <- trunc((birth_date %--% x_date) / years(1))
}

# i filter out the unique, non-duplicate alters;
# by excluding alters with empty strings for gender attribute
for ( i in 1:length(alterL)) {
  alterL[[i]] <- alterL[[i]][which(alterL[[i]]$alter_gender!=""),]
  # and replace the alter id: 1 : no. unique alters
  alterL[[i]]$alterid <- 1:nrow(alterL[[i]])
}

#dyadic similarity measures
for (i in 1:length(alterL))  { # for ego i
  # get attributes of ego
  agei <- alterL[[i]]$ego_age[1]
  genderi <- alterL[[i]]$ego_gender[1]
  edi <- alterL[[i]]$ego_educ[1]
  
  for (j in 1:max(alterL[[i]]$alterid)) { # for alter j
    # calculate "same gender" (0/1)
    genderj <- as.numeric(alterL[[i]]$alter_gender[j]) # get alter j gender
    same <- ifelse(genderi==genderj, 1, 0)
    alterL[[i]]$same_gender[which(alterL[[i]]$alterid==j)] <- same
    
    #same education
    eduj <- alterL[[i]]$alter_educ[j]
    same <- ifelse(eduj==edi, 1, 0)
    alterL[[i]]$sim_educ[which(alterL[[i]]$alterid==j)] <- same
    
    # calculate similarity for age as the absolute difference between alter and ego value
    #get alter j attributes
    agej <- as.numeric(alterL[[i]]$alter_age[j])
    #difference score
    difage <- abs(agei - agej)
    
    # so higher values represent *less* similarity
    if( !is.na (agej) ) { # age similarity only calculable if z_j is known!
        alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- difage
      } else { alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- NA}
  }
}

#based on this, i make a long dataframe with alters nested in ego.
#first, add an ego_id
for (i in 1:length(alterL)) {
  alterL[[i]]$ego <- i
  alterL[[i]]$respnr <- data$respnr[i]
}

#combine using rbind
df <- do.call(rbind,alterL)

#other ego variables:
#transitions
# we need info of waves 1 and 2, so i merge them
d <- merge(data3, data4, by="respnr")
d <- data.frame(d[order(d$respnr, decreasing = FALSE), ]) # and sort by respnr

#1. residential transitions
df$housing.t1a <- NA #wave 1, pre-transition
df$housing.t1b <- NA #wave 1, post-transition
df$housing.t2 <- NA #wave 2
df$housing.transition <- NA #no. of changes 

#make sure descriptions of living situation matches over waves
d$A7 <- ifelse(d$A7 == "inwonend bij ouder(s)/verzorger(s)", "inwonend bij je ouder(s)/verzorger(s)", d$A7)
d$A4d <- ifelse(d$A4d == "inwonend bij ouder(s)/verzorger(s)", "inwonend bij je ouder(s)/verzorger(s)", d$A4d)

for (i in unique(df$ego)) {
  df$housing.t1a[which(df$ego == i)] <- d$A4.x[i] # w1, half year prior to transition
    
  # if respondent had moved at the time of w1, get current living situation
  df$housing.t1b[which(df$ego==i)] <- ifelse(d$A6[i] == "Nee", d$A7[i], df$housing.t1a[which(df$ego==i)])
  
  #if they moved, this indicates a transition
  transition <- ifelse(d$A6[i]=="Nee", 1, 0)
  
  # if respondent has moved since then (in w2), ...
  df$housing.t2[which(df$ego==i)] <- ifelse(d$A4.y[i] == "Ja" | d$A4b[i] == "Ja" | d$A4c[i] == "Ja", d$A4d[i], df$housing.t1b[which(df$ego==i)]) 
  
  transition <- ifelse(d$A4.y[i] == "Ja" | d$A4b[i] == "Ja" | d$A4c[i] == "Ja", transition + 1, transition)
  
  # no. of housing transition
  df$housing.transition[which(df$ego==i)] <- transition
}

#also make it binary (yes/no)
df$housing.transition_bin <- ifelse(df$housing.transition > 0, 1, 0)

#2. transition to university
df$occupation.t1 <- NA
df$occupation.t2 <- "higher" #everybody is in school...
df$study.year <- NA
df$occupation.transition <- NA 

for (i in unique(df$ego)) {
  df$occupation.t1[which(df$ego==i)] <- ifelse(data$E1[i]=="Ik zat op de middelbare school", "secondary",
                             ifelse(data$E1[i]=="Ik deed een MBO-, HBO- of WO-opleiding","higher",
                                    ifelse(data$E1[i]=="Ik werkte","work",
                                             ifelse(data$E1[i]=="Ik had een tussenjaar","gap",
                                                    ifelse(data$E1[i]=="Overige","other",NA)))))

  df$study.year[which(df$ego==i)] <- ifelse(data$T3[i]=="in het eerste jaar", 1,
                                            ifelse(data$T3[i]=="in het tweede jaar", 2,
                                                   ifelse(data$T3[i]=="in het derde jaar of hoger", 3, NA)))
  
  #transition = 1 for those who are first years and who came from secondary school/gap year
  transition <- ifelse( data$T3[i]=="in het eerste jaar" & (data$E1[i]=="Ik zat op de middelbare school" | data$E2[i]=="Ik had een tussenjaar"), 1, 0)
  
  # did ego drop out/start a new study?
  transition <- ifelse(data4$G07Q90[i]== "Nee", transition + 1, transition)
  df$occupation.transition[which(df$ego==i)] <- transition
}

#also make binary
df$occupation.transition_bin <- ifelse(df$occupation.transition > 0, 1, 0)

#romantic relationship 
df$romantic <- NA
for (i in unique(df$ego)) {
  df$romantic[which(df$ego==i)] <- ifelse(data$R3[i] == "Ja", 1,0)
}

#psychological/other variables (wave 2)
#loneliness  
l <- cbind(d$stellingen.SQ001., d$stellingen.SQ002.)
l <- ifelse(l=="Helemaal oneens", 1, ifelse(l=="Mee oneens", 2,
                   # do not know = neutral... this category was sparsely populated 
                   ifelse(l=="Neutraal" | l=="Ik weet het niet", 3, 
                          ifelse(l=="Mee eens", 4, ifelse(l=="Helemaal eens", 5, l)))))
# turn the first indicator, so that higher scores indicate "more" loneliness.
l[,1] <- 6-as.numeric(l[,1])

#extraversion
e <- cbind(d$stellingen.SQ003., d$stellingen.SQ006.)
e <- ifelse(e=="Helemaal oneens", 1,
            ifelse(e=="Mee oneens", 2,
                   ifelse(e=="Neutraal" | e=="Ik weet het niet", 3, 
                          ifelse(e=="Mee eens", 4,
                                 ifelse(e=="Helemaal eens", 5, e)))))
# turn the second indicator
e[,2] <- 6-as.numeric(e[,2])

# calculate averages.
df$loneliness <- NA
df$extraversion <- NA
for (i in unique(df$ego)) {
  l_scores <- as.numeric(l[i,])
  e_scores <- as.numeric(e[i,])
  df$loneliness[which(df$ego==i)] <- mean(l_scores)
  df$extraversion[which(df$ego==i)] <- mean(e_scores)
}

# financial restrictions (for social interaction)
df$fin_restr <- NA
# 0. never; 1. sometimes; 2. often; 3. always
for (i in unique(df$ego)) {
  df$fin_restr[which(df$ego==i)] <- d$G01Q111[i]
}
df$fin_restr <- ifelse(df$fin_restr=="Nooit",0,
                       ifelse(df$fin_restr=="Soms",1,
                              ifelse(df$fin_restr=="Vaak",2,
                                     ifelse(df$fin_restr=="Altijd",3, NA))))

# parental SES / educational level:
for (i in unique(df$ego)) {
  df$educ_father[which(df$ego == i)] <- data4$G08Q56.SQ001.[i]
  df$educ_mother[which(df$ego == i)] <- data4$G08Q56.SQ002.[i]
}

# add network variables
# 1. size (no. of unique alters)
df$netsize <- NA
# also per egonet
df$cdn.size <- df$study.size <- df$csn.size <- df$bff.size <- 0

for (i in unique(df$ego)) {
  #no. of unique alters is simply the number of rows per ego
  df$netsize[which(df$ego==i)] <- nrow(df[which(df$ego==i),])
  
  #size of each egonet is the number of alters named in the name generators
  #thus, including 'duplicates'
  df$cdn.size[which(df$ego==i)] <- length(which(!df_names[i,c(1:5)]==""))
  df$study.size[which(df$ego==i)] <- length(which(!df_names[i,c(6:10)]==""))
  df$bff.size[which(df$ego==i)] <- length(which(!df_names[i,c(11:15)]==""))
  df$csn.size[which(df$ego==i)] <- length(which(!df_names[i,c(16:20)]==""))
}

# 2. density.
# this can only be calculated per egonet, as only between-alter ties within egonet are asked.
df$cdn.density <- df$study.density <- df$csn.density <- df$bff.density <- NA

for (i in unique(df$ego)) {
  #a. CDN
  size <- df$cdn.size[which(df$ego==i)][1]
  #no. possible ties between alters
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  #get no. of observed ties
  #only if ego named at least 2 alters in this name generator
  #because only then i can calculate density
  if(size>1) {
    #get alter-alter adjacency matrices of this particular egonet
    adjacency <- data[i,c(13:32)]
   
    #i want the correct ones, given the no. of listed alters in this name generator
    #i list the columns numbers of the variables corresponding to each matching matrix set,
    #and subset corresponding columns in `adjacency`
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$cdn.density[which(df$ego==i)] <- obs/pos  }
  
  #b. study
  size <- df$study.size[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data[i,c(53:72)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
  
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$study.density[which(df$ego==i)] <- obs/pos }
  
  #c. friends
  size <- df$bff.size[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data[i,c(169:188)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$bff.density[which(df$ego==i)] <- obs/pos }

  #d. sports partners
  size <- df$csn.size[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data[i,c(444:463)]

    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$csn.density[which(df$ego==i)] <- obs/pos }
}

#set NAs to 0.
df$cdn.density[is.na(df$cdn.density)] <- 0
df$study.density[is.na(df$study.density)] <- 0
df$bff.density[is.na(df$bff.density)] <- 0
df$csn.density[is.na(df$csn.density)] <- 0

#binary attributes indicating whether alters appeared in each of the ego-nets at w1.
for (i in unique(df$ego)) {  
    for (j in 1:nrow(df[which(df$ego==i),])) { # for alters nested in ego

    # find out if alter_id denoting alter j appear in the 4 egonets
    alter <- unlist(df[which(df$ego==i & df$alterid==j),][5:8], use.names = FALSE) # get name(s) of alter j
    alter <- alter[!is.na(alter)] # exclude NAs
    
    cdn <- alter %in% net1[i,]
    study <- alter %in% net2[i,]
    bff <- alter %in% net3[i,]
    csn <- alter %in% net4[i,]
    
    # and if so, give alter j score 1 on indicators; 0 otherwise
    df$cdn1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% cdn, 1, 0)
    df$study1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% study, 1, 0)
    df$bff1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% bff, 1, 0)
    df$csn1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% csn, 1, 0)
  }
}

#calculate multiplexity as an alter/tie attribute (i.e., number of *additional* networks j appeared in)
df$multiplex <- rowSums(df[,c(51:54)]) - 1 #minus one to reach a meaningful intercept
#psych::describe(df$multiplex) #M=.51; SD=.76

#now, find out whether w1-alters were maintainted in w2.
df$survive <- NA

# In wave 2, another matching procedure was used; in rows, alters named in w2 are listed; columns denote (unique) alters of w1, so the same one as our data-frame.
# again, i made multiple matrices, conditional on the no. of unique alters of w1 (15, to be precise)
# if, in one of these matrices, column j was marked as corresponding to a named alter in w2, then alter j did, indeed, survive

# subset w1-w2 matching matrices; and respnr
w1w2 <- data4[,c(756:1155,length(data4))]

for (i in unique(df$ego)) {  # for ego 
  matchingL <- vector("list", 20) #pre-allocate empty list of length 20, to store matching matrices

  {
    matchingL[[1]] <- w1w2[i,1:20]
    matchingL[[2]] <- w1w2[i,21:40]
    matchingL[[3]] <- w1w2[i,41:60]
    matchingL[[4]] <- w1w2[i,61:80]
    matchingL[[5]] <- w1w2[i,81:100]
    matchingL[[6]] <- w1w2[i,101:120]
    matchingL[[7]] <- w1w2[i,121:140]
    matchingL[[8]] <- w1w2[i,141:160]
    matchingL[[9]] <- w1w2[i,161:180]
    matchingL[[10]] <- w1w2[i,181:200]
    matchingL[[11]] <- w1w2[i,201:220]
    matchingL[[12]] <- w1w2[i,221:240]
    matchingL[[13]] <- w1w2[i,241:260]
    matchingL[[14]] <- w1w2[i,261:280]
    matchingL[[15]] <- w1w2[i,281:300]
    matchingL[[16]] <- w1w2[i,301:320]
    matchingL[[17]] <- w1w2[i,321:340]
    matchingL[[18]] <- w1w2[i,341:360]
    matchingL[[19]] <- w1w2[i,361:380]
    matchingL[[20]] <- w1w2[i,381:400]
  }
  
  #find the 'right' matching matrix in this list, by taking the one that contains answers
  ind <- NULL
  for (j in seq_along(matchingL)) {
    #check along the sequence of elements j in the matching matrix list if they are non-empty and not NA
    check <- FALSE
    for (col in matchingL[[j]]) {
      if (!all(is.na(col)) && any(nchar(col) > 0)) {
        check <- TRUE
        break  
      }
      }
    if (check) {
      ind <- j
      break  
    }
  }
  if(length(ind)>0) {
    # get the  corresponding matrix
    mm <- matchingL[[ind]]
    
    # unlist the answers given
    ans <- unlist(mm, use.names = FALSE)
    
    # use 'stringr' to extract numbers in these answers,
    # they refer to the id of "survived" alters
    id <- unlist(stringr::str_extract_all(ans,"\\(?[0-9,.]+\\)?"))
    id <- id[!is.na(id)] # exclude NA
    
    # if alter id of ego i corresponds to this id, they did indeed survive:
    df$survive[which(df$ego==i)] <- ifelse(df$alterid[which(df$ego==i)] %in% id, 1, 0)
  }
}

# in which egonet did they (re)appear?
df$cdn2 <- NA
df$study2 <- NA
df$bff2 <- NA
df$csn2 <- NA

# i use the matching matrices of w2; where columns refer to unique w1 alters (those of our constructed dataframe) and rows referring to unique w2 alters
# so, if w1-alter/column j is matched to w2-alter/row i, then w1-alter did survive (but we already knew that); but more importantly, j was the i^th alter mentioned
# and this allows me to infer in what network j (re-)appeared
# since cdn = 1-5; study = 6-10; bff = 11-15; csn = 16-20.
# however, since the rows only include non-duplicate w2-alters;
# if alter j at t1 was named more than once at t2, i only know the first
# egonet in which j was named...
# thus, before i proceed, i use the matching matrices that allowed ego to match alters from different egonets to make alter ids for w2 alters; and figure out to which egonets they belonged

# i make a dataframe of w2-alters, with potential duplicates...
df_names <- data.frame(p1 = data4$egonet1.SQ001.,p2 = data4$egonet1.SQ002.,p3 = data4$egonet1.SQ003.,p4 = data4$egonet1.SQ004.,p5 = data4$egonet1.SQ005.,p6 = data4$egonet2.SQ001.,p7 = data4$egonet2.SQ002.,p8 = data4$egonet2.SQ003.,p9 = data4$egonet2.SQ004.,p10= data4$egonet2.SQ005.,p11= data4$egonet3.SQ001.,p12= data4$egonet3.SQ002.,p13= data4$egonet3.SQ003.,p14= data4$egonet3.SQ004.,p15= data4$egonet3.SQ005.,p16= data4$egonet4.SQ001., p17= data4$egonet4.SQ002., p18= data4$egonet4.SQ003.,p19= data4$egonet4.SQ004.,p20= data4$egonet4.SQ005.)

#males
df_gender <- data.frame(p1 = data4$males.SQ001.,p2 = data4$males.SQ002.,p3 = data4$males.SQ003.,p4 = data4$males.SQ004.,p5 = data4$males.SQ005.,p6 = data4$males.SQ006.,p7 = data4$males.SQ007.,p8 = data4$males.SQ008.,p9 = data4$males.SQ009.,p10= data4$males.SQ010.,p11= data4$males.SQ011.,p12= data4$males.SQ012.,p13= data4$males.SQ013.,p14= data4$males.SQ014.,p15= data4$males.SQ015.,p16= data4$males.SQ016.,p17= data4$males.SQ017.,p18= data4$males.SQ018.,p19= data4$males.SQ019.,p20= data4$males.SQ020.)

# list of dataframes per ego with rows reflecting alters
# and columns indicating the name(s) of the particular alters
alterL <- list()
# loop over all egos
for ( i in 1:nrow(df_names)) {alterL[[i]] <- data.frame(
    alterid = 1:20,name1 = NA,name2 = NA, name3 = NA,name4 = NA, gender=NA,
    cdn_embed.t2 = NA, study_embed.t2 = NA, bff_embed.t2 = NA, csn_embed.t2 = NA)}

# fill the names based on the names data-frame
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=F)
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
  
  alterL[[i]]$gender <- unlist(df_gender[i,], use.names=F)
  alterL[[i]]$gender <- ifelse(alterL[[i]]$gender=="Ja", 0, #male = ref
                               ifelse(alterL[[i]]$gender=="Nee", 1, NA))
  
  }

# calculate netsize for each net to get the correct matching matrix
{
  net1 <- cbind(data4$egonet1.SQ001.,data4$egonet1.SQ002., data4$egonet1.SQ003.,data4$egonet1.SQ004., data4$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data4$egonet2.SQ001.,data4$egonet2.SQ002., data4$egonet2.SQ003.,data4$egonet2.SQ004., data4$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data4$egonet3.SQ001.,data4$egonet3.SQ002., data4$egonet3.SQ003.,data4$egonet3.SQ004., data4$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data4$egonet4.SQ001.,data4$egonet4.SQ002., data4$egonet4.SQ003.,data4$egonet4.SQ004., data4$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}

# construct the matching matrices list for egonet1-2
matchingList <- list()
for (i in 1:length(alterL)) {
  matchingL <- list()
  matchingL[[1]] <- cbind(data4$matching1N1.SQ001_SQ001.[i],data4$matching1N1.SQ002_SQ001.[i],data4$matching1N1.SQ003_SQ001.[i],data4$matching1N1.SQ004_SQ001.[i],data4$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data4$matching1N2.SQ001_SQ001.[i], data4$matching1N2.SQ002_SQ001.[i], data4$matching1N2.SQ003_SQ001.[i], data4$matching1N2.SQ004_SQ001.[i], data4$matching1N2.SQ005_SQ001.[i]),
    cbind(data4$matching1N2.SQ001_SQ002.[i], data4$matching1N2.SQ002_SQ002.[i], data4$matching1N2.SQ003_SQ002.[i], data4$matching1N2.SQ004_SQ002.[i], data4$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data4$matching1N3.SQ001_SQ001.[i], data4$matching1N3.SQ002_SQ001.[i], data4$matching1N3.SQ003_SQ001.[i], data4$matching1N3.SQ004_SQ001.[i], data4$matching1N3.SQ005_SQ001.[i]),
    cbind(data4$matching1N3.SQ001_SQ002.[i], data4$matching1N3.SQ002_SQ002.[i], data4$matching1N3.SQ003_SQ002.[i], data4$matching1N3.SQ004_SQ002.[i], data4$matching1N3.SQ005_SQ002.[i]),
    cbind(data4$matching1N3.SQ001_SQ003.[i], data4$matching1N3.SQ002_SQ003.[i], data4$matching1N3.SQ003_SQ003.[i], data4$matching1N3.SQ004_SQ003.[i], data4$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data4$matching1N4.SQ001_SQ001.[i], data4$matching1N4.SQ002_SQ001.[i], data4$matching1N4.SQ003_SQ001.[i], data4$matching1N4.SQ004_SQ001.[i], data4$matching1N4.SQ005_SQ001.[i]),
    cbind(data4$matching1N4.SQ001_SQ002.[i], data4$matching1N4.SQ002_SQ002.[i], data4$matching1N4.SQ003_SQ002.[i], data4$matching1N4.SQ004_SQ002.[i], data4$matching1N4.SQ005_SQ002.[i]),
    cbind(data4$matching1N4.SQ001_SQ003.[i], data4$matching1N4.SQ002_SQ003.[i], data4$matching1N4.SQ003_SQ003.[i], data4$matching1N4.SQ004_SQ003.[i], data4$matching1N4.SQ005_SQ003.[i]),
    cbind(data4$matching1N4.SQ001_SQ004.[i], data4$matching1N4.SQ002_SQ004.[i], data4$matching1N4.SQ003_SQ004.[i], data4$matching1N4.SQ004_SQ004.[i], data4$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data4$matching1N5.SQ001_SQ001.[i], data4$matching1N5.SQ002_SQ001.[i], data4$matching1N5.SQ003_SQ001.[i], data4$matching1N5.SQ004_SQ001.[i], data4$matching1N5.SQ005_SQ001.[i]),
    cbind(data4$matching1N5.SQ001_SQ002.[i], data4$matching1N5.SQ002_SQ002.[i], data4$matching1N5.SQ003_SQ002.[i], data4$matching1N5.SQ004_SQ002.[i], data4$matching1N5.SQ005_SQ002.[i]),
    cbind(data4$matching1N5.SQ001_SQ003.[i], data4$matching1N5.SQ002_SQ003.[i], data4$matching1N5.SQ003_SQ003.[i], data4$matching1N5.SQ004_SQ003.[i], data4$matching1N5.SQ005_SQ003.[i]),
    cbind(data4$matching1N5.SQ001_SQ004.[i], data4$matching1N5.SQ002_SQ004.[i], data4$matching1N5.SQ003_SQ004.[i], data4$matching1N5.SQ004_SQ004.[i], data4$matching1N5.SQ005_SQ004.[i]),
    cbind(data4$matching1N5.SQ001_SQ005.[i], data4$matching1N5.SQ002_SQ005.[i], data4$matching1N5.SQ003_SQ005.[i], data4$matching1N5.SQ004_SQ005.[i], data4$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
} # so... matchingL[[1]][[5]] is matchingmatrix 5 (i.e., 5 alters named in egonet2) for ego 1

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
}#ignore warning

# again, make a matching list for the second matching matrices
# (i.e., matching egonet3 alters to prev. alters);
matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data4$matching2N1.SQ001_SQ001.[i],data4$matching2N1.SQ002_SQ001.[i],data4$matching2N1.SQ003_SQ001.[i],data4$matching2N1.SQ004_SQ001.[i],data4$matching2N1.SQ005_SQ001.[i],data4$matching2N1.SQ006_SQ001.[i],data4$matching2N1.SQ007_SQ001.[i],data4$matching2N1.SQ008_SQ001.[i],data4$matching2N1.SQ009_SQ001.[i],data4$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data4$matching2N2.SQ001_SQ001.[i],data4$matching2N2.SQ002_SQ001.[i],data4$matching2N2.SQ003_SQ001.[i],data4$matching2N2.SQ004_SQ001.[i],data4$matching2N2.SQ005_SQ001.[i],data4$matching2N2.SQ006_SQ001.[i],data4$matching2N2.SQ007_SQ001.[i],data4$matching2N2.SQ008_SQ001.[i],data4$matching2N2.SQ009_SQ001.[i],data4$matching2N2.SQ010_SQ001.[i]),
    cbind(data4$matching2N2.SQ001_SQ002.[i],data4$matching2N2.SQ002_SQ002.[i],data4$matching2N2.SQ003_SQ002.[i],data4$matching2N2.SQ004_SQ002.[i],data4$matching2N2.SQ005_SQ002.[i],data4$matching2N2.SQ006_SQ002.[i],data4$matching2N2.SQ007_SQ002.[i],data4$matching2N2.SQ008_SQ002.[i],data4$matching2N2.SQ009_SQ002.[i],data4$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data4$matching2N3.SQ001_SQ001.[i],data4$matching2N3.SQ002_SQ001.[i],data4$matching2N3.SQ003_SQ001.[i],data4$matching2N3.SQ004_SQ001.[i],data4$matching2N3.SQ005_SQ001.[i],data4$matching2N3.SQ006_SQ001.[i],data4$matching2N3.SQ007_SQ001.[i],data4$matching2N3.SQ008_SQ001.[i],data4$matching2N3.SQ009_SQ001.[i],data4$matching2N3.SQ010_SQ001.[i]),
    cbind(data4$matching2N3.SQ001_SQ002.[i],data4$matching2N3.SQ002_SQ002.[i],data4$matching2N3.SQ003_SQ002.[i],data4$matching2N3.SQ004_SQ002.[i],data4$matching2N3.SQ005_SQ002.[i],data4$matching2N3.SQ006_SQ002.[i],data4$matching2N3.SQ007_SQ002.[i],data4$matching2N3.SQ008_SQ002.[i],data4$matching2N3.SQ009_SQ002.[i],data4$matching2N3.SQ010_SQ002.[i]),
    cbind(data4$matching2N3.SQ001_SQ003.[i],data4$matching2N3.SQ002_SQ003.[i],data4$matching2N3.SQ003_SQ003.[i],data4$matching2N3.SQ004_SQ003.[i],data4$matching2N3.SQ005_SQ003.[i],data4$matching2N3.SQ006_SQ003.[i],data4$matching2N3.SQ007_SQ003.[i],data4$matching2N3.SQ008_SQ003.[i],data4$matching2N3.SQ009_SQ003.[i],data4$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data4$matching2N4.SQ001_SQ001.[i],data4$matching2N4.SQ002_SQ001.[i],data4$matching2N4.SQ003_SQ001.[i],data4$matching2N4.SQ004_SQ001.[i],data4$matching2N4.SQ005_SQ001.[i],data4$matching2N4.SQ006_SQ001.[i],data4$matching2N4.SQ007_SQ001.[i],data4$matching2N4.SQ008_SQ001.[i],data4$matching2N4.SQ009_SQ001.[i],data4$matching2N4.SQ010_SQ001.[i]),
    cbind(data4$matching2N4.SQ001_SQ002.[i],data4$matching2N4.SQ002_SQ002.[i],data4$matching2N4.SQ003_SQ002.[i],data4$matching2N4.SQ004_SQ002.[i],data4$matching2N4.SQ005_SQ002.[i],data4$matching2N4.SQ006_SQ002.[i],data4$matching2N4.SQ007_SQ002.[i],data4$matching2N4.SQ008_SQ002.[i],data4$matching2N4.SQ009_SQ002.[i],data4$matching2N4.SQ010_SQ002.[i]),
    cbind(data4$matching2N4.SQ001_SQ003.[i],data4$matching2N4.SQ002_SQ003.[i],data4$matching2N4.SQ003_SQ003.[i],data4$matching2N4.SQ004_SQ003.[i],data4$matching2N4.SQ005_SQ003.[i],data4$matching2N4.SQ006_SQ003.[i],data4$matching2N4.SQ007_SQ003.[i],data4$matching2N4.SQ008_SQ003.[i],data4$matching2N4.SQ009_SQ003.[i],data4$matching2N4.SQ010_SQ003.[i]),
    cbind(data4$matching2N4.SQ001_SQ004.[i],data4$matching2N4.SQ002_SQ004.[i],data4$matching2N4.SQ003_SQ004.[i],data4$matching2N4.SQ004_SQ004.[i],data4$matching2N4.SQ005_SQ004.[i],data4$matching2N4.SQ006_SQ004.[i],data4$matching2N4.SQ007_SQ004.[i],data4$matching2N4.SQ008_SQ004.[i],data4$matching2N4.SQ009_SQ004.[i],data4$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data4$matching2N5.SQ001_SQ001.[i],data4$matching2N5.SQ002_SQ001.[i],data4$matching2N5.SQ003_SQ001.[i],data4$matching2N5.SQ004_SQ001.[i],data4$matching2N5.SQ005_SQ001.[i],data4$matching2N5.SQ006_SQ001.[i],data4$matching2N5.SQ007_SQ001.[i],data4$matching2N5.SQ008_SQ001.[i],data4$matching2N5.SQ009_SQ001.[i],data4$matching2N5.SQ010_SQ001.[i]),
    cbind(data4$matching2N5.SQ001_SQ002.[i],data4$matching2N5.SQ002_SQ002.[i],data4$matching2N5.SQ003_SQ002.[i],data4$matching2N5.SQ004_SQ002.[i],data4$matching2N5.SQ005_SQ002.[i],data4$matching2N5.SQ006_SQ002.[i],data4$matching2N5.SQ007_SQ002.[i],data4$matching2N5.SQ008_SQ002.[i],data4$matching2N5.SQ009_SQ002.[i],data4$matching2N5.SQ010_SQ002.[i]),
    cbind(data4$matching2N5.SQ001_SQ003.[i],data4$matching2N5.SQ002_SQ003.[i],data4$matching2N5.SQ003_SQ003.[i],data4$matching2N5.SQ004_SQ003.[i],data4$matching2N5.SQ005_SQ003.[i],data4$matching2N5.SQ006_SQ003.[i],data4$matching2N5.SQ007_SQ003.[i],data4$matching2N5.SQ008_SQ003.[i],data4$matching2N5.SQ009_SQ003.[i],data4$matching2N5.SQ010_SQ003.[i]),
    cbind(data4$matching2N5.SQ001_SQ004.[i],data4$matching2N5.SQ002_SQ004.[i],data4$matching2N5.SQ003_SQ004.[i],data4$matching2N5.SQ004_SQ004.[i],data4$matching2N5.SQ005_SQ004.[i],data4$matching2N5.SQ006_SQ004.[i],data4$matching2N5.SQ007_SQ004.[i],data4$matching2N5.SQ008_SQ004.[i],data4$matching2N5.SQ009_SQ004.[i],data4$matching2N5.SQ010_SQ004.[i]),
    cbind(data4$matching2N5.SQ001_SQ005.[i],data4$matching2N5.SQ002_SQ005.[i],data4$matching2N5.SQ003_SQ005.[i],data4$matching2N5.SQ004_SQ005.[i],data4$matching2N5.SQ005_SQ005.[i],data4$matching2N5.SQ006_SQ005.[i],data4$matching2N5.SQ007_SQ005.[i],data4$matching2N5.SQ008_SQ005.[i],data4$matching2N5.SQ009_SQ005.[i],data4$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

# same for matching 3 (i.e., egonet4 with egonets 1-3)
matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data4$matching3N1.SQ001_SQ001.[i],data4$matching3N1.SQ002_SQ001.[i],data4$matching3N1.SQ003_SQ001.[i],data4$matching3N1.SQ004_SQ001.[i],data4$matching3N1.SQ005_SQ001.[i],data4$matching3N1.SQ006_SQ001.[i],data4$matching3N1.SQ007_SQ001.[i],data4$matching3N1.SQ008_SQ001.[i],data4$matching3N1.SQ009_SQ001.[i],data4$matching3N1.SQ010_SQ001.[i], data4$matching3N1.SQ011_SQ001.[i], data4$matching3N1.SQ012_SQ001.[i], data4$matching3N1.SQ013_SQ001.[i], data4$matching3N1.SQ014_SQ001.[i], data4$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data4$matching3N2.SQ001_SQ001.[i],data4$matching3N2.SQ002_SQ001.[i],data4$matching3N2.SQ003_SQ001.[i],data4$matching3N2.SQ004_SQ001.[i],data4$matching3N2.SQ005_SQ001.[i],data4$matching3N2.SQ006_SQ001.[i],data4$matching3N2.SQ007_SQ001.[i],data4$matching3N2.SQ008_SQ001.[i],data4$matching3N2.SQ009_SQ001.[i],data4$matching3N2.SQ010_SQ001.[i], data4$matching3N2.SQ011_SQ001.[i], data4$matching3N2.SQ012_SQ001.[i], data4$matching3N2.SQ013_SQ001.[i], data4$matching3N2.SQ014_SQ001.[i], data4$matching3N2.SQ015_SQ001.[i]),
    cbind(data4$matching3N2.SQ001_SQ002.[i],data4$matching3N2.SQ002_SQ002.[i],data4$matching3N2.SQ003_SQ002.[i],data4$matching3N2.SQ004_SQ002.[i],data4$matching3N2.SQ005_SQ002.[i],data4$matching3N2.SQ006_SQ002.[i],data4$matching3N2.SQ007_SQ002.[i],data4$matching3N2.SQ008_SQ002.[i],data4$matching3N2.SQ009_SQ002.[i],data4$matching3N2.SQ010_SQ002.[i], data4$matching3N2.SQ011_SQ002.[i], data4$matching3N2.SQ012_SQ002.[i], data4$matching3N2.SQ013_SQ002.[i], data4$matching3N2.SQ014_SQ002.[i], data4$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data4$matching3N3.SQ001_SQ001.[i],data4$matching3N3.SQ002_SQ001.[i],data4$matching3N3.SQ003_SQ001.[i],data4$matching3N3.SQ004_SQ001.[i],data4$matching3N3.SQ005_SQ001.[i],data4$matching3N3.SQ006_SQ001.[i],data4$matching3N3.SQ007_SQ001.[i],data4$matching3N3.SQ008_SQ001.[i],data4$matching3N3.SQ009_SQ001.[i],data4$matching3N3.SQ010_SQ001.[i], data4$matching3N3.SQ011_SQ001.[i], data4$matching3N3.SQ012_SQ001.[i], data4$matching3N3.SQ013_SQ001.[i], data4$matching3N3.SQ014_SQ001.[i], data4$matching3N3.SQ015_SQ001.[i]),
    cbind(data4$matching3N3.SQ001_SQ002.[i],data4$matching3N3.SQ002_SQ002.[i],data4$matching3N3.SQ003_SQ002.[i],data4$matching3N3.SQ004_SQ002.[i],data4$matching3N3.SQ005_SQ002.[i],data4$matching3N3.SQ006_SQ002.[i],data4$matching3N3.SQ007_SQ002.[i],data4$matching3N3.SQ008_SQ002.[i],data4$matching3N3.SQ009_SQ002.[i],data4$matching3N3.SQ010_SQ002.[i], data4$matching3N3.SQ011_SQ002.[i], data4$matching3N3.SQ012_SQ002.[i], data4$matching3N3.SQ013_SQ002.[i], data4$matching3N3.SQ014_SQ002.[i], data4$matching3N3.SQ015_SQ002.[i]),
    cbind(data4$matching3N3.SQ001_SQ003.[i],data4$matching3N3.SQ002_SQ003.[i],data4$matching3N3.SQ003_SQ003.[i],data4$matching3N3.SQ004_SQ003.[i],data4$matching3N3.SQ005_SQ003.[i],data4$matching3N3.SQ006_SQ003.[i],data4$matching3N3.SQ007_SQ003.[i],data4$matching3N3.SQ008_SQ003.[i],data4$matching3N3.SQ009_SQ003.[i],data4$matching3N3.SQ010_SQ003.[i], data4$matching3N3.SQ011_SQ003.[i], data4$matching3N3.SQ012_SQ003.[i], data4$matching3N3.SQ013_SQ003.[i], data4$matching3N3.SQ014_SQ003.[i], data4$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data4$matching3N4.SQ001_SQ001.[i],data4$matching3N4.SQ002_SQ001.[i],data4$matching3N4.SQ003_SQ001.[i],data4$matching3N4.SQ004_SQ001.[i],data4$matching3N4.SQ005_SQ001.[i],data4$matching3N4.SQ006_SQ001.[i],data4$matching3N4.SQ007_SQ001.[i],data4$matching3N4.SQ008_SQ001.[i],data4$matching3N4.SQ009_SQ001.[i],data4$matching3N4.SQ010_SQ001.[i], data4$matching3N4.SQ011_SQ001.[i], data4$matching3N4.SQ012_SQ001.[i], data4$matching3N4.SQ013_SQ001.[i], data4$matching3N4.SQ014_SQ001.[i], data4$matching3N4.SQ015_SQ001.[i]),
    cbind(data4$matching3N4.SQ001_SQ002.[i],data4$matching3N4.SQ002_SQ002.[i],data4$matching3N4.SQ003_SQ002.[i],data4$matching3N4.SQ004_SQ002.[i],data4$matching3N4.SQ005_SQ002.[i],data4$matching3N4.SQ006_SQ002.[i],data4$matching3N4.SQ007_SQ002.[i],data4$matching3N4.SQ008_SQ002.[i],data4$matching3N4.SQ009_SQ002.[i],data4$matching3N4.SQ010_SQ002.[i], data4$matching3N4.SQ011_SQ002.[i], data4$matching3N4.SQ012_SQ002.[i], data4$matching3N4.SQ013_SQ002.[i], data4$matching3N4.SQ014_SQ002.[i], data4$matching3N4.SQ015_SQ002.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ004.[i],data4$matching3N4.SQ002_SQ004.[i],data4$matching3N4.SQ003_SQ004.[i],data4$matching3N4.SQ004_SQ004.[i],data4$matching3N4.SQ005_SQ004.[i],data4$matching3N4.SQ006_SQ004.[i],data4$matching3N4.SQ007_SQ004.[i],data4$matching3N4.SQ008_SQ004.[i],data4$matching3N4.SQ009_SQ004.[i],data4$matching3N4.SQ010_SQ004.[i], data4$matching3N4.SQ011_SQ004.[i], data4$matching3N4.SQ012_SQ004.[i], data4$matching3N4.SQ013_SQ004.[i], data4$matching3N4.SQ014_SQ004.[i], data4$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data4$matching3N5.SQ001_SQ001.[i],data4$matching3N5.SQ002_SQ001.[i],data4$matching3N5.SQ003_SQ001.[i],data4$matching3N5.SQ004_SQ001.[i],data4$matching3N5.SQ005_SQ001.[i],data4$matching3N5.SQ006_SQ001.[i],data4$matching3N5.SQ007_SQ001.[i],data4$matching3N5.SQ008_SQ001.[i],data4$matching3N5.SQ009_SQ001.[i],data4$matching3N5.SQ010_SQ001.[i], data4$matching3N5.SQ011_SQ001.[i], data4$matching3N5.SQ012_SQ001.[i], data4$matching3N5.SQ013_SQ001.[i], data4$matching3N5.SQ014_SQ001.[i], data4$matching3N5.SQ015_SQ001.[i]),
    cbind(data4$matching3N5.SQ001_SQ002.[i],data4$matching3N5.SQ002_SQ002.[i],data4$matching3N5.SQ003_SQ002.[i],data4$matching3N5.SQ004_SQ002.[i],data4$matching3N5.SQ005_SQ002.[i],data4$matching3N5.SQ006_SQ002.[i],data4$matching3N5.SQ007_SQ002.[i],data4$matching3N5.SQ008_SQ002.[i],data4$matching3N5.SQ009_SQ002.[i],data4$matching3N5.SQ010_SQ002.[i], data4$matching3N5.SQ011_SQ002.[i], data4$matching3N5.SQ012_SQ002.[i], data4$matching3N5.SQ013_SQ002.[i], data4$matching3N5.SQ014_SQ002.[i], data4$matching3N5.SQ015_SQ002.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ005.[i],data4$matching3N5.SQ002_SQ005.[i],data4$matching3N5.SQ003_SQ005.[i],data4$matching3N5.SQ004_SQ005.[i],data4$matching3N5.SQ005_SQ005.[i],data4$matching3N5.SQ006_SQ005.[i],data4$matching3N5.SQ007_SQ005.[i],data4$matching3N5.SQ008_SQ005.[i],data4$matching3N5.SQ009_SQ005.[i],data4$matching3N5.SQ010_SQ005.[i], data4$matching3N5.SQ011_SQ005.[i], data4$matching3N5.SQ012_SQ005.[i], data4$matching3N5.SQ013_SQ005.[i], data4$matching3N5.SQ014_SQ005.[i], data4$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

#let us here also attach structural embeddedness. because the 'network environment' for maintained alters may change in t2

#again, fix the irregularities in the data.
data4 <- data4 %>%
  rename (adj3N2a.SQ001. = adj3N1a.SQ001., adj3N3a.SQ001. = adj3N2a.SQ001.,adj3N3a.SQ002. = adj3N2a.SQ002.,adj3N3b.SQ001. = adj3N2b.SQ001.,adj3N4a.SQ001. = adj3N3a.SQ001.,adj3N4a.SQ002. = adj3N3a.SQ002.,adj3N4a.SQ003. = adj3N3a.SQ003.,adj3N4b.SQ001. = adj3N3b.SQ001.,adj3N4b.SQ002. = adj3N3b.SQ002.,   adj3N4c.SQ001. = adj3N3c.SQ001.,adj3N5a.SQ001. = adj3N4a.SQ001.,adj3N5a.SQ002. = adj3N4a.SQ002.,adj3N5a.SQ003. = adj3N4a.SQ003.,adj3N5a.SQ004. = adj3N4a.SQ004.,adj3N5b.SQ001. = adj3N4b.SQ001.,adj3N5b.SQ002. = adj3N4b.SQ002.,adj3N5b.SQ003. = adj3N4b.SQ003.,adj3N5c.SQ001. = adj3N4c.SQ001.,adj3N5c.SQ002. = adj3N4c.SQ002.,adj3N5d.SQ001. = adj3N4d.SQ001.)

data4$adj4N3b.SQ001. <- data4$adj4N3b.SQ002.

for (i in 1:length(alterL)) { # for ego i

  #get number of names listed in the egonets
  nnames_cdn <- length(which(data4[i,c("egonet1.SQ001.","egonet1.SQ002.","egonet1.SQ003.","egonet1.SQ004.","egonet1.SQ005.")]!=""))
  nnames_study <- length(which(data4[i,c("egonet2.SQ001.","egonet2.SQ002.","egonet2.SQ003.","egonet2.SQ004.","egonet2.SQ005.")]!=""))
  nnames_bff <- length(which(data4[i,c("egonet3.SQ001.","egonet3.SQ002.","egonet3.SQ003.","egonet3.SQ004.","egonet3.SQ005.")]!=""))
  nnames_csn <- length(which(data4[i,c("egonet4.SQ001.","egonet4.SQ002.","egonet4.SQ003.","egonet4.SQ004.","egonet4.SQ005.")]!=""))
  
  #CDN
  
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj1N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj1N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj1N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj1N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj1N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj1N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj1N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj1N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj1N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj1N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj1N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj1N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj1N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj1N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj1N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj1N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj1N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj1N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj1N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj1N5d.SQ001.[i]
  }

  if(nnames_cdn>1) { #we only know alter-alter relations for nnames>1
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_cdn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
  
  #CDN: 1:nnames_CDN
  alterL[[i]]$cdn_embed.t2[1:nnames_cdn] <- embed 
  }
  
  #STUDY

  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj2N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj2N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj2N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj2N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj2N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj2N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj2N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj2N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj2N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj2N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj2N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj2N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj2N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj2N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj2N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj2N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj2N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj2N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj2N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj2N5d.SQ001.[i]
  }

  if(nnames_study>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_study]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$study_embed.t2[6:(5+nnames_study)] <- embed 
  }
  
  # BEST FRIENDS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj3N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj3N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj3N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj3N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj3N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj3N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj3N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj3N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj3N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj3N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj3N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj3N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj3N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj3N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj3N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj3N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj3N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj3N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj3N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj3N5d.SQ001.[i]
  }

  if(nnames_bff>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_bff]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$bff_embed.t2[11:(10+nnames_bff)] <- embed 
  }
  
  # SPORTS PARTNERS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj4N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj4N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj4N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj4N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj4N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj4N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj4N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj4N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj4N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj4N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj4N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj4N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj4N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj4N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj4N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj4N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj4N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj4N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj4N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj4N5d.SQ001.[i]
  }

  if(nnames_csn>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_csn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$csn_embed.t2[16:(15+nnames_csn)] <- embed 
  }
}

#alters could be named in multiple name generators. each element of alterL contains rows corresponding to alters that may be 'duplicates'.
#i match the structural embeddedness measures

for (i in 1:length(alterL)) { #for ego i 
  for (j in 1:nrow(alterL[[i]])) { #for alter j
        
    #if name2 is not empty, alter j was matched to the study network; and more precisely, to the study partner with name1: alterL[[i]]$name2[j]
    alterL[[i]]$study_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name2[j]), 
                                            alterL[[i]]$study_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name2[j] & !is.na(alterL[[i]]$study_embed.t2))],
                                            alterL[[i]]$study_embed.t2[j])
    
    #if name3 is not empty, alter j was matched to the friends network; and more precisely, to the friend with name1: alterL[[i]]$name3[j]
    alterL[[i]]$bff_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name3[j]), 
                                          alterL[[i]]$bff_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name3[j] & !is.na(alterL[[i]]$bff_embed.t2))],
                                          alterL[[i]]$bff_embed.t2[j])
    
    #if name4 is not empty, alter j was matched to the sports network; and more precisely, to the sports partner with name1: alterL[[i]]$name4[j]
    alterL[[i]]$csn_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name4[j]), 
                                          alterL[[i]]$csn_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name4[j] & !is.na(alterL[[i]]$csn_embed.t2))],     
                                          alterL[[i]]$csn_embed.t2[j] )
  }
}

# make long df with alters in ego,
# and indicators for 4 egonets;
df2 <- data.frame(
  ego=rep(1:length(alterL),each=20),alter=rep(1:20),cdn=NA, study=NA,bff=NA,csn=NA,
  cdn_embed.t2=NA, study_embed.t2=NA, bff_embed.t2=NA, csn_embed.t2=NA)

for (i in unique(df2$ego)) {  
      for (j in 1:20) { # for alters nested in ego
                 
        # find out if names denoting alter j appear in the 4 egonets
        alter <- unlist(alterL[[i]][j,c(2:5)], use.names = FALSE)
        alter <- alter[!is.na(alter)] # exclude NAs
        
        cdn <- alter %in% net1[i,]
        study <- alter %in% net2[i,]
        bff <- alter %in% net3[i,]
        csn <- alter %in% net4[i,]
        
        # and if so, give alter j score 1 on indicators; 0 otherwise
        df2$cdn[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% cdn, 1, 0)
        df2$study[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% study, 1, 0)
        df2$bff[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% bff, 1, 0)
        df2$csn[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% csn, 1, 0)
        
        #and attach embeddedness t2 scores
        df2$cdn_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$cdn_embed.t2[j]
        df2$study_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$study_embed.t2[j]
        df2$bff_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$bff_embed.t2[j]
        df2$csn_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$csn_embed.t2[j]
      }
}

# now that i have, for each alter of ego (including duplicates) at t2,
# the nets to which they belong..
# i continue with the w1-w2 matching matrices

# we already subsetted the matching matrices (in `w1w2`)
for (i in unique(df$ego)) {  # for ego i
  
  matchingL <- vector("list", 20) #pre-allocate empty list of length 15, to store matching matrices
  {
    matchingL[[1]] <- w1w2[i,1:20]
    matchingL[[2]] <- w1w2[i,21:40]
    matchingL[[3]] <- w1w2[i,41:60]
    matchingL[[4]] <- w1w2[i,61:80]
    matchingL[[5]] <- w1w2[i,81:100]
    matchingL[[6]] <- w1w2[i,101:120]
    matchingL[[7]] <- w1w2[i,121:140]
    matchingL[[8]] <- w1w2[i,141:160]
    matchingL[[9]] <- w1w2[i,161:180]
    matchingL[[10]] <- w1w2[i,181:200]
    matchingL[[11]] <- w1w2[i,201:220]
    matchingL[[12]] <- w1w2[i,221:240]
    matchingL[[13]] <- w1w2[i,241:260]
    matchingL[[14]] <- w1w2[i,261:280]
    matchingL[[15]] <- w1w2[i,281:300]
    matchingL[[16]] <- w1w2[i,301:320]
    matchingL[[17]] <- w1w2[i,321:340]
    matchingL[[18]] <- w1w2[i,341:360]
    matchingL[[19]] <- w1w2[i,361:380]
    matchingL[[20]] <- w1w2[i,381:400]
  }
  
  #find the 'right' matching matrix in this list, by taking the one that contains answers
  ind <- NULL
  for (j in seq_along(matchingL)) {
    #check along the sequence of elements j in the matching matrix list if they are non-empty and not NA
    check <- FALSE
    for (col in matchingL[[j]]) {
      if (!all(is.na(col)) && any(nchar(col) > 0)) {
        check <- TRUE
        break  
      }
      }
    if (check) {
      ind <- j
      break  
    }
  }
  
  if(length(ind)>0) {
    # get the  corresponding matrix
    mm <- matchingL[[ind]]
    
    # extract alter ids
    ans <- stringr::str_extract_all(mm,"\\(?[0-9,.]+\\)?") 
    # here, the object refers to the w1-alter j;
    # and the element indicator in the list refers to the w2-alter 
    
    for (j in unique(df$alterid[which(df$ego==i)])) { # for w1-alter j,
      
      # to which row/w2-alter was alter j matched?
      match <- which(ans==j)
      # and in which networks did this w2-alter belong?
      nets <- df2[which(df2$ego==i & df2$alter==match[1]),] #add [1] to match, for the few cases where multiple w2-alters were matched to the same w1-alter (i.e., when ego did not correclty perform match w2 alters to each other.)
    
    if(length(match)>0) { # if j was matched to w2-alters...
      # assign to alter j the networks in which j reappeared
      df$cdn2[which(df$ego==i & df$alterid==j)] <- nets$cdn[1]
      df$study2[which(df$ego==i & df$alterid==j)] <- nets$study[1]
      df$bff2[which(df$ego==i & df$alterid==j)] <- nets$bff[1]
      df$csn2[which(df$ego==i & df$alterid==j)] <- nets$csn[1]
      
      #and attach their t2 embeddednes score
      df$cdn_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$cdn_embed.t2[1]
      df$study_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$study_embed.t2[1]
      df$bff_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$bff_embed.t2[1]
      df$csn_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$csn_embed.t2[1]
    }
  }
  }
}

# currently; a lot of NAs in the attributes indicating whether alter j was member of the respective networks.
# those are alters that were not renamed, and thus have no 'chance' to re-appear. so, 0s are alters that were once part of a specific network, but re-appeared not in that particular network.
# naturally, for analyses, NAs can be set to 0.

# as a second tie change indicator, i look at whether (un-listed) alters and ego were still in touch
df$frequency.t2 <- NA
df$closeness.t2 <- NA

# subset w2 name interpreters on contact freq. of w1 alters (sq021 - sq040!)
freq <- data4[,c(1176:1195)]
# also closeness
close <- data4[,c(1216:1235)]

# recode into numeric values;
freq <- ifelse(freq=="(Bijna) elke dag",7,
                       ifelse(freq=="1-2 keer per week",6,
                              ifelse(freq=="Aantal keer per maand",5, 
                                     ifelse(freq=="Ong. 1 keer per maand",4, 
                                            ifelse(freq=="Aantal keer per jaar",3,
                                                   ifelse(freq=="Ong. 1 keer per jaar",2,
                                                          ifelse(freq=="Nooit",1,
                                                                      NA   )))))))
close <- ifelse(close=="Heel erg hecht", 4, ifelse(close=="Hecht", 3, ifelse(close=="Enigszins hecht",2, ifelse(close=="Niet hecht", 1, NA))))

for (i in unique(df$ego)) {
  for (j in unique(df$alterid[which(df$ego==i)])) {
    df$frequency.t2[which(df$ego==i & df$alterid==j)] <- freq[i,][!is.na(freq[i,])][j]
    df$closeness.t2[which(df$ego==i & df$alterid==j)] <- close[i,][!is.na(close[i,])][j]
  }
}

# size of each egonet at t2
df$cdn.size2 <- NA
df$study.size2 <- NA
df$bff.size2 <- NA
df$csn.size2 <- NA

for (i in unique(df$ego)) {
  df$cdn.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet1.SQ001.[i],data4$egonet1.SQ002.[i],data4$egonet1.SQ003.[i],data4$egonet1.SQ004.[i],data4$egonet1.SQ005.[i])!="")
    df$study.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet2.SQ001.[i],data4$egonet2.SQ002.[i],data4$egonet2.SQ003.[i],data4$egonet2.SQ004.[i],data4$egonet2.SQ005.[i])!="")
    df$bff.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet3.SQ001.[i],data4$egonet3.SQ002.[i],data4$egonet3.SQ003.[i],data4$egonet3.SQ004.[i],data4$egonet3.SQ005.[i])!="")
    df$csn.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet4.SQ001.[i],data4$egonet4.SQ002.[i],data4$egonet4.SQ003.[i],data4$egonet4.SQ004.[i],data4$egonet4.SQ005.[i])!="") }

#and density at t2.
df$cdn.density2 <- df$study.density2 <- df$csn.density2 <- df$bff.density2 <- NA

for (i in unique(df$ego)) {
  #a. CDN
  size <- df$cdn.size2[which(df$ego==i)][1]
  #no. possible ties between alters
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  #get no. of observed ties
  #only if ego named at least 2 alters in this name generator
  #because only then i can calculate density
  if(size>1) {
    #get alter-alter adjacency matrices of this particular egonet
    adjacency <- data4[i,c(8:27)]

    #i want the correct ones, given the no. of listed alters in this name generator
    #i list the columns numbers of the variables corresponding to each matching matrix set,
    #and subset corresponding columns in `adjacency`
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
   
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$cdn.density2[which(df$ego==i)] <- obs/pos  }
  
  #b. study
  size <- df$study.size2[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data4[i,c(33:52)]
    #reorder columns.... messy!
    adjacency <- adjacency[,c("adj2N2a.SQ001.", "adj2N3a.SQ001.", "adj2N3a.SQ002.", "adj2N3b.SQ001.", "adj2N4a.SQ001.", "adj2N4a.SQ002.",
                       "adj2N4a.SQ003.", "adj2N4b.SQ001.", "adj2N4b.SQ002.", "adj2N4c.SQ001.", "adj2N5a.SQ001.", "adj2N5a.SQ002.",
                       "adj2N5a.SQ003.", "adj2N5a.SQ004.", "adj2N5b.SQ001.", "adj2N5b.SQ002.", "adj2N5b.SQ003.", "adj2N5c.SQ001.",
                       "adj2N5c.SQ002.", "adj2N5d.SQ001.")]
    
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
  
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$study.density2[which(df$ego==i)] <- obs/pos }
  
  #c. friends
  size <- df$bff.size2[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data4[i,c(156:175)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$bff.density2[which(df$ego==i)] <- obs/pos }

  #d. sports partners
  size <- df$csn.size2[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data4[i,c(401:420)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
  
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$csn.density2[which(df$ego==i)] <- obs/pos }
}

#set NAs to 0.
df$cdn.density2[is.na(df$cdn.density2)] <- 0
df$study.density2[is.na(df$study.density2)] <- 0
df$bff.density2[is.na(df$bff.density2)] <- 0
df$csn.density2[is.na(df$csn.density2)] <- 0

#add Wave 2 status
df$statusW2 <- NA
df$statusW2 <- ifelse(df$survive==1, "Maintained", "Dropped")

## methodological controls

#include whether max number of names was reached at t2
df$cdn.max2 <- ifelse(df$cdn.size2==5,1,0)
df$study.max2 <- ifelse(df$study.size2==5,1,0)
df$bff.max2 <- ifelse(df$bff.size2==5,1,0)
df$csn.max2 <- ifelse(df$csn.size2==5,1,0)

#survey time 
df$survey2time <- NA

#time of day
df$beforenine2 <- NA
df$ninetofive2 <- NA
df$afterfive2 <- NA

#filled out on the weekend (sa/su)?
df$weekend2 <- NA

#time per name generator
df$cdn.time2 <- NA
df$study.time2 <- NA
df$bff.time2 <- NA
df$csn.time2 <- NA

#time till name generator
df$cdn.timetill2 <- NA
df$study.timetill2 <- NA
df$bff.timetill2 <- NA
df$csn.timetill2 <- NA

for (i in unique(df$ego)) {
  
  #survey time at wave 2
  #convert start and submitdate to POSIXct and calculate difference
  #in minutes; extract number.
  df$survey2time[which(df$ego==i)] <- stringr::str_extract_all(trunc(as.POSIXct(data4$submitdate[i]) - as.POSIXct(data4$startdate[i]), "mins"),"\\(?[0-9,.]+\\)?")[[1]]
  
  #time of day of survey.
  #table(substr(data.table::as.ITime(as.POSIXct(data4$startdate)),1,2)) #distribution
  df$beforenine2[which(df$ego==i)] <- ifelse ( data.table::as.ITime(as.POSIXct(data4$startdate[i])) < data.table::as.ITime("9:00:00"), 1, 0)
  df$ninetofive2[which(df$ego==i)] <- ifelse ( data.table::as.ITime(as.POSIXct(data4$startdate[i])) > data.table::as.ITime("9:00:00") &
                                                 data.table::as.ITime(as.POSIXct(data4$startdate[i])) < data.table::as.ITime("17:00:00"), 1, 0)
  df$afterfive2[which(df$ego==i)] <- ifelse ( data.table::as.ITime(as.POSIXct(data4$startdate[i])) > data.table::as.ITime("17:00:00"), 1, 0)
  
  #time per name generator question, in minutes
 
  df$cdn.time2[which(df$ego==i)] <- data4$groupTime718[i]/60
  df$study.time2[which(df$ego==i)] <- data4$groupTime722[i]/60
  df$bff.time2[which(df$ego==i)] <- data4$groupTime730[i]/60
  df$csn.time2[which(df$ego==i)] <- data4$groupTime737[i]/60
  
  #also get weekday/weekend
  #counting friday as weekend, after 5pm
  
  #table(wday(ymd_hms(as.POSIXct(data4$submitdate)),label = TRUE)) #distribution
  #extract weekday
  weekday <- wday(ymd_hms(as.POSIXct(data4$submitdate[i])),label = TRUE)
  #and time
  time <- hour(as.POSIXct(data4$submitdate[i]))
  df$weekend2[which(df$ego==i)] <- ifelse( (weekday %in% c("za", "zo")) | (weekday == "vr" & time >= 17), 1, 0)
  
  #time till name generator question, in minutes
  # wave 2: question groups 717 CDN718 720 STUDY722 723 724 725 748 793 726 749 729 740 BFF730 731 732 734 735 736 743 CSN737
  df$cdn.timetill2[which(df$ego==i)] <- sum(data4[,c("groupTime717")][i])/60
  df$study.timetill2[which(df$ego==i)] <- rowSums(data4[,c("groupTime717", "groupTime718", "groupTime720")][i,], na.rm = TRUE)/60
  df$bff.timetill2[which(df$ego==i)] <- rowSums(data4[,c("groupTime717", "groupTime718", "groupTime720", "groupTime722", "groupTime723", "groupTime724", "groupTime725", "groupTime748", "groupTime793", "groupTime726", "groupTime749","groupTime729", "groupTime740")][i,], na.rm = TRUE)/60
  df$csn.timetill2[which(df$ego==i)] <- rowSums(data4[,c("groupTime717", "groupTime718", "groupTime720", "groupTime722", "groupTime723", "groupTime724", "groupTime725", "groupTime748", "groupTime793", "groupTime726", "groupTime749","groupTime729", "groupTime740","groupTime730", "groupTime731", "groupTime732", "groupTime734", "groupTime735", "groupTime736", "groupTime743")][i,], na.rm = TRUE)/60
  
}

df_maintained <- df
#fix(df_maintained)
#str(df_maintained)
#last make gender numeric..
df_maintained$alter_gender <- as.numeric(df_maintained$alter_gender)

wave 2 created

#create a new alter List
alterL <- vector("list", nrow(data))

# loop over all egos
for ( i in 1:length(alterL)) {
    alterL[[i]] <- data.frame(
      ego_gender = NA, ego_age = NA, ego_educ = NA,
      alterid = 1:20, name1 = NA, name2 = NA, name3 = NA, name4 = NA,
      alter_gender = NA, alter_age = NA, alter_educ=NA,
      same_gender = NA, dif_age = NA, sim_educ = NA,
      cdn_embed.t2 = NA, study_embed.t2 = NA, bff_embed.t2 = NA, csn_embed.t2 = NA)
}

#fill in names based on names_df
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=FALSE)
}

# replace empty strings with <NA>
for (i in 1:length(alterL)) {
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
}

#matching
{
  net1 <- cbind(data4$egonet1.SQ001.,data4$egonet1.SQ002., data4$egonet1.SQ003.,data4$egonet1.SQ004., data4$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data4$egonet2.SQ001.,data4$egonet2.SQ002., data4$egonet2.SQ003.,data4$egonet2.SQ004., data4$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data4$egonet3.SQ001.,data4$egonet3.SQ002., data4$egonet3.SQ003.,data4$egonet3.SQ004., data4$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data4$egonet4.SQ001.,data4$egonet4.SQ002., data4$egonet4.SQ003.,data4$egonet4.SQ004., data4$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}

matchingList <- list()
for (i in 1:length(alterL)) { # for ego i
  matchingL <- list()
  # make 5 seperate matching matrices. naturally, only 1 is relevant... but i will select that later.
  matchingL[[1]] <- cbind(data4$matching1N1.SQ001_SQ001.[i],data4$matching1N1.SQ002_SQ001.[i],data4$matching1N1.SQ003_SQ001.[i],data4$matching1N1.SQ004_SQ001.[i],data4$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data4$matching1N2.SQ001_SQ001.[i], data4$matching1N2.SQ002_SQ001.[i], data4$matching1N2.SQ003_SQ001.[i], data4$matching1N2.SQ004_SQ001.[i], data4$matching1N2.SQ005_SQ001.[i]),
    cbind(data4$matching1N2.SQ001_SQ002.[i], data4$matching1N2.SQ002_SQ002.[i], data4$matching1N2.SQ003_SQ002.[i], data4$matching1N2.SQ004_SQ002.[i], data4$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data4$matching1N3.SQ001_SQ001.[i], data4$matching1N3.SQ002_SQ001.[i], data4$matching1N3.SQ003_SQ001.[i], data4$matching1N3.SQ004_SQ001.[i], data4$matching1N3.SQ005_SQ001.[i]),
    cbind(data4$matching1N3.SQ001_SQ002.[i], data4$matching1N3.SQ002_SQ002.[i], data4$matching1N3.SQ003_SQ002.[i], data4$matching1N3.SQ004_SQ002.[i], data4$matching1N3.SQ005_SQ002.[i]),
    cbind(data4$matching1N3.SQ001_SQ003.[i], data4$matching1N3.SQ002_SQ003.[i], data4$matching1N3.SQ003_SQ003.[i], data4$matching1N3.SQ004_SQ003.[i], data4$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data4$matching1N4.SQ001_SQ001.[i], data4$matching1N4.SQ002_SQ001.[i], data4$matching1N4.SQ003_SQ001.[i], data4$matching1N4.SQ004_SQ001.[i], data4$matching1N4.SQ005_SQ001.[i]),
    cbind(data4$matching1N4.SQ001_SQ002.[i], data4$matching1N4.SQ002_SQ002.[i], data4$matching1N4.SQ003_SQ002.[i], data4$matching1N4.SQ004_SQ002.[i], data4$matching1N4.SQ005_SQ002.[i]),
    cbind(data4$matching1N4.SQ001_SQ003.[i], data4$matching1N4.SQ002_SQ003.[i], data4$matching1N4.SQ003_SQ003.[i], data4$matching1N4.SQ004_SQ003.[i], data4$matching1N4.SQ005_SQ003.[i]),
    cbind(data4$matching1N4.SQ001_SQ004.[i], data4$matching1N4.SQ002_SQ004.[i], data4$matching1N4.SQ003_SQ004.[i], data4$matching1N4.SQ004_SQ004.[i], data4$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data4$matching1N5.SQ001_SQ001.[i], data4$matching1N5.SQ002_SQ001.[i], data4$matching1N5.SQ003_SQ001.[i], data4$matching1N5.SQ004_SQ001.[i], data4$matching1N5.SQ005_SQ001.[i]),
    cbind(data4$matching1N5.SQ001_SQ002.[i], data4$matching1N5.SQ002_SQ002.[i], data4$matching1N5.SQ003_SQ002.[i], data4$matching1N5.SQ004_SQ002.[i], data4$matching1N5.SQ005_SQ002.[i]),
    cbind(data4$matching1N5.SQ001_SQ003.[i], data4$matching1N5.SQ002_SQ003.[i], data4$matching1N5.SQ003_SQ003.[i], data4$matching1N5.SQ004_SQ003.[i], data4$matching1N5.SQ005_SQ003.[i]),
    cbind(data4$matching1N5.SQ001_SQ004.[i], data4$matching1N5.SQ002_SQ004.[i], data4$matching1N5.SQ003_SQ004.[i], data4$matching1N5.SQ004_SQ004.[i], data4$matching1N5.SQ005_SQ004.[i]),
    cbind(data4$matching1N5.SQ001_SQ005.[i], data4$matching1N5.SQ002_SQ005.[i], data4$matching1N5.SQ003_SQ005.[i], data4$matching1N5.SQ004_SQ005.[i], data4$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
}

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
}

matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data4$matching2N1.SQ001_SQ001.[i],data4$matching2N1.SQ002_SQ001.[i],data4$matching2N1.SQ003_SQ001.[i],data4$matching2N1.SQ004_SQ001.[i],data4$matching2N1.SQ005_SQ001.[i],data4$matching2N1.SQ006_SQ001.[i],data4$matching2N1.SQ007_SQ001.[i],data4$matching2N1.SQ008_SQ001.[i],data4$matching2N1.SQ009_SQ001.[i],data4$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data4$matching2N2.SQ001_SQ001.[i],data4$matching2N2.SQ002_SQ001.[i],data4$matching2N2.SQ003_SQ001.[i],data4$matching2N2.SQ004_SQ001.[i],data4$matching2N2.SQ005_SQ001.[i],data4$matching2N2.SQ006_SQ001.[i],data4$matching2N2.SQ007_SQ001.[i],data4$matching2N2.SQ008_SQ001.[i],data4$matching2N2.SQ009_SQ001.[i],data4$matching2N2.SQ010_SQ001.[i]),
    cbind(data4$matching2N2.SQ001_SQ002.[i],data4$matching2N2.SQ002_SQ002.[i],data4$matching2N2.SQ003_SQ002.[i],data4$matching2N2.SQ004_SQ002.[i],data4$matching2N2.SQ005_SQ002.[i],data4$matching2N2.SQ006_SQ002.[i],data4$matching2N2.SQ007_SQ002.[i],data4$matching2N2.SQ008_SQ002.[i],data4$matching2N2.SQ009_SQ002.[i],data4$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data4$matching2N3.SQ001_SQ001.[i],data4$matching2N3.SQ002_SQ001.[i],data4$matching2N3.SQ003_SQ001.[i],data4$matching2N3.SQ004_SQ001.[i],data4$matching2N3.SQ005_SQ001.[i],data4$matching2N3.SQ006_SQ001.[i],data4$matching2N3.SQ007_SQ001.[i],data4$matching2N3.SQ008_SQ001.[i],data4$matching2N3.SQ009_SQ001.[i],data4$matching2N3.SQ010_SQ001.[i]),
    cbind(data4$matching2N3.SQ001_SQ002.[i],data4$matching2N3.SQ002_SQ002.[i],data4$matching2N3.SQ003_SQ002.[i],data4$matching2N3.SQ004_SQ002.[i],data4$matching2N3.SQ005_SQ002.[i],data4$matching2N3.SQ006_SQ002.[i],data4$matching2N3.SQ007_SQ002.[i],data4$matching2N3.SQ008_SQ002.[i],data4$matching2N3.SQ009_SQ002.[i],data4$matching2N3.SQ010_SQ002.[i]),
    cbind(data4$matching2N3.SQ001_SQ003.[i],data4$matching2N3.SQ002_SQ003.[i],data4$matching2N3.SQ003_SQ003.[i],data4$matching2N3.SQ004_SQ003.[i],data4$matching2N3.SQ005_SQ003.[i],data4$matching2N3.SQ006_SQ003.[i],data4$matching2N3.SQ007_SQ003.[i],data4$matching2N3.SQ008_SQ003.[i],data4$matching2N3.SQ009_SQ003.[i],data4$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data4$matching2N4.SQ001_SQ001.[i],data4$matching2N4.SQ002_SQ001.[i],data4$matching2N4.SQ003_SQ001.[i],data4$matching2N4.SQ004_SQ001.[i],data4$matching2N4.SQ005_SQ001.[i],data4$matching2N4.SQ006_SQ001.[i],data4$matching2N4.SQ007_SQ001.[i],data4$matching2N4.SQ008_SQ001.[i],data4$matching2N4.SQ009_SQ001.[i],data4$matching2N4.SQ010_SQ001.[i]),
    cbind(data4$matching2N4.SQ001_SQ002.[i],data4$matching2N4.SQ002_SQ002.[i],data4$matching2N4.SQ003_SQ002.[i],data4$matching2N4.SQ004_SQ002.[i],data4$matching2N4.SQ005_SQ002.[i],data4$matching2N4.SQ006_SQ002.[i],data4$matching2N4.SQ007_SQ002.[i],data4$matching2N4.SQ008_SQ002.[i],data4$matching2N4.SQ009_SQ002.[i],data4$matching2N4.SQ010_SQ002.[i]),
    cbind(data4$matching2N4.SQ001_SQ003.[i],data4$matching2N4.SQ002_SQ003.[i],data4$matching2N4.SQ003_SQ003.[i],data4$matching2N4.SQ004_SQ003.[i],data4$matching2N4.SQ005_SQ003.[i],data4$matching2N4.SQ006_SQ003.[i],data4$matching2N4.SQ007_SQ003.[i],data4$matching2N4.SQ008_SQ003.[i],data4$matching2N4.SQ009_SQ003.[i],data4$matching2N4.SQ010_SQ003.[i]),
    cbind(data4$matching2N4.SQ001_SQ004.[i],data4$matching2N4.SQ002_SQ004.[i],data4$matching2N4.SQ003_SQ004.[i],data4$matching2N4.SQ004_SQ004.[i],data4$matching2N4.SQ005_SQ004.[i],data4$matching2N4.SQ006_SQ004.[i],data4$matching2N4.SQ007_SQ004.[i],data4$matching2N4.SQ008_SQ004.[i],data4$matching2N4.SQ009_SQ004.[i],data4$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data4$matching2N5.SQ001_SQ001.[i],data4$matching2N5.SQ002_SQ001.[i],data4$matching2N5.SQ003_SQ001.[i],data4$matching2N5.SQ004_SQ001.[i],data4$matching2N5.SQ005_SQ001.[i],data4$matching2N5.SQ006_SQ001.[i],data4$matching2N5.SQ007_SQ001.[i],data4$matching2N5.SQ008_SQ001.[i],data4$matching2N5.SQ009_SQ001.[i],data4$matching2N5.SQ010_SQ001.[i]),
    cbind(data4$matching2N5.SQ001_SQ002.[i],data4$matching2N5.SQ002_SQ002.[i],data4$matching2N5.SQ003_SQ002.[i],data4$matching2N5.SQ004_SQ002.[i],data4$matching2N5.SQ005_SQ002.[i],data4$matching2N5.SQ006_SQ002.[i],data4$matching2N5.SQ007_SQ002.[i],data4$matching2N5.SQ008_SQ002.[i],data4$matching2N5.SQ009_SQ002.[i],data4$matching2N5.SQ010_SQ002.[i]),
    cbind(data4$matching2N5.SQ001_SQ003.[i],data4$matching2N5.SQ002_SQ003.[i],data4$matching2N5.SQ003_SQ003.[i],data4$matching2N5.SQ004_SQ003.[i],data4$matching2N5.SQ005_SQ003.[i],data4$matching2N5.SQ006_SQ003.[i],data4$matching2N5.SQ007_SQ003.[i],data4$matching2N5.SQ008_SQ003.[i],data4$matching2N5.SQ009_SQ003.[i],data4$matching2N5.SQ010_SQ003.[i]),
    cbind(data4$matching2N5.SQ001_SQ004.[i],data4$matching2N5.SQ002_SQ004.[i],data4$matching2N5.SQ003_SQ004.[i],data4$matching2N5.SQ004_SQ004.[i],data4$matching2N5.SQ005_SQ004.[i],data4$matching2N5.SQ006_SQ004.[i],data4$matching2N5.SQ007_SQ004.[i],data4$matching2N5.SQ008_SQ004.[i],data4$matching2N5.SQ009_SQ004.[i],data4$matching2N5.SQ010_SQ004.[i]),
    cbind(data4$matching2N5.SQ001_SQ005.[i],data4$matching2N5.SQ002_SQ005.[i],data4$matching2N5.SQ003_SQ005.[i],data4$matching2N5.SQ004_SQ005.[i],data4$matching2N5.SQ005_SQ005.[i],data4$matching2N5.SQ006_SQ005.[i],data4$matching2N5.SQ007_SQ005.[i],data4$matching2N5.SQ008_SQ005.[i],data4$matching2N5.SQ009_SQ005.[i],data4$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data4$matching3N1.SQ001_SQ001.[i],data4$matching3N1.SQ002_SQ001.[i],data4$matching3N1.SQ003_SQ001.[i],data4$matching3N1.SQ004_SQ001.[i],data4$matching3N1.SQ005_SQ001.[i],data4$matching3N1.SQ006_SQ001.[i],data4$matching3N1.SQ007_SQ001.[i],data4$matching3N1.SQ008_SQ001.[i],data4$matching3N1.SQ009_SQ001.[i],data4$matching3N1.SQ010_SQ001.[i], data4$matching3N1.SQ011_SQ001.[i], data4$matching3N1.SQ012_SQ001.[i], data4$matching3N1.SQ013_SQ001.[i], data4$matching3N1.SQ014_SQ001.[i], data4$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data4$matching3N2.SQ001_SQ001.[i],data4$matching3N2.SQ002_SQ001.[i],data4$matching3N2.SQ003_SQ001.[i],data4$matching3N2.SQ004_SQ001.[i],data4$matching3N2.SQ005_SQ001.[i],data4$matching3N2.SQ006_SQ001.[i],data4$matching3N2.SQ007_SQ001.[i],data4$matching3N2.SQ008_SQ001.[i],data4$matching3N2.SQ009_SQ001.[i],data4$matching3N2.SQ010_SQ001.[i], data4$matching3N2.SQ011_SQ001.[i], data4$matching3N2.SQ012_SQ001.[i], data4$matching3N2.SQ013_SQ001.[i], data4$matching3N2.SQ014_SQ001.[i], data4$matching3N2.SQ015_SQ001.[i]),
    cbind(data4$matching3N2.SQ001_SQ002.[i],data4$matching3N2.SQ002_SQ002.[i],data4$matching3N2.SQ003_SQ002.[i],data4$matching3N2.SQ004_SQ002.[i],data4$matching3N2.SQ005_SQ002.[i],data4$matching3N2.SQ006_SQ002.[i],data4$matching3N2.SQ007_SQ002.[i],data4$matching3N2.SQ008_SQ002.[i],data4$matching3N2.SQ009_SQ002.[i],data4$matching3N2.SQ010_SQ002.[i], data4$matching3N2.SQ011_SQ002.[i], data4$matching3N2.SQ012_SQ002.[i], data4$matching3N2.SQ013_SQ002.[i], data4$matching3N2.SQ014_SQ002.[i], data4$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data4$matching3N3.SQ001_SQ001.[i],data4$matching3N3.SQ002_SQ001.[i],data4$matching3N3.SQ003_SQ001.[i],data4$matching3N3.SQ004_SQ001.[i],data4$matching3N3.SQ005_SQ001.[i],data4$matching3N3.SQ006_SQ001.[i],data4$matching3N3.SQ007_SQ001.[i],data4$matching3N3.SQ008_SQ001.[i],data4$matching3N3.SQ009_SQ001.[i],data4$matching3N3.SQ010_SQ001.[i], data4$matching3N3.SQ011_SQ001.[i], data4$matching3N3.SQ012_SQ001.[i], data4$matching3N3.SQ013_SQ001.[i], data4$matching3N3.SQ014_SQ001.[i], data4$matching3N3.SQ015_SQ001.[i]),
    cbind(data4$matching3N3.SQ001_SQ002.[i],data4$matching3N3.SQ002_SQ002.[i],data4$matching3N3.SQ003_SQ002.[i],data4$matching3N3.SQ004_SQ002.[i],data4$matching3N3.SQ005_SQ002.[i],data4$matching3N3.SQ006_SQ002.[i],data4$matching3N3.SQ007_SQ002.[i],data4$matching3N3.SQ008_SQ002.[i],data4$matching3N3.SQ009_SQ002.[i],data4$matching3N3.SQ010_SQ002.[i], data4$matching3N3.SQ011_SQ002.[i], data4$matching3N3.SQ012_SQ002.[i], data4$matching3N3.SQ013_SQ002.[i], data4$matching3N3.SQ014_SQ002.[i], data4$matching3N3.SQ015_SQ002.[i]),
    cbind(data4$matching3N3.SQ001_SQ003.[i],data4$matching3N3.SQ002_SQ003.[i],data4$matching3N3.SQ003_SQ003.[i],data4$matching3N3.SQ004_SQ003.[i],data4$matching3N3.SQ005_SQ003.[i],data4$matching3N3.SQ006_SQ003.[i],data4$matching3N3.SQ007_SQ003.[i],data4$matching3N3.SQ008_SQ003.[i],data4$matching3N3.SQ009_SQ003.[i],data4$matching3N3.SQ010_SQ003.[i], data4$matching3N3.SQ011_SQ003.[i], data4$matching3N3.SQ012_SQ003.[i], data4$matching3N3.SQ013_SQ003.[i], data4$matching3N3.SQ014_SQ003.[i], data4$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data4$matching3N4.SQ001_SQ001.[i],data4$matching3N4.SQ002_SQ001.[i],data4$matching3N4.SQ003_SQ001.[i],data4$matching3N4.SQ004_SQ001.[i],data4$matching3N4.SQ005_SQ001.[i],data4$matching3N4.SQ006_SQ001.[i],data4$matching3N4.SQ007_SQ001.[i],data4$matching3N4.SQ008_SQ001.[i],data4$matching3N4.SQ009_SQ001.[i],data4$matching3N4.SQ010_SQ001.[i], data4$matching3N4.SQ011_SQ001.[i], data4$matching3N4.SQ012_SQ001.[i], data4$matching3N4.SQ013_SQ001.[i], data4$matching3N4.SQ014_SQ001.[i], data4$matching3N4.SQ015_SQ001.[i]),
    cbind(data4$matching3N4.SQ001_SQ002.[i],data4$matching3N4.SQ002_SQ002.[i],data4$matching3N4.SQ003_SQ002.[i],data4$matching3N4.SQ004_SQ002.[i],data4$matching3N4.SQ005_SQ002.[i],data4$matching3N4.SQ006_SQ002.[i],data4$matching3N4.SQ007_SQ002.[i],data4$matching3N4.SQ008_SQ002.[i],data4$matching3N4.SQ009_SQ002.[i],data4$matching3N4.SQ010_SQ002.[i], data4$matching3N4.SQ011_SQ002.[i], data4$matching3N4.SQ012_SQ002.[i], data4$matching3N4.SQ013_SQ002.[i], data4$matching3N4.SQ014_SQ002.[i], data4$matching3N4.SQ015_SQ002.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ004.[i],data4$matching3N4.SQ002_SQ004.[i],data4$matching3N4.SQ003_SQ004.[i],data4$matching3N4.SQ004_SQ004.[i],data4$matching3N4.SQ005_SQ004.[i],data4$matching3N4.SQ006_SQ004.[i],data4$matching3N4.SQ007_SQ004.[i],data4$matching3N4.SQ008_SQ004.[i],data4$matching3N4.SQ009_SQ004.[i],data4$matching3N4.SQ010_SQ004.[i], data4$matching3N4.SQ011_SQ004.[i], data4$matching3N4.SQ012_SQ004.[i], data4$matching3N4.SQ013_SQ004.[i], data4$matching3N4.SQ014_SQ004.[i], data4$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data4$matching3N5.SQ001_SQ001.[i],data4$matching3N5.SQ002_SQ001.[i],data4$matching3N5.SQ003_SQ001.[i],data4$matching3N5.SQ004_SQ001.[i],data4$matching3N5.SQ005_SQ001.[i],data4$matching3N5.SQ006_SQ001.[i],data4$matching3N5.SQ007_SQ001.[i],data4$matching3N5.SQ008_SQ001.[i],data4$matching3N5.SQ009_SQ001.[i],data4$matching3N5.SQ010_SQ001.[i], data4$matching3N5.SQ011_SQ001.[i], data4$matching3N5.SQ012_SQ001.[i], data4$matching3N5.SQ013_SQ001.[i], data4$matching3N5.SQ014_SQ001.[i], data4$matching3N5.SQ015_SQ001.[i]),
    cbind(data4$matching3N5.SQ001_SQ002.[i],data4$matching3N5.SQ002_SQ002.[i],data4$matching3N5.SQ003_SQ002.[i],data4$matching3N5.SQ004_SQ002.[i],data4$matching3N5.SQ005_SQ002.[i],data4$matching3N5.SQ006_SQ002.[i],data4$matching3N5.SQ007_SQ002.[i],data4$matching3N5.SQ008_SQ002.[i],data4$matching3N5.SQ009_SQ002.[i],data4$matching3N5.SQ010_SQ002.[i], data4$matching3N5.SQ011_SQ002.[i], data4$matching3N5.SQ012_SQ002.[i], data4$matching3N5.SQ013_SQ002.[i], data4$matching3N5.SQ014_SQ002.[i], data4$matching3N5.SQ015_SQ002.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ005.[i],data4$matching3N5.SQ002_SQ005.[i],data4$matching3N5.SQ003_SQ005.[i],data4$matching3N5.SQ004_SQ005.[i],data4$matching3N5.SQ005_SQ005.[i],data4$matching3N5.SQ006_SQ005.[i],data4$matching3N5.SQ007_SQ005.[i],data4$matching3N5.SQ008_SQ005.[i],data4$matching3N5.SQ009_SQ005.[i],data4$matching3N5.SQ010_SQ005.[i], data4$matching3N5.SQ011_SQ005.[i], data4$matching3N5.SQ012_SQ005.[i], data4$matching3N5.SQ013_SQ005.[i], data4$matching3N5.SQ014_SQ005.[i], data4$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

#i also calculate structural embeddedness (alter-level) per ego-net.
for (i in 1:length(alterL)) { # for ego i

  #get number of names listed in the egonets
  nnames_cdn <- length(which(data4[i,c("egonet1.SQ001.","egonet1.SQ002.","egonet1.SQ003.","egonet1.SQ004.","egonet1.SQ005.")]!=""))
  nnames_study <- length(which(data4[i,c("egonet2.SQ001.","egonet2.SQ002.","egonet2.SQ003.","egonet2.SQ004.","egonet2.SQ005.")]!=""))
  nnames_bff <- length(which(data4[i,c("egonet3.SQ001.","egonet3.SQ002.","egonet3.SQ003.","egonet3.SQ004.","egonet3.SQ005.")]!=""))
  nnames_csn <- length(which(data4[i,c("egonet4.SQ001.","egonet4.SQ002.","egonet4.SQ003.","egonet4.SQ004.","egonet4.SQ005.")]!=""))
  
  #CDN
  
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj1N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj1N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj1N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj1N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj1N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj1N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj1N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj1N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj1N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj1N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj1N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj1N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj1N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj1N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj1N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj1N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj1N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj1N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj1N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj1N5d.SQ001.[i]
  }

  if(nnames_cdn>1) { #we only know alter-alter relations for nnames>1
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_cdn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
  
  #CDN: 1:nnames_CDN
  alterL[[i]]$cdn_embed.t2[1:nnames_cdn] <- embed 
  }
  
  #STUDY

  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj2N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj2N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj2N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj2N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj2N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj2N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj2N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj2N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj2N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj2N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj2N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj2N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj2N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj2N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj2N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj2N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj2N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj2N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj2N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj2N5d.SQ001.[i]
  }

  if(nnames_study>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_study]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$study_embed.t2[6:(5+nnames_study)] <- embed 
  }
  
  # BEST FRIENDS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj3N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj3N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj3N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj3N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj3N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj3N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj3N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj3N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj3N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj3N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj3N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj3N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj3N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj3N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj3N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj3N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj3N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj3N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj3N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj3N5d.SQ001.[i]
  }

  if(nnames_bff>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_bff]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$bff_embed.t2[11:(10+nnames_bff)] <- embed 
  }

  # SPORTS PARTNERS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj4N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj4N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj4N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj4N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj4N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj4N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj4N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj4N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj4N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj4N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj4N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj4N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj4N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj4N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj4N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj4N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj4N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj4N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj4N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj4N5d.SQ001.[i]
  }

  if(nnames_csn>1) { #we only know alter-alter relations for nnames>2
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_csn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$csn_embed.t2[16:(15+nnames_csn)] <- embed 
  }
}

#alters could be named in multiple name generators. each element of alterL contains rows corresponding to alters that may be 'duplicates'.
#i match the structural embeddedness measures

for (i in 1:length(alterL)) { #for ego i 
  for (j in 1:nrow(alterL[[i]])) { #for alter j
 
    #if name2 is not empty, alter j was matched to the study network; and more precisely, to the study partner with name1: alterL[[i]]$name2[j]
    alterL[[i]]$study_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name2[j]), 
                                            alterL[[i]]$study_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name2[j] & !is.na(alterL[[i]]$study_embed.t2))], 
                                            alterL[[i]]$study_embed.t2[j])
    
    #if name3 is not empty, alter j was matched to the friends network; and more precisely, to the friend with name1: alterL[[i]]$name3[j]
    alterL[[i]]$bff_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name3[j]), 
                                          alterL[[i]]$bff_embed.t2[which(alterL[[i]]$name1 == alterL[[i]]$name3[j] & !is.na(alterL[[i]]$bff_embed.t2))], 
                                          alterL[[i]]$bff_embed.t2[j])
    
    #if name4 is not empty, alter j was matched to the sports network; and more precisely, to the sports partner with name1: alterL[[i]]$name4[j]
    alterL[[i]]$csn_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name4[j]), 
                                          alterL[[i]]$csn_embed.t2[which(alterL[[i]]$name1 == alterL[[i]]$name4[j] & !is.na(alterL[[i]]$csn_embed.t2))],
                                          alterL[[i]]$csn_embed.t2[j] )

  }
}

#while in wave 1, only the unique alters were listed in the gender-name interpreters
#in wave 2, only the NEW alters were listed (since it is a stable characteristic)
#so only the non-matched alters that did not appear in w1
#i asked respondents to 'check' males.

#gender
df_males <- data.frame(p1 = data4$males.SQ001.,p2 = data4$males.SQ002.,p3 = data4$males.SQ003.,p4 = data4$males.SQ004.,p5 = data4$males.SQ005.,p6 = data4$males.SQ006.,p7 = data4$males.SQ007.,p8 = data4$males.SQ008.,p9 = data4$males.SQ009.,p10= data4$males.SQ010.,p11= data4$males.SQ011.,p12= data4$males.SQ012.,p13= data4$males.SQ013.,p14= data4$males.SQ014.,p15= data4$males.SQ015.,p16= data4$males.SQ016.,p17= data4$males.SQ017.,p18= data4$males.SQ018.,p19= data4$males.SQ019.,p20= data4$males.SQ020.)

#kin 
df_kin <- data.frame(p1 = data4$kin.SQ001.,p2 = data4$kin.SQ002.,p3 = data4$kin.SQ003.,p4 = data4$kin.SQ004.,p5 = data4$kin.SQ005.,p6 = data4$kin.SQ006.,p7 = data4$kin.SQ007.,p8 = data4$kin.SQ008.,p9 = data4$kin.SQ009., p10= data4$kin.SQ010., p11= data4$kin.SQ011., p12= data4$kin.SQ012., p13= data4$kin.SQ013., p14= data4$kin.SQ014.,p15= data4$kin.SQ015., p16= data4$kin.SQ016., p17= data4$kin.SQ017., p18= data4$kin.SQ018., p19= data4$kin.SQ019., p20= data4$kin.SQ020.)

#age 
df_age <- data.frame(p1 = data4$age.SQ001.,p2 = data4$age.SQ002.,p3 = data4$age.SQ003.,p4 = data4$age.SQ004.,p5 = data4$age.SQ005.,p6 = data4$age.SQ006.,p7 = data4$age.SQ007.,p8 = data4$age.SQ008.,p9 = data4$age.SQ009., p10= data4$age.SQ010., p11= data4$age.SQ011., p12= data4$age.SQ012., p13= data4$age.SQ013., p14= data4$age.SQ014.,p15= data4$age.SQ015., p16= data4$age.SQ016., p17= data4$age.SQ017., p18= data4$age.SQ018., p19= data4$age.SQ019., p20= data4$age.SQ020.)

#educ 
df_educ <- data.frame(p1 = data4$educ.SQ001.,p2 = data4$educ.SQ002.,p3 = data4$educ.SQ003.,p4 = data4$educ.SQ004.,p5 = data4$educ.SQ005.,p6 = data4$educ.SQ006.,p7 = data4$educ.SQ007.,p8 = data4$educ.SQ008.,p9 = data4$educ.SQ009., p10= data4$educ.SQ010., p11= data4$educ.SQ011., p12= data4$educ.SQ012., p13= data4$educ.SQ013., p14= data4$educ.SQ014.,p15= data4$educ.SQ015., p16= data4$educ.SQ016., p17= data4$educ.SQ017., p18= data4$educ.SQ018., p19= data4$educ.SQ019., p20= data4$educ.SQ020.)

#duration 
df_duration <- data.frame(p1 = data4$duur.SQ001.,p2 = data4$duur.SQ002.,p3 = data4$duur.SQ003.,p4 = data4$duur.SQ004.,p5 = data4$duur.SQ005.,p6 = data4$duur.SQ006.,p7 = data4$duur.SQ007.,p8 = data4$duur.SQ008.,p9 = data4$duur.SQ009.,p10= data4$duur.SQ010.,p11= data4$duur.SQ011.,p12= data4$duur.SQ012.,p13= data4$duur.SQ013.,p14= data4$duur.SQ014.,p15= data4$duur.SQ015.,p16= data4$duur.SQ016.,p17= data4$duur.SQ017.,p18= data4$duur.SQ018.,p19= data4$duur.SQ019.,p20= data4$duur.SQ020.)

#proximity
df_proximity <- data.frame(p1 = data4$prox.SQ001.,p2 = data4$prox.SQ002.,p3 = data4$prox.SQ003.,p4 = data4$prox.SQ004.,p5 = data4$prox.SQ005.,p6 = data4$prox.SQ006.,p7 = data4$prox.SQ007.,p8 = data4$prox.SQ008., p9 = data4$prox.SQ009.,p10= data4$prox.SQ010., p11= data4$prox.SQ011., p12= data4$prox.SQ012., p13= data4$prox.SQ013., p14= data4$prox.SQ014., p15= data4$prox.SQ015., p16= data4$prox.SQ016., p17= data4$prox.SQ017., p18= data4$prox.SQ018., p19= data4$prox.SQ019., p20= data4$prox.SQ020.)

#and dynamic
#communication frequency
df_freq <- data.frame(p1 = data4$freq.SQ001.,p2 = data4$freq.SQ002.,p3 = data4$freq.SQ003.,p4 = data4$freq.SQ004.,p5 = data4$freq.SQ005.,p6 = data4$freq.SQ006.,p7 = data4$freq.SQ007.,p8 = data4$freq.SQ008., p9 = data4$freq.SQ009.,p10= data4$freq.SQ010., p11= data4$freq.SQ011., p12= data4$freq.SQ012., p13= data4$freq.SQ013., p14= data4$freq.SQ014., p15= data4$freq.SQ015., p16= data4$freq.SQ016., p17= data4$freq.SQ017., p18= data4$freq.SQ018., p19= data4$freq.SQ019., p20= data4$freq.SQ020.)

#closeness
df_close <- data.frame(p1 = data4$close.SQ001.,p2 = data4$close.SQ002.,p3 = data4$close.SQ003.,p4 = data4$close.SQ004.,p5 = data4$close.SQ005.,p6 = data4$close.SQ006.,p7 = data4$close.SQ007.,p8 = data4$close.SQ008., p9 = data4$close.SQ009.,p10= data4$close.SQ010., p11= data4$close.SQ011., p12= data4$close.SQ012., p13= data4$close.SQ013., p14= data4$close.SQ014., p15= data4$close.SQ015., p16= data4$close.SQ016., p17= data4$close.SQ017., p18= data4$close.SQ018., p19= data4$close.SQ019., p20= data4$close.SQ020.)

for (i in 1:length(alterL)) {
  
  alterL[[i]]$alter_gender <- ifelse(unlist(df_males[i,], use.names=F)=="Ja", 0, ifelse(unlist(df_males[i,], use.names=F)=="Nee", 1, NA))
  
  alterL[[i]]$alter_age <- 
    ifelse(unlist(df_age[i,], use.names=F)=="Jonger dan 18 jaar", 16, #???
           ifelse(unlist(df_age[i,], use.names=F)=="18 tot 21 jaar", 20,
                  ifelse(unlist(df_age[i,], use.names=F)=="22 tot 25 jaar", 23,
                         ifelse(unlist(df_age[i,], use.names=F)=="26 tot 30 jaar", 28,
                                ifelse(unlist(df_age[i,], use.names=F)=="31 tot 40 jaar", 35,
                                       ifelse(unlist(df_age[i,], use.names=F)=="Ouder dan 40 jaar", 45, #???
                                              ifelse(unlist(df_age[i,], use.names=F)=="Weet ik niet", NA, # i don't know = missing
                                                     unlist(df_age[i,], use.names=F))))))))
  alterL[[i]]$kin <- ifelse(unlist(df_kin[i,], use.names=F)=="Ja", 1, ifelse(unlist(df_kin[i,], use.names=F)=="Nee", 0, ""))
  
  alterL[[i]]$alter_educ <- ifelse(unlist(df_educ[i,],use.names=F)=="lagere school", 1,
                               ifelse(unlist(df_educ[i,],use.names=F)=="vmbo, mavo", 2,
                                      ifelse(unlist(df_educ[i,],use.names=F)=="mbo", 3,
                                         ifelse(unlist(df_educ[i,],use.names=F)=="havo", 4,
                                            ifelse(unlist(df_educ[i,],use.names=F)=="vwo / gymnasium", 5,
                                                 ifelse(unlist(df_educ[i,],use.names=F)=="hbo", 6,
                                                     ifelse(unlist(df_educ[i,],use.names=F)=="universiteit", 7, NA))))))) 
  
  alterL[[i]]$proximity <- ifelse(unlist(df_proximity[i,], use.names=FALSE) == "In hetzelfde huis", "roommate",
                                  ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In dezelfde buurt" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde straat" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde gemeente", "close",
                                    ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In hetzelfde land" |
                                             unlist(df_proximity[i,], use.names = FALSE) == "In een ander land","far", NA)))
  
  
  
  alterL[[i]]$duration <- ifelse(unlist(df_duration[i,],use.names=F)=="Minder dan 1 jaar", 0, 
                               ifelse(unlist(df_duration[i,],use.names=F)=="1 tot 3 jaar", 2,
                                      ifelse(unlist(df_duration[i,],use.names=F)=="4 tot 8 jaar", 6,
                                             ifelse(unlist(df_duration[i,],use.names=F)=="9 tot 15 jaar", 12,
                                                    ifelse(unlist(df_duration[i,],use.names=F)=="Meer dan 15 jaar", 15,NA )))))
  
  alterL[[i]]$frequency.t2 <- ifelse(unlist(df_freq[i,],use.names=F)=="(Bijna) elke dag", 7,
                               ifelse(unlist(df_freq[i,],use.names=F)=="1-2 keer per week",6,
                                      ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per maand",5, 
                                         ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per maand",4, 
                                            ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per jaar",3,
                                                 ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per jaar",2,
                                                     ifelse(unlist(df_freq[i,],use.names=F)=="Nooit",1, NA )))))))
  alterL[[i]]$closeness.t2 <- ifelse(unlist(df_close[i,],use.names=F)=="Niet hecht", 1, 
                               ifelse(unlist(df_close[i,],use.names=F)=="Enigszins hecht", 2,
                                      ifelse(unlist(df_close[i,],use.names=F)=="Hecht", 3,
                                             ifelse(unlist(df_close[i,],use.names=F)=="Heel erg hecht", 4, NA ))))
}

#match ego-covars (based on 'df'.) 
for (i in 1:length(alterL)) {
  alterL[[i]]$ego_gender <- df$ego_gender[which(df$ego==1)][1]
  alterL[[i]]$ego_age <- df$ego_age[which(df$ego==1)][1]
  alterL[[i]]$ego_educ <- df$ego_educ[which(df$ego==1)][1]
}

#sameness
for (i in 1:length(alterL))  { # for ego i
  # get attributes of ego
  agei <- alterL[[i]]$ego_age[1]
  genderi <- alterL[[i]]$ego_gender[1]
  edi <- alterL[[i]]$ego_educ[1]
  
  for (j in 1:max(alterL[[i]]$alterid)) { # for alter j
    # calculate "same gender" (0/1)
    genderj <- as.numeric(alterL[[i]]$alter_gender[j]) # get alter j gender
    same <- ifelse(genderi==genderj, 1, 0)
    alterL[[i]]$same_gender[which(alterL[[i]]$alterid==j)] <- same
    
    #same education
    eduj <- alterL[[i]]$alter_educ[j]
    same <- ifelse(eduj==edi, 1, 0)
    alterL[[i]]$sim_educ[which(alterL[[i]]$alterid==j)] <- same
    
    # calculate similarity for age as the absolute difference between alter and ego value
    #get alter j attributes
    agej <- as.numeric(alterL[[i]]$alter_age[j])
    #difference score
    difage <- abs(agei - agej)
    
    # so higher values represent *less* similarity
    if( !is.na (agej) ) { # age similarity only calculable if z_j is known!
        alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- difage
      } else { alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- NA}
  }
}
  
#in which egonets did altesr appear at t2?
for (i in 1:length(alterL))  { 
  for (j in 1:nrow(alterL[[i]])) {
    
    #get name(s) of j
    alter <- unlist(alterL[[i]][j,c(5:8)], use.names=FALSE)
    alter <- alter[!is.na(alter)] # exclude NAs
    
    #find out alter j appears in the 4 egonets
    cdn <- alter %in% net1[i,]
    study <- alter %in% net2[i,]
    bff <- alter %in% net3[i,]
    csn <- alter %in% net4[i,]
    
    # and if so, give alter j score 1 on indicators; 0 otherwise
    alterL[[i]]$cdn2[j] <- ifelse("TRUE" %in% cdn, 1, 0)
    alterL[[i]]$study2[j] <- ifelse("TRUE" %in% study, 1, 0)
    alterL[[i]]$bff2[j] <- ifelse("TRUE" %in% bff, 1, 0)
    alterL[[i]]$csn2[j] <- ifelse("TRUE" %in% csn, 1, 0)
  }
}

#calculate multiplexity as an alter/tie attribute (i.e., number of *additional* networks j appeared in)
for (i in 1:length(alterL))  {
  for (j in 1:nrow(alterL[[i]])) {
    alterL[[i]]$multiplex.t2[j] <- rowSums(alterL[[i]][j,c(24:27)]) - 1 #minus one to reach a meaningful intercept
  }
}

# i filter out the maintained alters
# by excluding alters with empty strings for gender attribute
for ( i in 1:length(alterL)) {
  alterL[[i]] <- alterL[[i]][which(!is.na(alterL[[i]]$alter_gender) & alterL[[i]]$alter_gender!=""),]
  
    if (nrow(alterL[[i]])>0){ #if no. of created alters > 0
      # replace the alter id: 1 : no. unique alters
      alterL[[i]]$alterid <- 1:nrow(alterL[[i]])
  }
}

#based on this, i make a long dataframe with alters nested in ego.
#first, add an ego_id
for (i in 1:length(alterL)) {
  if(nrow(alterL[[i]]>0)) {
    alterL[[i]]$ego <- i
    alterL[[i]]$respnr <- data$respnr[i]
    }
}

#combine using rbind
df <- do.call(rbind,alterL)

#add network variables; match based on df_maintained
df$cdn.size2 <- df$study.size2 <- df$bff.size2 <- df$csn.size2 <- NA
df$cdn.density2 <- df$study.density2 <- df$bff.density2 <- df$csn.density2 <- NA

for ( i in unique(df$ego)) { 
  df$cdn.size2[which(df$ego == i)] <- df_maintained$cdn.size2[which(df_maintained$ego == i)][1]
  df$study.size2[which(df$ego == i)] <- df_maintained$study.size2[which(df_maintained$ego == i)][1]
  df$bff.size2[which(df$ego == i)] <- df_maintained$bff.size2[which(df_maintained$ego == i)][1]
  df$csn.size2[which(df$ego == i)] <- df_maintained$csn.size2[which(df_maintained$ego == i)][1]
  
  df$cdn.density2[which(df$ego == i)] <- df_maintained$cdn.density2[which(df_maintained$ego == i)][1]
  df$study.density2[which(df$ego == i)] <- df_maintained$study.density2[which(df_maintained$ego == i)][1]
  df$bff.density2[which(df$ego == i)] <- df_maintained$bff.density2[which(df_maintained$ego == i)][1]
  df$csn.density2[which(df$ego == i)] <- df_maintained$csn.density2[which(df_maintained$ego == i)][1]
}

#methodological controls??

df$statusW2 <- "Created"
df_created <- df

#some empty name generator entries were saved, due to a bug.
#empties <- which(is.na(df$name1)) #done using the data with names instead of alter_ids
#save(empties,file="./data/empty_entries_indicators.Rda")
empties <- fload("./data_shared/empty_entries_indicators.Rda")

df_created <- df_created[-empties,]

# bind them together
df_alters <- dplyr::bind_rows(df_maintained, df_created)

# arrange, by ego, and status
df_alters <- df_alters %>%
    arrange(ego, factor(statusW2, levels = c("Maintained", "Created", "Dropped")))
row.names(df_alters) <- 1:nrow(df_alters)

wave 2 –> wave 3

Let’s see whether wave 2 alters were maintained into wave 3.

# first subset from `df_alters` w2-maintained or w2-created alters
test <- df_alters[which(df_alters$statusW2 == "Maintained" | df_alters$statusW2 == "Created"), ]

# make variable indicating whether alter was (re-)named in w3
test$surviveW3 <- NA

# some alters had no 'chance' to reappear, because ego did fill out the last survey
test$w3participation <- NA
for (i in unique(test$respnr)) {
    test$w3participation[which(test$respnr == i)] <- ifelse(i %in% data5$respnr, 1, 0)
}

# fix alter ids, 1:N
for (i in unique(test$respnr)) {
    test$alterid[which(test$respnr == i)] <- 1:length(test$alterid[which(test$respnr == i)])
}

# subset wave 5 matching matrix (matching wave 3 alters to wave 2 alters - either created or
# maintained ) also add respnr
w1w2 <- data5[, c(486:885, ncol(data5))]

for (i in unique(test$respnr)) {
    # for all egos (both those that participated in w3, and those that didnt)

    matchingL <- vector("list", 20)  #pre-allocate empty list of length 20, to store matching matrices
    {
        matchingL[[1]] <- w1w2[which(w1w2$respnr == i), 1:20]
        matchingL[[2]] <- w1w2[which(w1w2$respnr == i), 21:40]
        matchingL[[3]] <- w1w2[which(w1w2$respnr == i), 41:60]
        matchingL[[4]] <- w1w2[which(w1w2$respnr == i), 61:80]
        matchingL[[5]] <- w1w2[which(w1w2$respnr == i), 81:100]
        matchingL[[6]] <- w1w2[which(w1w2$respnr == i), 101:120]
        matchingL[[7]] <- w1w2[which(w1w2$respnr == i), 121:140]
        matchingL[[8]] <- w1w2[which(w1w2$respnr == i), 141:160]
        matchingL[[9]] <- w1w2[which(w1w2$respnr == i), 161:180]
        matchingL[[10]] <- w1w2[which(w1w2$respnr == i), 181:200]
        matchingL[[11]] <- w1w2[which(w1w2$respnr == i), 201:220]
        matchingL[[12]] <- w1w2[which(w1w2$respnr == i), 221:240]
        matchingL[[13]] <- w1w2[which(w1w2$respnr == i), 241:260]
        matchingL[[14]] <- w1w2[which(w1w2$respnr == i), 261:280]
        matchingL[[15]] <- w1w2[which(w1w2$respnr == i), 281:300]
        matchingL[[16]] <- w1w2[which(w1w2$respnr == i), 301:320]
        matchingL[[17]] <- w1w2[which(w1w2$respnr == i), 321:340]
        matchingL[[18]] <- w1w2[which(w1w2$respnr == i), 341:360]
        matchingL[[19]] <- w1w2[which(w1w2$respnr == i), 361:380]
        matchingL[[20]] <- w1w2[which(w1w2$respnr == i), 381:400]
    }
    # find the 'right' matching matrix in this list, by taking the one that contains answers
    ind <- NULL
    for (j in seq_along(matchingL)) {
        # check along the sequence of elements j in the matching matrix list if they are non-empty
        # and not NA
        check <- FALSE
        for (col in matchingL[[j]]) {
            if (!all(is.na(col)) && any(nchar(col) > 0)) {
                check <- TRUE
                break
            }
        }
        if (check) {
            ind <- j
            break
        }
    }
    if (length(ind) > 0) {
        # get the matrix
        mm <- matchingL[[ind]]
        # unlist the answers given
        ans <- unlist(mm, use.names = FALSE)
        # use `stringr` to extract numbers in these answers
        id <- unlist(stringr::str_extract_all(ans, "\\(?[0-9,.]+\\)?"))
        id <- id[!is.na(id)]  #remove NA

        for (k in 1:length(which(test$respnr == i))) {
            # for w2-alters of ego i, with ids 1:k, if alter was renamed, give score 1, if he/she
            # was not renamed give 0, alters who had no chance to reappear due to ego not filling
            # out the survey keep NA.
            test$surviveW3[which(test$respnr == i & test$alterid == k)] <- ifelse(test$alterid[which(test$respnr ==
                i & test$alterid == k)] %in% id, 1, 0)
        }
    }
}

# remaining NAs on `survivew3` among egos who did fill out w3, are due to ego not listing any
# alters.. any(test$w3participation==1 & is.na(test$surviveW3))
# test$respnr[which(test$w3participation==1 & is.na(test$surviveW3))] so, set to 0...
test$surviveW3[which(is.na(test$surviveW3) & test$w3participation == 1)] <- 0

# in which egonet did they (re)appear?
test$cdn3 <- NA
test$study3 <- NA
test$bff3 <- NA
test$csn3 <- NA

# i use the matching matrices of w3; where columns refer to unique w2 alters (maintained and
# created; those of our constructed dataframe) and rows referring to unique w3 alters so, if
# w2-alter/column j is matched to w3-alter/row i, then w2-alter did survive (but we already knew
# that); but more importantly, j was the i^th alter mentioned and this allows me to infer in what
# network j (re-)appeared since cdn = 1-5; study = 6-10; bff = 11-15; csn = 16-20.  however, since
# the rows only include non-duplicate w3-alters; if alter j at t2 was named more than once at t3, i
# only know the first egonet in which j was named...  thus, before i proceed, i use the matching
# matrices that allowed ego to match alters from different egonets to make alter ids for w3 alters;
# and figure out to which egonets they belonged

# first, subset data5, bc we only want names data of egos with w1/w2 alters who participated in w3
data5a <- data5[which(data5$respnr %in% unique(test$respnr[which(test$w3participation == 1)])), ]

# i make a dataframe of w3-alters, with potential duplicates...
df_names <- data.frame(p1 = data5a$egonet1.SQ001., p2 = data5a$egonet1.SQ002., p3 = data5a$egonet1.SQ003.,
    p4 = data5a$egonet1.SQ004., p5 = data5a$egonet1.SQ005., p6 = data5a$egonet2.SQ001., p7 = data5a$egonet2.SQ002.,
    p8 = data5a$egonet2.SQ003., p9 = data5a$egonet2.SQ004., p10 = data5a$egonet2.SQ005., p11 = data5a$egonet3.SQ001.,
    p12 = data5a$egonet3.SQ002., p13 = data5a$egonet3.SQ003., p14 = data5a$egonet3.SQ004., p15 = data5a$egonet3.SQ005.,
    p16 = data5a$egonet4.SQ001., p17 = data5a$egonet4.SQ002., p18 = data5a$egonet4.SQ003., p19 = data5a$egonet4.SQ004.,
    p20 = data5a$egonet4.SQ005.)

# list of dataframes per ego with rows reflecting alters and columns indicating the name(s) of the
# particular alters
alterL <- list()
# loop over all egos, that filled out w3
for (i in 1:nrow(data5a)) {
    alterL[[i]] <- data.frame(alterid = 1:20, name1 = NA, name2 = NA, name3 = NA, name4 = NA)
}

# fill the names based on the names data-frame
for (i in 1:length(alterL)) {
    alterL[[i]]$name1 <- unlist(df_names[i, ], use.names = FALSE)
    alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1 == "", NA, alterL[[i]]$name1)
}

# calculate netsize for each net to get the correct matching matrix
{
    net1 <- cbind(data5a$egonet1.SQ001., data5a$egonet1.SQ002., data5a$egonet1.SQ003., data5a$egonet1.SQ004.,
        data5a$egonet1.SQ005.)
    net1 <- ifelse(net1 == "", NA, net1)
    ns1 <- vector()
    for (i in 1:nrow(net1)) {
        ns1[i] <- length(net1[i, ][which(!is.na(net1[i, ]))])
    }
    net2 <- cbind(data5a$egonet2.SQ001., data5a$egonet2.SQ002., data5a$egonet2.SQ003., data5a$egonet2.SQ004.,
        data5a$egonet2.SQ005.)
    net2 <- ifelse(net2 == "", NA, net2)
    ns2 <- vector()
    for (i in 1:nrow(net2)) {
        ns2[i] <- length(net2[i, ][which(!is.na(net2[i, ]))])
    }
    net3 <- cbind(data5a$egonet3.SQ001., data5a$egonet3.SQ002., data5a$egonet3.SQ003., data5a$egonet3.SQ004.,
        data5a$egonet3.SQ005.)
    net3 <- ifelse(net3 == "", NA, net3)
    ns3 <- vector()
    for (i in 1:nrow(net3)) {
        ns3[i] <- length(net3[i, ][which(!is.na(net3[i, ]))])
    }
    net4 <- cbind(data5a$egonet4.SQ001., data5a$egonet4.SQ002., data5a$egonet4.SQ003., data5a$egonet4.SQ004.,
        data5a$egonet4.SQ005.)
    net4 <- ifelse(net4 == "", NA, net4)
    ns4 <- vector()
    for (i in 1:nrow(net4)) {
        ns4[i] <- length(net4[i, ][which(!is.na(net4[i, ]))])
    }
}
# construct the matching matrices list for egonet1-2
matchingList <- list()
for (i in 1:length(alterL)) {
    matchingL <- list()
    matchingL[[1]] <- cbind(data5a$matching1N1.SQ001_SQ001.[i], data5a$matching1N1.SQ002_SQ001.[i], data5a$matching1N1.SQ003_SQ001.[i],
        data5a$matching1N1.SQ004_SQ001.[i], data5a$matching1N1.SQ005_SQ001.[i])
    matchingL[[2]] <- rbind(cbind(data5a$matching1N2.SQ001_SQ001.[i], data5a$matching1N2.SQ002_SQ001.[i],
        data5a$matching1N2.SQ003_SQ001.[i], data5a$matching1N2.SQ004_SQ001.[i], data5a$matching1N2.SQ005_SQ001.[i]),
        cbind(data5a$matching1N2.SQ001_SQ002.[i], data5a$matching1N2.SQ002_SQ002.[i], data5a$matching1N2.SQ003_SQ002.[i],
            data5a$matching1N2.SQ004_SQ002.[i], data5a$matching1N2.SQ005_SQ002.[i]))
    matchingL[[3]] <- rbind(cbind(data5a$matching1N3.SQ001_SQ001.[i], data5a$matching1N3.SQ002_SQ001.[i],
        data5a$matching1N3.SQ003_SQ001.[i], data5a$matching1N3.SQ004_SQ001.[i], data5a$matching1N3.SQ005_SQ001.[i]),
        cbind(data5a$matching1N3.SQ001_SQ002.[i], data5a$matching1N3.SQ002_SQ002.[i], data5a$matching1N3.SQ003_SQ002.[i],
            data5a$matching1N3.SQ004_SQ002.[i], data5a$matching1N3.SQ005_SQ002.[i]), cbind(data5a$matching1N3.SQ001_SQ003.[i],
            data5a$matching1N3.SQ002_SQ003.[i], data5a$matching1N3.SQ003_SQ003.[i], data5a$matching1N3.SQ004_SQ003.[i],
            data5a$matching1N3.SQ005_SQ003.[i]))
    matchingL[[4]] <- rbind(cbind(data5a$matching1N4.SQ001_SQ001.[i], data5a$matching1N4.SQ002_SQ001.[i],
        data5a$matching1N4.SQ003_SQ001.[i], data5a$matching1N4.SQ004_SQ001.[i], data5a$matching1N4.SQ005_SQ001.[i]),
        cbind(data5a$matching1N4.SQ001_SQ002.[i], data5a$matching1N4.SQ002_SQ002.[i], data5a$matching1N4.SQ003_SQ002.[i],
            data5a$matching1N4.SQ004_SQ002.[i], data5a$matching1N4.SQ005_SQ002.[i]), cbind(data5a$matching1N4.SQ001_SQ003.[i],
            data5a$matching1N4.SQ002_SQ003.[i], data5a$matching1N4.SQ003_SQ003.[i], data5a$matching1N4.SQ004_SQ003.[i],
            data5a$matching1N4.SQ005_SQ003.[i]), cbind(data5a$matching1N4.SQ001_SQ004.[i], data5a$matching1N4.SQ002_SQ004.[i],
            data5a$matching1N4.SQ003_SQ004.[i], data5a$matching1N4.SQ004_SQ004.[i], data5a$matching1N4.SQ005_SQ004.[i]))
    matchingL[[5]] <- rbind(cbind(data5a$matching1N5.SQ001_SQ001.[i], data5a$matching1N5.SQ002_SQ001.[i],
        data5a$matching1N5.SQ003_SQ001.[i], data5a$matching1N5.SQ004_SQ001.[i], data5a$matching1N5.SQ005_SQ001.[i]),
        cbind(data5a$matching1N5.SQ001_SQ002.[i], data5a$matching1N5.SQ002_SQ002.[i], data5a$matching1N5.SQ003_SQ002.[i],
            data5a$matching1N5.SQ004_SQ002.[i], data5a$matching1N5.SQ005_SQ002.[i]), cbind(data5a$matching1N5.SQ001_SQ003.[i],
            data5a$matching1N5.SQ002_SQ003.[i], data5a$matching1N5.SQ003_SQ003.[i], data5a$matching1N5.SQ004_SQ003.[i],
            data5a$matching1N5.SQ005_SQ003.[i]), cbind(data5a$matching1N5.SQ001_SQ004.[i], data5a$matching1N5.SQ002_SQ004.[i],
            data5a$matching1N5.SQ003_SQ004.[i], data5a$matching1N5.SQ004_SQ004.[i], data5a$matching1N5.SQ005_SQ004.[i]),
        cbind(data5a$matching1N5.SQ001_SQ005.[i], data5a$matching1N5.SQ002_SQ005.[i], data5a$matching1N5.SQ003_SQ005.[i],
            data5a$matching1N5.SQ004_SQ005.[i], data5a$matching1N5.SQ005_SQ005.[i]))
    matchingList[[i]] <- matchingL
}  # so... matchingL[[1]][[5]] is matchingmatrix 5 (i.e., 5 alters named in egonet2) for ego 1

# the combination of the matching matrices and the netsizes for ego allows me to match the names
# myself.  for ego i
for (i in 1:length(matchingList)) {
    mL <- matchingList[[i]]  # get the matching list
    ns <- ns2[[i]]  # get the size of egonet2
    if (ns > 0) {
        # if ns=0, no matching was done!
        mm <- as.matrix(mL[[ns]])  # retrieve the corresponding matrix
        matched <- which(mm == 1, arr.ind = T)  # retrieve array indices
        net <- net2[i, ]
        if (length(matched) > 0) {
            # if matching was performed!
            alterL[[i]]$name2[which(alterL[[i]]$alterid == matched[, 2])] <- net[matched[, 1]]
        }
    }
}  #ignore warning

# again, make a matching list for the second matching matrices (i.e., matching egonet3 alters to
# prev. alters);
matchingList2 <- list()
for (i in 1:length(alterL)) {
    # for ego i
    matching2L <- list()
    matching2L[[1]] <- cbind(data5a$matching2N1.SQ001_SQ001.[i], data5a$matching2N1.SQ002_SQ001.[i],
        data5a$matching2N1.SQ003_SQ001.[i], data5a$matching2N1.SQ004_SQ001.[i], data5a$matching2N1.SQ005_SQ001.[i],
        data5a$matching2N1.SQ006_SQ001.[i], data5a$matching2N1.SQ007_SQ001.[i], data5a$matching2N1.SQ008_SQ001.[i],
        data5a$matching2N1.SQ009_SQ001.[i], data5a$matching2N1.SQ010_SQ001.[i])
    matching2L[[2]] <- rbind(cbind(data5a$matching2N2.SQ001_SQ001.[i], data5a$matching2N2.SQ002_SQ001.[i],
        data5a$matching2N2.SQ003_SQ001.[i], data5a$matching2N2.SQ004_SQ001.[i], data5a$matching2N2.SQ005_SQ001.[i],
        data5a$matching2N2.SQ006_SQ001.[i], data5a$matching2N2.SQ007_SQ001.[i], data5a$matching2N2.SQ008_SQ001.[i],
        data5a$matching2N2.SQ009_SQ001.[i], data5a$matching2N2.SQ010_SQ001.[i]), cbind(data5a$matching2N2.SQ001_SQ002.[i],
        data5a$matching2N2.SQ002_SQ002.[i], data5a$matching2N2.SQ003_SQ002.[i], data5a$matching2N2.SQ004_SQ002.[i],
        data5a$matching2N2.SQ005_SQ002.[i], data5a$matching2N2.SQ006_SQ002.[i], data5a$matching2N2.SQ007_SQ002.[i],
        data5a$matching2N2.SQ008_SQ002.[i], data5a$matching2N2.SQ009_SQ002.[i], data5a$matching2N2.SQ010_SQ002.[i]))
    matching2L[[3]] <- rbind(cbind(data5a$matching2N3.SQ001_SQ001.[i], data5a$matching2N3.SQ002_SQ001.[i],
        data5a$matching2N3.SQ003_SQ001.[i], data5a$matching2N3.SQ004_SQ001.[i], data5a$matching2N3.SQ005_SQ001.[i],
        data5a$matching2N3.SQ006_SQ001.[i], data5a$matching2N3.SQ007_SQ001.[i], data5a$matching2N3.SQ008_SQ001.[i],
        data5a$matching2N3.SQ009_SQ001.[i], data5a$matching2N3.SQ010_SQ001.[i]), cbind(data5a$matching2N3.SQ001_SQ002.[i],
        data5a$matching2N3.SQ002_SQ002.[i], data5a$matching2N3.SQ003_SQ002.[i], data5a$matching2N3.SQ004_SQ002.[i],
        data5a$matching2N3.SQ005_SQ002.[i], data5a$matching2N3.SQ006_SQ002.[i], data5a$matching2N3.SQ007_SQ002.[i],
        data5a$matching2N3.SQ008_SQ002.[i], data5a$matching2N3.SQ009_SQ002.[i], data5a$matching2N3.SQ010_SQ002.[i]),
        cbind(data5a$matching2N3.SQ001_SQ003.[i], data5a$matching2N3.SQ002_SQ003.[i], data5a$matching2N3.SQ003_SQ003.[i],
            data5a$matching2N3.SQ004_SQ003.[i], data5a$matching2N3.SQ005_SQ003.[i], data5a$matching2N3.SQ006_SQ003.[i],
            data5a$matching2N3.SQ007_SQ003.[i], data5a$matching2N3.SQ008_SQ003.[i], data5a$matching2N3.SQ009_SQ003.[i],
            data5a$matching2N3.SQ010_SQ003.[i]))
    matching2L[[4]] <- rbind(cbind(data5a$matching2N4.SQ001_SQ001.[i], data5a$matching2N4.SQ002_SQ001.[i],
        data5a$matching2N4.SQ003_SQ001.[i], data5a$matching2N4.SQ004_SQ001.[i], data5a$matching2N4.SQ005_SQ001.[i],
        data5a$matching2N4.SQ006_SQ001.[i], data5a$matching2N4.SQ007_SQ001.[i], data5a$matching2N4.SQ008_SQ001.[i],
        data5a$matching2N4.SQ009_SQ001.[i], data5a$matching2N4.SQ010_SQ001.[i]), cbind(data5a$matching2N4.SQ001_SQ002.[i],
        data5a$matching2N4.SQ002_SQ002.[i], data5a$matching2N4.SQ003_SQ002.[i], data5a$matching2N4.SQ004_SQ002.[i],
        data5a$matching2N4.SQ005_SQ002.[i], data5a$matching2N4.SQ006_SQ002.[i], data5a$matching2N4.SQ007_SQ002.[i],
        data5a$matching2N4.SQ008_SQ002.[i], data5a$matching2N4.SQ009_SQ002.[i], data5a$matching2N4.SQ010_SQ002.[i]),
        cbind(data5a$matching2N4.SQ001_SQ003.[i], data5a$matching2N4.SQ002_SQ003.[i], data5a$matching2N4.SQ003_SQ003.[i],
            data5a$matching2N4.SQ004_SQ003.[i], data5a$matching2N4.SQ005_SQ003.[i], data5a$matching2N4.SQ006_SQ003.[i],
            data5a$matching2N4.SQ007_SQ003.[i], data5a$matching2N4.SQ008_SQ003.[i], data5a$matching2N4.SQ009_SQ003.[i],
            data5a$matching2N4.SQ010_SQ003.[i]), cbind(data5a$matching2N4.SQ001_SQ004.[i], data5a$matching2N4.SQ002_SQ004.[i],
            data5a$matching2N4.SQ003_SQ004.[i], data5a$matching2N4.SQ004_SQ004.[i], data5a$matching2N4.SQ005_SQ004.[i],
            data5a$matching2N4.SQ006_SQ004.[i], data5a$matching2N4.SQ007_SQ004.[i], data5a$matching2N4.SQ008_SQ004.[i],
            data5a$matching2N4.SQ009_SQ004.[i], data5a$matching2N4.SQ010_SQ004.[i]))
    matching2L[[5]] <- rbind(cbind(data5a$matching2N5.SQ001_SQ001.[i], data5a$matching2N5.SQ002_SQ001.[i],
        data5a$matching2N5.SQ003_SQ001.[i], data5a$matching2N5.SQ004_SQ001.[i], data5a$matching2N5.SQ005_SQ001.[i],
        data5a$matching2N5.SQ006_SQ001.[i], data5a$matching2N5.SQ007_SQ001.[i], data5a$matching2N5.SQ008_SQ001.[i],
        data5a$matching2N5.SQ009_SQ001.[i], data5a$matching2N5.SQ010_SQ001.[i]), cbind(data5a$matching2N5.SQ001_SQ002.[i],
        data5a$matching2N5.SQ002_SQ002.[i], data5a$matching2N5.SQ003_SQ002.[i], data5a$matching2N5.SQ004_SQ002.[i],
        data5a$matching2N5.SQ005_SQ002.[i], data5a$matching2N5.SQ006_SQ002.[i], data5a$matching2N5.SQ007_SQ002.[i],
        data5a$matching2N5.SQ008_SQ002.[i], data5a$matching2N5.SQ009_SQ002.[i], data5a$matching2N5.SQ010_SQ002.[i]),
        cbind(data5a$matching2N5.SQ001_SQ003.[i], data5a$matching2N5.SQ002_SQ003.[i], data5a$matching2N5.SQ003_SQ003.[i],
            data5a$matching2N5.SQ004_SQ003.[i], data5a$matching2N5.SQ005_SQ003.[i], data5a$matching2N5.SQ006_SQ003.[i],
            data5a$matching2N5.SQ007_SQ003.[i], data5a$matching2N5.SQ008_SQ003.[i], data5a$matching2N5.SQ009_SQ003.[i],
            data5a$matching2N5.SQ010_SQ003.[i]), cbind(data5a$matching2N5.SQ001_SQ004.[i], data5a$matching2N5.SQ002_SQ004.[i],
            data5a$matching2N5.SQ003_SQ004.[i], data5a$matching2N5.SQ004_SQ004.[i], data5a$matching2N5.SQ005_SQ004.[i],
            data5a$matching2N5.SQ006_SQ004.[i], data5a$matching2N5.SQ007_SQ004.[i], data5a$matching2N5.SQ008_SQ004.[i],
            data5a$matching2N5.SQ009_SQ004.[i], data5a$matching2N5.SQ010_SQ004.[i]), cbind(data5a$matching2N5.SQ001_SQ005.[i],
            data5a$matching2N5.SQ002_SQ005.[i], data5a$matching2N5.SQ003_SQ005.[i], data5a$matching2N5.SQ004_SQ005.[i],
            data5a$matching2N5.SQ005_SQ005.[i], data5a$matching2N5.SQ006_SQ005.[i], data5a$matching2N5.SQ007_SQ005.[i],
            data5a$matching2N5.SQ008_SQ005.[i], data5a$matching2N5.SQ009_SQ005.[i], data5a$matching2N5.SQ010_SQ005.[i]))
    matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {
    # for ego i
    mL <- matchingList2[[i]]  # get the matching list 2
    ns <- ns3[[i]]  # get the size of egonet3
    if (ns > 0) {
        # if ns=0, no matching was done!
        mm <- as.matrix(mL[[ns]])  # and the corresponding matrix
        matched <- which(mm == 1, arr.ind = T)  # retrieve array indices
        net <- net3[i, ]
        if (length(matched) > 0) {
            # if matching was performed!
            alterL[[i]]$name3[matched[, 2]] <- net[matched[, 1]]
        }
    }
}

# same for matching 3 (i.e., egonet4 with egonets 1-3)
matchingList3 <- list()
for (i in 1:length(alterL)) {
    # for ego i
    matching3L <- list()
    matching3L[[1]] <- cbind(data5a$matching3N1.SQ001_SQ001.[i], data5a$matching3N1.SQ002_SQ001.[i],
        data5a$matching3N1.SQ003_SQ001.[i], data5a$matching3N1.SQ004_SQ001.[i], data5a$matching3N1.SQ005_SQ001.[i],
        data5a$matching3N1.SQ006_SQ001.[i], data5a$matching3N1.SQ007_SQ001.[i], data5a$matching3N1.SQ008_SQ001.[i],
        data5a$matching3N1.SQ009_SQ001.[i], data5a$matching3N1.SQ010_SQ001.[i], data5a$matching3N1.SQ011_SQ001.[i],
        data5a$matching3N1.SQ012_SQ001.[i], data5a$matching3N1.SQ013_SQ001.[i], data5a$matching3N1.SQ014_SQ001.[i],
        data5a$matching3N1.SQ015_SQ001.[i])
    matching3L[[2]] <- rbind(cbind(data5a$matching3N2.SQ001_SQ001.[i], data5a$matching3N2.SQ002_SQ001.[i],
        data5a$matching3N2.SQ003_SQ001.[i], data5a$matching3N2.SQ004_SQ001.[i], data5a$matching3N2.SQ005_SQ001.[i],
        data5a$matching3N2.SQ006_SQ001.[i], data5a$matching3N2.SQ007_SQ001.[i], data5a$matching3N2.SQ008_SQ001.[i],
        data5a$matching3N2.SQ009_SQ001.[i], data5a$matching3N2.SQ010_SQ001.[i], data5a$matching3N2.SQ011_SQ001.[i],
        data5a$matching3N2.SQ012_SQ001.[i], data5a$matching3N2.SQ013_SQ001.[i], data5a$matching3N2.SQ014_SQ001.[i],
        data5a$matching3N2.SQ015_SQ001.[i]), cbind(data5a$matching3N2.SQ001_SQ002.[i], data5a$matching3N2.SQ002_SQ002.[i],
        data5a$matching3N2.SQ003_SQ002.[i], data5a$matching3N2.SQ004_SQ002.[i], data5a$matching3N2.SQ005_SQ002.[i],
        data5a$matching3N2.SQ006_SQ002.[i], data5a$matching3N2.SQ007_SQ002.[i], data5a$matching3N2.SQ008_SQ002.[i],
        data5a$matching3N2.SQ009_SQ002.[i], data5a$matching3N2.SQ010_SQ002.[i], data5a$matching3N2.SQ011_SQ002.[i],
        data5a$matching3N2.SQ012_SQ002.[i], data5a$matching3N2.SQ013_SQ002.[i], data5a$matching3N2.SQ014_SQ002.[i],
        data5a$matching3N2.SQ015_SQ002.[i]))
    matching3L[[3]] <- rbind(cbind(data5a$matching3N3.SQ001_SQ001.[i], data5a$matching3N3.SQ002_SQ001.[i],
        data5a$matching3N3.SQ003_SQ001.[i], data5a$matching3N3.SQ004_SQ001.[i], data5a$matching3N3.SQ005_SQ001.[i],
        data5a$matching3N3.SQ006_SQ001.[i], data5a$matching3N3.SQ007_SQ001.[i], data5a$matching3N3.SQ008_SQ001.[i],
        data5a$matching3N3.SQ009_SQ001.[i], data5a$matching3N3.SQ010_SQ001.[i], data5a$matching3N3.SQ011_SQ001.[i],
        data5a$matching3N3.SQ012_SQ001.[i], data5a$matching3N3.SQ013_SQ001.[i], data5a$matching3N3.SQ014_SQ001.[i],
        data5a$matching3N3.SQ015_SQ001.[i]), cbind(data5a$matching3N3.SQ001_SQ002.[i], data5a$matching3N3.SQ002_SQ002.[i],
        data5a$matching3N3.SQ003_SQ002.[i], data5a$matching3N3.SQ004_SQ002.[i], data5a$matching3N3.SQ005_SQ002.[i],
        data5a$matching3N3.SQ006_SQ002.[i], data5a$matching3N3.SQ007_SQ002.[i], data5a$matching3N3.SQ008_SQ002.[i],
        data5a$matching3N3.SQ009_SQ002.[i], data5a$matching3N3.SQ010_SQ002.[i], data5a$matching3N3.SQ011_SQ002.[i],
        data5a$matching3N3.SQ012_SQ002.[i], data5a$matching3N3.SQ013_SQ002.[i], data5a$matching3N3.SQ014_SQ002.[i],
        data5a$matching3N3.SQ015_SQ002.[i]), cbind(data5a$matching3N3.SQ001_SQ003.[i], data5a$matching3N3.SQ002_SQ003.[i],
        data5a$matching3N3.SQ003_SQ003.[i], data5a$matching3N3.SQ004_SQ003.[i], data5a$matching3N3.SQ005_SQ003.[i],
        data5a$matching3N3.SQ006_SQ003.[i], data5a$matching3N3.SQ007_SQ003.[i], data5a$matching3N3.SQ008_SQ003.[i],
        data5a$matching3N3.SQ009_SQ003.[i], data5a$matching3N3.SQ010_SQ003.[i], data5a$matching3N3.SQ011_SQ003.[i],
        data5a$matching3N3.SQ012_SQ003.[i], data5a$matching3N3.SQ013_SQ003.[i], data5a$matching3N3.SQ014_SQ003.[i],
        data5a$matching3N3.SQ015_SQ003.[i]))
    matching3L[[4]] <- rbind(cbind(data5a$matching3N4.SQ001_SQ001.[i], data5a$matching3N4.SQ002_SQ001.[i],
        data5a$matching3N4.SQ003_SQ001.[i], data5a$matching3N4.SQ004_SQ001.[i], data5a$matching3N4.SQ005_SQ001.[i],
        data5a$matching3N4.SQ006_SQ001.[i], data5a$matching3N4.SQ007_SQ001.[i], data5a$matching3N4.SQ008_SQ001.[i],
        data5a$matching3N4.SQ009_SQ001.[i], data5a$matching3N4.SQ010_SQ001.[i], data5a$matching3N4.SQ011_SQ001.[i],
        data5a$matching3N4.SQ012_SQ001.[i], data5a$matching3N4.SQ013_SQ001.[i], data5a$matching3N4.SQ014_SQ001.[i],
        data5a$matching3N4.SQ015_SQ001.[i]), cbind(data5a$matching3N4.SQ001_SQ002.[i], data5a$matching3N4.SQ002_SQ002.[i],
        data5a$matching3N4.SQ003_SQ002.[i], data5a$matching3N4.SQ004_SQ002.[i], data5a$matching3N4.SQ005_SQ002.[i],
        data5a$matching3N4.SQ006_SQ002.[i], data5a$matching3N4.SQ007_SQ002.[i], data5a$matching3N4.SQ008_SQ002.[i],
        data5a$matching3N4.SQ009_SQ002.[i], data5a$matching3N4.SQ010_SQ002.[i], data5a$matching3N4.SQ011_SQ002.[i],
        data5a$matching3N4.SQ012_SQ002.[i], data5a$matching3N4.SQ013_SQ002.[i], data5a$matching3N4.SQ014_SQ002.[i],
        data5a$matching3N4.SQ015_SQ002.[i]), cbind(data5a$matching3N4.SQ001_SQ003.[i], data5a$matching3N4.SQ002_SQ003.[i],
        data5a$matching3N4.SQ003_SQ003.[i], data5a$matching3N4.SQ004_SQ003.[i], data5a$matching3N4.SQ005_SQ003.[i],
        data5a$matching3N4.SQ006_SQ003.[i], data5a$matching3N4.SQ007_SQ003.[i], data5a$matching3N4.SQ008_SQ003.[i],
        data5a$matching3N4.SQ009_SQ003.[i], data5a$matching3N4.SQ010_SQ003.[i], data5a$matching3N4.SQ011_SQ003.[i],
        data5a$matching3N4.SQ012_SQ003.[i], data5a$matching3N4.SQ013_SQ003.[i], data5a$matching3N4.SQ014_SQ003.[i],
        data5a$matching3N4.SQ015_SQ003.[i]), cbind(data5a$matching3N4.SQ001_SQ003.[i], data5a$matching3N4.SQ002_SQ003.[i],
        data5a$matching3N4.SQ003_SQ003.[i], data5a$matching3N4.SQ004_SQ003.[i], data5a$matching3N4.SQ005_SQ003.[i],
        data5a$matching3N4.SQ006_SQ003.[i], data5a$matching3N4.SQ007_SQ003.[i], data5a$matching3N4.SQ008_SQ003.[i],
        data5a$matching3N4.SQ009_SQ003.[i], data5a$matching3N4.SQ010_SQ003.[i], data5a$matching3N4.SQ011_SQ003.[i],
        data5a$matching3N4.SQ012_SQ003.[i], data5a$matching3N4.SQ013_SQ003.[i], data5a$matching3N4.SQ014_SQ003.[i],
        data5a$matching3N4.SQ015_SQ003.[i]), cbind(data5a$matching3N4.SQ001_SQ004.[i], data5a$matching3N4.SQ002_SQ004.[i],
        data5a$matching3N4.SQ003_SQ004.[i], data5a$matching3N4.SQ004_SQ004.[i], data5a$matching3N4.SQ005_SQ004.[i],
        data5a$matching3N4.SQ006_SQ004.[i], data5a$matching3N4.SQ007_SQ004.[i], data5a$matching3N4.SQ008_SQ004.[i],
        data5a$matching3N4.SQ009_SQ004.[i], data5a$matching3N4.SQ010_SQ004.[i], data5a$matching3N4.SQ011_SQ004.[i],
        data5a$matching3N4.SQ012_SQ004.[i], data5a$matching3N4.SQ013_SQ004.[i], data5a$matching3N4.SQ014_SQ004.[i],
        data5a$matching3N4.SQ015_SQ004.[i]))
    matching3L[[5]] <- rbind(cbind(data5a$matching3N5.SQ001_SQ001.[i], data5a$matching3N5.SQ002_SQ001.[i],
        data5a$matching3N5.SQ003_SQ001.[i], data5a$matching3N5.SQ004_SQ001.[i], data5a$matching3N5.SQ005_SQ001.[i],
        data5a$matching3N5.SQ006_SQ001.[i], data5a$matching3N5.SQ007_SQ001.[i], data5a$matching3N5.SQ008_SQ001.[i],
        data5a$matching3N5.SQ009_SQ001.[i], data5a$matching3N5.SQ010_SQ001.[i], data5a$matching3N5.SQ011_SQ001.[i],
        data5a$matching3N5.SQ012_SQ001.[i], data5a$matching3N5.SQ013_SQ001.[i], data5a$matching3N5.SQ014_SQ001.[i],
        data5a$matching3N5.SQ015_SQ001.[i]), cbind(data5a$matching3N5.SQ001_SQ002.[i], data5a$matching3N5.SQ002_SQ002.[i],
        data5a$matching3N5.SQ003_SQ002.[i], data5a$matching3N5.SQ004_SQ002.[i], data5a$matching3N5.SQ005_SQ002.[i],
        data5a$matching3N5.SQ006_SQ002.[i], data5a$matching3N5.SQ007_SQ002.[i], data5a$matching3N5.SQ008_SQ002.[i],
        data5a$matching3N5.SQ009_SQ002.[i], data5a$matching3N5.SQ010_SQ002.[i], data5a$matching3N5.SQ011_SQ002.[i],
        data5a$matching3N5.SQ012_SQ002.[i], data5a$matching3N5.SQ013_SQ002.[i], data5a$matching3N5.SQ014_SQ002.[i],
        data5a$matching3N5.SQ015_SQ002.[i]), cbind(data5a$matching3N5.SQ001_SQ003.[i], data5a$matching3N5.SQ002_SQ003.[i],
        data5a$matching3N5.SQ003_SQ003.[i], data5a$matching3N5.SQ004_SQ003.[i], data5a$matching3N5.SQ005_SQ003.[i],
        data5a$matching3N5.SQ006_SQ003.[i], data5a$matching3N5.SQ007_SQ003.[i], data5a$matching3N5.SQ008_SQ003.[i],
        data5a$matching3N5.SQ009_SQ003.[i], data5a$matching3N5.SQ010_SQ003.[i], data5a$matching3N5.SQ011_SQ003.[i],
        data5a$matching3N5.SQ012_SQ003.[i], data5a$matching3N5.SQ013_SQ003.[i], data5a$matching3N5.SQ014_SQ003.[i],
        data5a$matching3N5.SQ015_SQ003.[i]), cbind(data5a$matching3N5.SQ001_SQ003.[i], data5a$matching3N5.SQ002_SQ003.[i],
        data5a$matching3N5.SQ003_SQ003.[i], data5a$matching3N5.SQ004_SQ003.[i], data5a$matching3N5.SQ005_SQ003.[i],
        data5a$matching3N5.SQ006_SQ003.[i], data5a$matching3N5.SQ007_SQ003.[i], data5a$matching3N5.SQ008_SQ003.[i],
        data5a$matching3N5.SQ009_SQ003.[i], data5a$matching3N5.SQ010_SQ003.[i], data5a$matching3N5.SQ011_SQ003.[i],
        data5a$matching3N5.SQ012_SQ003.[i], data5a$matching3N5.SQ013_SQ003.[i], data5a$matching3N5.SQ014_SQ003.[i],
        data5a$matching3N5.SQ015_SQ003.[i]), cbind(data5a$matching3N5.SQ001_SQ005.[i], data5a$matching3N5.SQ002_SQ005.[i],
        data5a$matching3N5.SQ003_SQ005.[i], data5a$matching3N5.SQ004_SQ005.[i], data5a$matching3N5.SQ005_SQ005.[i],
        data5a$matching3N5.SQ006_SQ005.[i], data5a$matching3N5.SQ007_SQ005.[i], data5a$matching3N5.SQ008_SQ005.[i],
        data5a$matching3N5.SQ009_SQ005.[i], data5a$matching3N5.SQ010_SQ005.[i], data5a$matching3N5.SQ011_SQ005.[i],
        data5a$matching3N5.SQ012_SQ005.[i], data5a$matching3N5.SQ013_SQ005.[i], data5a$matching3N5.SQ014_SQ005.[i],
        data5a$matching3N5.SQ015_SQ005.[i]))
    matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {
    # for ego i
    mL <- matchingList3[[i]]  # get the matching list 2
    ns <- ns4[[i]]  # get the size of egonet4
    if (ns > 0) {
        # if ns=0, no matching was done!
        mm <- as.matrix(mL[[ns]])  # and the corresponding matrix
        matched <- which(mm == 1, arr.ind = T)  # retrieve array indices
        net <- net4[i, ]
        if (length(matched) > 0) {
            # if matching was performed!
            alterL[[i]]$name4[matched[, 2]] <- net[matched[, 1]]
        }
    }
}

# make long df with w3 alters in ego, and indicators for 4 egonets;
df2 <- data.frame(ego = rep(1:length(alterL), each = 20), alter = rep(1:20), cdn = NA, study = NA, bff = NA,
    csn = NA)

for (i in unique(df2$ego)) {
    for (j in 1:20) {
        # for alters nested in ego find out if names denoting alter j appear in the 4 egonets
        alter <- unlist(alterL[[i]][j, -1], use.names = F)  # get names alter j
        alter <- alter[!is.na(alter)]  # exclude NAs
        cdn <- alter %in% net1[i, ]
        study <- alter %in% net2[i, ]
        bff <- alter %in% net3[i, ]
        csn <- alter %in% net4[i, ]
        # and if so, give alter j score 1 on indicators; 0 otherwise
        df2$cdn[which(df2$ego == i & df2$alter == j)] <- ifelse("TRUE" %in% cdn, 1, 0)
        df2$study[which(df2$ego == i & df2$alter == j)] <- ifelse("TRUE" %in% study, 1, 0)
        df2$bff[which(df2$ego == i & df2$alter == j)] <- ifelse("TRUE" %in% bff, 1, 0)
        df2$csn[which(df2$ego == i & df2$alter == j)] <- ifelse("TRUE" %in% csn, 1, 0)
    }
}

# now that i have, for each alter of ego (including duplicates) at t2, the nets to which they
# belong..  i continue with the w1-w2 matching matrices

# we already subsetted the matching matrices (in `w1w2`), but now we take only the rows
# corresponding to w3 egos who have w2 maintained/created alters
w1w2 <- w1w2[which(w1w2$respnr %in% unique(test$respnr[which(test$w3participation == 1)])), ]

for (i in 1:length(alterL)) {

    matchingL <- vector("list", 20)  #pre-allocate empty list of length 20, to store matching matrices
    {
        matchingL[[1]] <- w1w2[i, 1:20]
        matchingL[[2]] <- w1w2[i, 21:40]
        matchingL[[3]] <- w1w2[i, 41:60]
        matchingL[[4]] <- w1w2[i, 61:80]
        matchingL[[5]] <- w1w2[i, 81:100]
        matchingL[[6]] <- w1w2[i, 101:120]
        matchingL[[7]] <- w1w2[i, 121:140]
        matchingL[[8]] <- w1w2[i, 141:160]
        matchingL[[9]] <- w1w2[i, 161:180]
        matchingL[[10]] <- w1w2[i, 181:200]
        matchingL[[11]] <- w1w2[i, 201:220]
        matchingL[[12]] <- w1w2[i, 221:240]
        matchingL[[13]] <- w1w2[i, 241:260]
        matchingL[[14]] <- w1w2[i, 261:280]
        matchingL[[15]] <- w1w2[i, 281:300]
        matchingL[[16]] <- w1w2[i, 301:320]
        matchingL[[17]] <- w1w2[i, 321:340]
        matchingL[[18]] <- w1w2[i, 341:360]
        matchingL[[19]] <- w1w2[i, 361:380]
        matchingL[[20]] <- w1w2[i, 381:400]
    }

    # find the 'right' matching matrix in this list, by taking the one that contains answers
    ind <- NULL
    for (j in seq_along(matchingL)) {
        # check along the sequence of elements j in the matching matrix list if they are non-empty
        # and not NA
        check <- FALSE
        for (col in matchingL[[j]]) {
            if (!all(is.na(col)) && any(nchar(col) > 0)) {
                check <- TRUE
                break
            }
        }
        if (check) {
            ind <- j
            break
        }
    }
    if (length(ind) > 0) {
        # get the matrix
        mm <- matchingL[[ind]]

        # extract alter ids
        ans <- stringr::str_extract_all(mm, "\\(?[0-9,.]+\\)?")
        # here, the object refers to the w2-alter j; and the element indicator in the list refers
        # to the w3-alter

        for (j in unique(test$alterid[which(test$ego == unique(test$ego[which(test$w3participation ==
            1)])[i])])) {
            # for wave-2 alter j

            # to which row/w3-alter was alter j matched?
            match <- which(ans == j)

            # and in which networks did this w3-alter belong?
            nets <- df2[which(df2$ego == i & df2$alter == match[1]), ]

            if (length(match) > 0) {
                # if j was matched to w3-alters...

                # assign to alter j the networks in which the w3-alter appeared
                test$cdn3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid ==
                  j)] <- nets$cdn[1]
                test$study3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] &
                  test$alterid == j)] <- nets$study[1]
                test$bff3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid ==
                  j)] <- nets$bff[1]
                test$csn3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid ==
                  j)] <- nets$csn[1]
            }
        }
    }
}

# also add dynamic relational info on closeness and communication frequency in w3, this was only
# asked for alters from w2 (those maintained from w1 or w2-created)!
test$frequency.t3 <- NA
test$closeness.t3 <- NA

# subset w3 name interpreters on contact freq. of w1 alters (sq021 - sq040!)
# tail(names(data5a),234)
freq <- data5a[, c(1106:1125)]
close <- data5a[, c(1126:1145)]

# recode into numeric values;
freq <- ifelse(freq == "(Bijna) elke dag", 7, ifelse(freq == "1-2 keer per week", 6, ifelse(freq == "Aantal keer per maand",
    5, ifelse(freq == "Ong. 1 keer per maand", 4, ifelse(freq == "Aantal keer per jaar", 3, ifelse(freq ==
        "Ong. 1 keer per jaar", 2, ifelse(freq == "Nooit", 1, NA)))))))
close <- ifelse(close == "Heel erg hecht", 4, ifelse(close == "Hecht", 3, ifelse(close == "Enigszins hecht",
    2, ifelse(close == "Niet hecht", 1, NA))))

for (i in 1:length(alterL)) {
    # for ego i for wave-2 alter j
    for (j in unique(test$alterid[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i])])) {

        test$frequency.t3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid ==
            j)] <- freq[i, ][!is.na(freq[i, ])][j]

        test$closeness.t3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid ==
            j)] <- close[i, ][!is.na(close[i, ])][j]
    }
}

# last, add respnodent's reasons for not renaming a w2-alter. this was only asked for w2 alters who
# were not relisted
test$reason <- NA

# subset reasons given for not renaming a w2-alter tail(names(data5a),243)
reasons <- data5a[, c(1146:1165)]
# unique(reasons$vergeten.SQ001.)  if, according to ego, the relationship with alter changed ('Onze
# relatie is veranderd.'), we probed further
reasons2 <- data5a[, c(1186:1205)]

for (i in 1:length(alterL)) {
    # for ego i

    # if any w2-alter of i went unlisted (thus, scored 0 on `surviveW3`)...

    if (0 %in% test$surviveW3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i])]) {

        # for *non-renamed* alter j
        for (j in unique(test$alterid[which(test$ego == unique(test$ego[which(test$w3participation ==
            1)])[i] & test$surviveW3 == 0)])) {

            # get respondent i's reason for not re-naming j,
            reasonij <- reasons[i, j]
            # if it was due to a change in relationship, probe further
            reasonij <- ifelse(reasonij == "Onze relatie is veranderd.", reasons2[i, j], reasonij)
            # and if the relationship changed in other ways than we offered as choices, just set to
            # 'other'
            reasonij <- ifelse(reasonij == "De relatie is op een andere manier veranderd.", "Andere reden.",
                reasonij)

            test$reason[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid ==
                j)] <- reasonij
        }
    }
}

# recode... unique(test$reason)
test$reason <- ifelse(test$reason == "Andere reden.", "miscellaneous", ifelse(test$reason == "Er is geen gelegenheid geweest voor ons om contact te hebben.",
    "no occasion to get together", ifelse(test$reason == "Eén van ons is verhuisd.", "someone moved",
        ifelse(test$reason == "Eén van ons heeft een belangrijke verandering doorgemaakt  (zoals het stoppen met studeren, het aangaan/beëindigen van een relatie, het krijgen van een kind, enzovoort).",
            "other life event", ifelse(test$reason == "Eén van ons heeft gezondheidsproblemen.", "other life event",
                ifelse(test$reason == "We zijn uit elkaar gegroeid/de relatie is verwaterd.", "drifted apart",
                  ifelse(test$reason == "We hadden meningsverschillen of ruzie.", "disagreement / conflict",
                    ifelse(test$reason == "Ik ben simpelweg vergeten om deze persoon opnieuw te noemen.",
                      "forgotten", test$reason))))))))

data_alters23 <- test

# now 'bind back together'
test <- df_alters[which(df_alters$statusW2 == "Dropped"), ]
df_alters <- bind_rows(test, data_alters23)

# rearrange, by ego, and status
df_alters <- df_alters %>%
    arrange(ego, factor(statusW2, levels = c("Maintained", "Created", "Dropped")))
row.names(df_alters) <- 1:nrow(df_alters)

# new alter ids, 1:n_alter
for (i in unique(df_alters$ego)) {
    df_alters$alterid[which(df_alters$ego == i)] <- 1:length(df_alters$alterid[which(df_alters$ego ==
        i)])
}

# exclude 'names'
df_alters.nn <- df_alters[, -c(5:8)]

# save the resulting dataframe (exclude names)
fsave(df_alters.nn, "tie_maintenance.RDa")

data_nested.Rda

To account for the nesting structure of our data (i.e., ties are nested in dyads and networks, that are both nested in ego), we use a 3-level design. So transform the data accordingly. We do this separately for wave 1 –> wave 2, and wave 2 –> wave 3.

First, tidy up the data-frame (deleting stuff that we won’t need + re-ordering columns)

# in our analyses, we only include non-kin
df <- df_alters.nn[which(df_alters.nn$kin == 0), ]
nrow(df)  #n=4653

# set gender to binary
df$ego_female <- ifelse(df$ego_gender == 2, 1, df$ego_gender)
df$alter_female <- ifelse(df$alter_gender == 2, 1, df$alter_gender)

# drop columns we won't need
names(df)
delete <- c(1, 5, 15, 21:25, 27, 28, 30, 33, 38, 52, 72:88)
df <- df[, -delete]

# reorder
names(df)

order <- c("ego", "ego_female", "ego_age", "ego_educ", "study.year", "romantic", "extraversion", "fin_restr",
    "housing.transition_bin", "occupation.transition_bin", "educ_father", "educ_mother", "cdn.size",
    "cdn.size2", "study.size", "study.size2", "bff.size", "bff.size2", "csn.size", "csn.size2", "cdn.density",
    "cdn.density2", "study.density", "study.density2", "bff.density", "bff.density2", "csn.density",
    "csn.density2", "alterid", "alter_female", "alter_age", "alter_educ", "same_gender", "dif_age", "sim_educ",
    "proximity", "duration", "frequency.t1", "frequency.t2", "frequency.t3", "closeness.t1", "closeness.t2",
    "closeness.t3", "multiplex", "multiplex.t2", "cdn1", "cdn2", "cdn3", "study1", "study2", "study3",
    "bff1", "bff2", "bff3", "csn1", "csn2", "csn3", "cdn_embed.t1", "cdn_embed.t2", "study_embed.t1",
    "study_embed.t2", "bff_embed.t1", "bff_embed.t2", "csn_embed.t1", "csn_embed.t2", "statusW2", "w3participation",
    "reason")

df <- df[, order]

# multiplex t2
df$multiplex.t2 <- rowSums(df[, c("cdn2", "study2", "bff2", "csn2")]) - 1

wave 1 –> wave 2

# in analyzing waves 1 and 2, we look at w1 alters and see whether they are maintained or dropped
# so, drop w2-created alters
df1 <- df[-which(df$statusW2 == "Created"), ]
nrow(df1)  #n=3174

# make seperate datasets for each relational dimension (at t1)
df.cdn <- df1[which(df1$cdn1 == 1), ]
df.study <- df1[which(df1$study1 == 1), ]
df.bff <- df1[which(df1$bff1 == 1), ]
df.csn <- df1[which(df1$csn1 == 1), ]

# rbind the multiple datasets

# but first, a. make new Y (i.e., remain in *same* egonet); make new variables for features
# depending on the particular egonet (eg size, embeddedness)
df.cdn$Y <- ifelse(is.na(df.cdn$cdn2), 0, df.cdn$cdn2)
df.study$Y <- ifelse(is.na(df.study$study2), 0, df.study$study2)
df.bff$Y <- ifelse(is.na(df.bff$bff2), 0, df.bff$bff2)
df.csn$Y <- ifelse(is.na(df.csn$csn2), 0, df.csn$csn2)

df.cdn$tie <- "Confidant"
df.study$tie <- "Study"
df.bff$tie <- "Friend"
df.csn$tie <- "Sport"

df.cdn$density <- df.cdn$cdn.density
df.bff$density <- df.bff$bff.density
df.study$density <- df.study$study.density
df.csn$density <- df.csn$csn.density

df.cdn$size <- df.cdn$cdn.size
df.bff$size <- df.bff$bff.size
df.study$size <- df.study$study.size
df.csn$size <- df.csn$csn.size

df.cdn$embed <- df.cdn$cdn_embed.t1
df.bff$embed <- df.bff$bff_embed.t1
df.study$embed <- df.study$study_embed.t1
df.csn$embed <- df.csn$csn_embed.t1

# also include 'extended' embeddedness, or a the extent to which an alter appearing in the egonet
# of ego is also embedded in additional layers of ego's network. we calculate the sum of
# embeddedness scores in additional networks, where NA = 0.
df.cdn$embed.ext <- rowSums(cbind(ifelse(is.na(df.cdn$bff_embed.t1), 0, df.cdn$bff_embed.t1), ifelse(is.na(df.cdn$study_embed.t1),
    0, df.cdn$study_embed.t1), ifelse(is.na(df.cdn$csn_embed.t1), 0, df.cdn$csn_embed.t1)))
df.bff$embed.ext <- rowSums(cbind(ifelse(is.na(df.bff$cdn_embed.t1), 0, df.bff$cdn_embed.t1), ifelse(is.na(df.bff$study_embed.t1),
    0, df.bff$study_embed.t1), ifelse(is.na(df.bff$csn_embed.t1), 0, df.bff$csn_embed.t1)))
df.study$embed.ext <- rowSums(cbind(ifelse(is.na(df.study$cdn_embed.t1), 0, df.study$cdn_embed.t1), ifelse(is.na(df.study$bff_embed.t1),
    0, df.study$bff_embed.t1), ifelse(is.na(df.study$csn_embed.t1), 0, df.study$csn_embed.t1)))
df.csn$embed.ext <- rowSums(cbind(ifelse(is.na(df.csn$cdn_embed.t1), 0, df.csn$cdn_embed.t1), ifelse(is.na(df.csn$bff_embed.t1),
    0, df.csn$bff_embed.t1), ifelse(is.na(df.csn$study_embed.t1), 0, df.csn$study_embed.t1)))

data <- rbind(df.cdn, df.bff, df.study, df.csn)

# add combination of tie and ego, to make a 'network' variable (i.e., networks are nested in ego,
# and thus unique to ego)
data$network <- paste0(data$tie, data$ego)

# reorder by ego and alter
data <- data[order(data$ego, data$alterid), ]
row.names(data) <- 1:nrow(data)

# set NAs on str. embededness to 0 (for those that were named as the only alter)
data$embed[is.na(data$embed)] <- 0

# keep only columns of interest
names(data)
data <- data[, c(1:12, 29:38, 41, 44, 46, 49, 52, 55, 66, 69:75)]

# make unique alter id
data$alterid <- paste0(data$ego, "X", data$alterid)

# listwise deletion
colSums(is.na(data))  #missings per variable

# table(data$proximity, useNA='always') prop.table(table(data$proximity, useNA='always'))
data$proximity <- ifelse(is.na(data$proximity), "far", data$proximity)  #some don't knows on alter proximity, set to 'far'

data <- data[which(complete.cases(data)), ]

data12 <- data

wave 2 –> wave 3

# in analyzing waves 2 and 3, we look at w2-alters (maintained / created) so, drop w2-dropped and
# only of egos who participated in w3
df2 <- df[-which(df$statusW2 == "Dropped"), ]
df2 <- df2[df2$w3participation == 1, ]

# make seperate datasets for each relational dimension (at t2)
df.cdn <- df2[which(df2$cdn2 == 1), ]
df.study <- df2[which(df2$study2 == 1), ]
df.bff <- df2[which(df2$bff2 == 1), ]
df.csn <- df2[which(df2$csn2 == 1), ]

# rbind the multiple datasets

# but first, a. make new Y (i.e., remain in *same* egonet); make new variables for features
# depending on the particular egonet (eg size, embeddedness)

df.cdn$Y <- ifelse(is.na(df.cdn$cdn3), 0, df.cdn$cdn3)
df.study$Y <- ifelse(is.na(df.study$study3), 0, df.study$study3)
df.bff$Y <- ifelse(is.na(df.bff$bff3), 0, df.bff$bff3)
df.csn$Y <- ifelse(is.na(df.csn$csn3), 0, df.csn$csn3)

df.cdn$tie <- "Confidant"
df.study$tie <- "Study"
df.bff$tie <- "Friend"
df.csn$tie <- "Sport"

df.cdn$density <- df.cdn$cdn.density2
df.bff$density <- df.bff$bff.density2
df.study$density <- df.study$study.density2
df.csn$density <- df.csn$csn.density2

df.cdn$size <- df.cdn$cdn.size2
df.bff$size <- df.bff$bff.size2
df.study$size <- df.study$study.size2
df.csn$size <- df.csn$csn.size2

df.cdn$embed <- df.cdn$cdn_embed.t2
df.bff$embed <- df.bff$bff_embed.t2
df.study$embed <- df.study$study_embed.t2
df.csn$embed <- df.csn$csn_embed.t2

# also include 'extended' embeddedness, or a the extent to which an alter appearing in the egonet
# of ego is also embedded in additional layers of ego's network. we calculate the sum of
# embeddedness scores in additional networks, where NA = 0.
df.cdn$embed.ext <- rowSums(cbind(ifelse(is.na(df.cdn$bff_embed.t2), 0, df.cdn$bff_embed.t2), ifelse(is.na(df.cdn$study_embed.t2),
    0, df.cdn$study_embed.t2), ifelse(is.na(df.cdn$csn_embed.t2), 0, df.cdn$csn_embed.t2)))
df.bff$embed.ext <- rowSums(cbind(ifelse(is.na(df.bff$cdn_embed.t2), 0, df.bff$cdn_embed.t2), ifelse(is.na(df.bff$study_embed.t2),
    0, df.bff$study_embed.t2), ifelse(is.na(df.bff$csn_embed.t2), 0, df.bff$csn_embed.t2)))
df.study$embed.ext <- rowSums(cbind(ifelse(is.na(df.study$cdn_embed.t2), 0, df.study$cdn_embed.t2), ifelse(is.na(df.study$bff_embed.t2),
    0, df.study$bff_embed.t2), ifelse(is.na(df.study$csn_embed.t2), 0, df.study$csn_embed.t2)))
df.csn$embed.ext <- rowSums(cbind(ifelse(is.na(df.csn$cdn_embed.t2), 0, df.csn$cdn_embed.t2), ifelse(is.na(df.csn$bff_embed.t2),
    0, df.csn$bff_embed.t2), ifelse(is.na(df.csn$study_embed.t2), 0, df.csn$study_embed.t2)))

data <- rbind(df.cdn, df.bff, df.study, df.csn)

# add combination of tie and ego, to make a 'network' variable (i.e., networks are nested in ego,
# and thus unique to ego)
data$network <- paste0(data$tie, data$ego)

# reorder by ego and alter
data <- data[order(data$ego, data$alterid), ]
row.names(data) <- 1:nrow(data)

# set NAs on str. embededness to 0 (for those that were named as the only alter)
data$embed[is.na(data$embed)] <- 0

# keep only columns of interest
names(data)

data <- data[, c(1:12, 29:37, 39, 42, 45, 47, 50, 53, 56, 66, 68:75)]
# here include also `statusW2` (created vs maintained) + `reason` for not renaming alter

data$multiplex <- data$multiplex.t2
names(data)
data <- data[, -24]

# make unique alter id
data$alterid <- paste0(data$ego, "X", data$alterid)

# listwise deletion
colSums(is.na(data))  #missings per variable
data$proximity <- ifelse(is.na(data$proximity), "far", data$proximity)

i = unique(data$ego)[1]

# ego covariates among 'created' alters are still empty, so match based on ego id note that
# transitions were not measured between w2w3, so they are NAs
for (i in unique(data$ego)) {

    data$educ_father[data$ego == i & data$statusW2 == "Created"] <- df$educ_father[df$ego == i & (df$statusW2 ==
        "Maintained" | df$statusW2 == "Dropped")][1]
    data$educ_mother[data$ego == i & data$statusW2 == "Created"] <- df$educ_mother[df$ego == i & (df$statusW2 ==
        "Maintained" | df$statusW2 == "Dropped")][1]
    data$study.year[data$ego == i & data$statusW2 == "Created"] <- df$study.year[df$ego == i & (df$statusW2 ==
        "Maintained" | df$statusW2 == "Dropped")][1]
    data$extraversion[data$ego == i & data$statusW2 == "Created"] <- df$extraversion[df$ego == i & (df$statusW2 ==
        "Maintained" | df$statusW2 == "Dropped")][1]
    data$fin_restr[data$ego == i & data$statusW2 == "Created"] <- df$fin_restr[df$ego == i & (df$statusW2 ==
        "Maintained" | df$statusW2 == "Dropped")][1]
    data$romantic[data$ego == i & data$statusW2 == "Created"] <- df$romantic[df$ego == i & (df$statusW2 ==
        "Maintained" | df$statusW2 == "Dropped")][1]
    data$housing.transition_bin[data$ego == i & data$statusW2 == "Created"] <- df$housing.transition_bin[df$ego ==
        i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
    data$occupation.transition_bin[data$ego == i & data$statusW2 == "Created"] <- df$occupation.transition_bin[df$ego ==
        i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
}

# missings on 'reason' are no missings, but refer to alters that were renamed
data$reason[is.na(data$reason)] <- "-"
data <- data[which(complete.cases(data)), ]

# transitions are not measured, set to 0
data$housing.transition <- 0
data$occupation.transition <- 0

data23 <- data

length(unique(data$ego))  #N_ego = 281
length(unique(data$alterid))  #N_alter = 1859

Make some last alterations, list the data-frames for waves 1-2 and 2-3. Also row-bind the data, and add a column indicating when the tie appeared and when it (did not) reappear.

# make colnames similar
colnames(data12)[colnames(data12) == "frequency.t1"] <- "frequency.t"
colnames(data12)[colnames(data12) == "closeness.t1"] <- "closeness.t"

colnames(data23)[colnames(data23) == "frequency.t2"] <- "frequency.t"
colnames(data23)[colnames(data23) == "closeness.t2"] <- "closeness.t"

colnames(data12)[25:28] <- sub("1$", "", colnames(data12)[25:28])
colnames(data23)[24:27] <- sub("2$", "", colnames(data23)[24:27])

data12$housing.transition <- data12$housing.transition_bin
data12$occupation.transition <- data12$occupation.transition_bin

# 2 list data-frames
dataList <- list(data12, data23)
fsave(dataList, "data_nested_List.Rda")

# 3 row-bind
data12$period <- "w1 -> w2"
# data23$period <- ifelse(data23$statusW2 == 'Created', 'w2 -> w3', 'w1 - w2 -> w3')
data23$period <- "w2 -> w3"

data <- bind_rows(data12, data23)

length(unique(data$ego))
length(unique(data$alterid))
nrow(data)

names(data)
data <- data[, -c(9, 10, 29)]

fsave(data, "data_nested.Rda")

References

Franken, Rob, Hidde Bekhuis, and Jochem Tolsma. “The Unstable Social Networks of Students: Where Does Dissimilarity Drive Tie Dissolution?” Sociological Forum.

———. 2023. “The Sports and Friendships Study 2021-2023 (Data File).” DANS Data Station Social Sciences and Humanities. https://doi.org/10.17026/SS/GODKDR.

---
title: "Data preparation"
bibliography: references.bib
link-citations: true
date: "Last compiled on `r format(Sys.time(), '%B, %Y')`"
output: 
  html_document:
    css: tweaks.css
    toc:  true
    toc_float: true
    number_sections: false
    toc_depth: 2
    code_folding: show
    code_download: yes
---


```{r, globalsettings, echo=FALSE, warning=FALSE, results='hide', message=FALSE}
library(knitr)
library(tidyverse)
knitr::opts_chunk$set(echo = TRUE)
opts_chunk$set(tidy.opts=list(width.cutoff=100),tidy=TRUE, warning = FALSE, message = FALSE,comment = "#>", cache=TRUE, class.source=c("test"), class.output=c("test3"))
options(width = 100)
rgl::setupKnitr()

colorize <- function(x, color) {sprintf("<span style='color: %s;'>%s</span>", color, x) }
```


```{r klippy, echo=FALSE, include=TRUE}
klippy::klippy(position = c('top', 'right'))
#klippy::klippy(color = 'darkred')
#klippy::klippy(tooltip_message = 'Click to copy', tooltip_success = 'Done')
```


---  

The following scripts can be used to replicate the data-set of @Franken2022. It may also be obtained by downloading: `r xfun::embed_file("./data_shared/data_nested.RDa")`



----

<br>

# Getting started

To copy the code, click the button in the upper right corner of the code-chunks.

## clean up

```{r, results='hide'}
rm(list=ls())
gc()
```

<br>

## general custom functions

- `fpackage.check`: Check if packages are installed (and install if not) in R
- `fsave`: Function to save data with time stamp in correct directory
- `fload`: Load R-objects under new names
- `fshowdf`: Print objects (`tibble` / `data.frame`) nicely on screen in `.Rmd`.

```{r, eval=FALSE}
fpackage.check <- function(packages) {
    lapply(packages, FUN = function(x) {
        if (!require(x, character.only = TRUE)) {
            install.packages(x, dependencies = TRUE)
            library(x, character.only = TRUE)
        }
    })
}

fsave <- function(x, file, location = "./data/processed/", ...) {
    if (!dir.exists(location))
        dir.create(location)
    datename <- substr(gsub("[:-]", "", Sys.time()), 1, 8)
    totalname <- paste(location, datename, file, sep = "")
    print(paste("SAVED: ", totalname, sep = ""))
    save(x, file = totalname)
}


fload  <- function(fileName){
  load(fileName)
  get(ls()[ls() != "fileName"])
}

fshowdf <- function(x, digits = 2, ...) {
    knitr::kable(x, digits = digits, "html", ...) %>%
        kableExtra::kable_styling(bootstrap_options = c("striped", "hover")) %>%
        kableExtra::scroll_box(width = "100%", height = "300px")
}
```

<br>

## necessary packages

- `dplyr`: package for data wrangling
- `lubridate`: parse and manipulate dates
- `stringr`: string manipulation

```{r, eval=FALSE}
packages = c("dplyr", "lubridate", "stringr")
fpackage.check(packages)
rm(packages)
```

<br>


<script>
function myFunction() {

            var btn = document.getElementById("myButton");
            //to make it fancier
            if (btn.value == "Click to hide code") {
                btn.value = "Show code";
                btn.innerHTML = "Show code";
            }
            else {
                btn.value = "Click to hide code";
                btn.innerHTML = "Hide code";
            }
            //this is what you're looking for
            var x = document.getElementById("myDIV");
            if (x.style.display === "none") {
                x.style.display = "block";
            } else {
                x.style.display = "none";
            }
        }
          
</script>


# Download data

Anonymized data-sets of the 'Sports and Friendships' [@data2022] study are deposited in DANS EASY. For this study we use waves 3-5 (Cohort II):

- `r xfun::embed_file("./data/wave3_public.RDa")`
- `r xfun::embed_file("./data/wave4_public.RDa")`
- `r xfun::embed_file("./data/wave5_public.RDa")`

Download the data-files, and put them in the `./data/` folder. But first, make a `./data/` folder: 

```{r, eval=F}
ifelse(!dir.exists("data"), dir.create("data"), FALSE)
``` 

<br>

To see the code used to anonymize the raw data files, click the button below.

<button class=button1 onclick="myFunction()" id="myButton" value="Click To Open Instructions">Show code: from raw to public data</button>

<div style="display:none;" id="myDIV">
<br>


### load in raw data

```{r eval=FALSE}
data1 <- read.csv("./rawdata/results-survey938243.csv") #wave1
data2 <- read.csv("./rawdata/results-survey868275.csv") #wave2
data3 <- read.csv("./rawdata/results-survey493144.csv") #wave3
data4a <- read.csv("./rawdata/results-survey952354a.csv") #wave4; the data was exported in the collection period of this wave, to restore a bug in limesurvey.
data4b <- read.csv("./rawdata/results-survey952354b.csv")
data5 <- read.csv("./rawdata/results-survey356557.csv") #wave 5
```

<br>

### make public data-sets to deposit {.tabset .tabset-fade} 

The following scripts were used to delete sensitive information from the respective survey rounds (e.g., information that could identify the respondent, names of alters, etc.) and to make public data-sets for each wave.

#### Cohort I: wave 1
```{r, eval=F}
#copy dataframe
df <- data1

# first, delete fake responses
fakeid <- c(337, 575, 576) #id of fake responses

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=170

#make new id variable
df$respnr <- 1:nrow(df)

#add 10000, sygnifying the cohort
df$respnr <- df$respnr + 10000

#remove first 10 columns (e.g,. token, submitdate, refurl, etc.)
#i dont need these
head(colnames(df),25)
df <- df[,-c(1:10)]
#now, column 1 is the first question of the survey

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
study <- 6 # study program
occupation <- 37
residency <- c(153,154,157,158) #residence (municipality name; 4digit postal code)
exclude_colnumber <- c(study,occupation,residency)
#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]

#i also need to deal with the name-generator data;
#i replace all listed names with a unique, anonymous alter-id,

#first, get all column numbers corresponding to names
net1 <- c(10:14)
net2 <- c(50:54)
net3 <- c(166:170)
net4 <- c(440:444)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, we delete columns that are irrelevant
df<-df[,-c(1:4)]

#and the last columns (excluding the very last, which denotes `respnr`)
#but first, we subset the respnr-email combination, so that we can merge with wave2
merge <- df[,c("respnr", "email")]
excl <- c(869:length(df))
excl <- excl[-length(excl)]
df <- df[, -excl]

#fix(df)#check if everything went alright
row.names(df) <- 1:nrow(df)

#save the dataframe to the processed data-folder
fsave(df, "wave1_public.RDa")
``` 

#### Cohort I: wave 2

```{r, eval=F}
#copy dataframe
df <- data2

# first, delete fake response
fakeid <- c(61) #id of fake response

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=93

#attach respnr
df <- merge(df,merge,by="email")

#reorder by respr
df <- df[order(df$respnr),]
rownames(df) <- 1:nrow(df)

#delete first columns
df <- df[,-c(1:12)]

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
#in wave 2, only residency variables
exclude_colnumber <- c(130,131)
#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]

#again, replace named persons with unique alter_id p_i
net1 <- c(1:5)
net2 <- c(26:30)
net3 <- c(142:146)
net4 <- c(336:340)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, remove all irrelevant variables,
#from "interviewtime" on.
#excluding the last column, as this is respnr
#which( colnames(df)=="interviewtime" )
excl <- c( which( colnames(df)=="interviewtime" ):length(df))
excl <- excl[-length(excl)]
df <- df[, -excl]

fsave(df, "wave2_public.RDa")
``` 
 
#### Cohort II: wave 1


```{r, eval=F}
#copy dataframe
df <- data3

# first, delete fake respons;
fakeid <- which(df$token  %in% "test")

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=1135

#make new id variable
#add 20000, so we have a signifier for the 2nd cohort
df$respnr <- 20000 + 1:nrow(df)

#remove first columns (e.g,. token, refurl, etc.)
#i dont need these
#but do keep submit and startdate
head(colnames(df),25)
df <- df[,-c(1,3:6,8:10)]

#now, column 3 is the first question of the survey

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
study <- 6 # study program
occupation <- 37 #other category 
residency <- c(153,154,157,158) #residence (municipality name; 4digit postal code)
exclude_colnumber <- c(study,occupation,residency)

#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]
#i also need to deal with the name-generator data;
#i replace all listed names with a unique, anonymous alter-id,

#first, get all column numbers corresponding to names
net1 <- c(10:14)
net2 <- c(50:54)
net3 <- c(166:170)
net4 <- c(441:445)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, we delete columns that are irrelevant
df<-df[,-c(3,4)]

#we also delete a lot of columns (excluding the very last, which denotes `respnr`; and variables indicating group times)

#but first, we subset the respnr-email combination, so that we can merge with wave2
merge <- df[,c("respnr", "email")]

fix(merge)

#fsave(merge, file = "sleutelbestand_respnr_email.RDa")

excl <- c(995:length(df))
excl <- excl[-length(excl)] #minus last columns (respnr)

#also keep group times
excl <- excl[ ! excl %in% which(grepl("groupTime", names(df)))]

df <- df[, -excl]

#fix(df)#check if everything went alright

#save the dataframe to the processed data-folder
fsave(df, "wave3_public.RDa")
```

#### Cohort II: wave 2

While conducting this survey round, I encountered an issue with a matching matrix not loading (`matching1N1.SQ001_SQ001` did not load; see the <a href="https://netchange.netlify.app/codebook.pdf" target="_blank">codebook</a> for more info). As a result, the survey was shut down temporarily to fix the bug, and then reactivated, thus resulting in two distinct data-files. 

Before making the public data-set for this wave, I (a) delete participants that had to be presented `matching1N1.SQ001_SQ001` (i.e., those who named at least one CDN and 1 study partner), and (b) merge the data-files.


```{r eval=FALSE}
# participants that named one or more CDN alters, AND 1 study partner, were not presented with a matching matrix. 
length(which(data4a$egonet1.SQ001.!="" & data4a$egonet2.SQ001.!="" & data4a$egonet2.SQ002.=="")) # 88 particpants. 

#i delete those...
data4a. <- data4a[-which(data4a$egonet1.SQ001.!="" & data4a$egonet2.SQ001.!="" & data4a$egonet2.SQ002.==""),]
data4a.$matching1N1.SQ001_SQ001. <- NA

#in order to merge the files, make sure the have the same columns
data4a.. <- data4a.[,which(names(data4a.) %in% names(data4b))]

#rowbind the separate data-files
data4 <- rbind(data4a.., data4b)
nrow(data4) #n=520
```

```{r, eval=F}
#copy dataframe
df <- data4

# first, delete fake/test response
fakeid <- c(3,4) #id of fake response

#remove rows where column id is equal to fakeid
df <- subset(df, !(id %in% fakeid))
nrow(df) #N=518

#attach respnr
df <- merge(df,merge,by="email")

#reorder by respr
df <- df[order(df$respnr),]
rownames(df) <- 1:nrow(df)

#delete first columns
#but keep start/submit date
#head(names(df),12)

df <- df[,-c(1,2,4:7,9:12)]

#i create a vector containing column numbers of variables that need to be excluded
#i.e., (string) variables that can identify a respondent
#study: 129
#residency variables: 138, 139

exclude_colnumber <- c(129, 138,139)
#exclude variables corresponding to these column numbers
df <- df[,-exclude_colnumber]

#again, replace named persons with unique alter_id p_i
net1 <- c(3:7)
net2 <- c(28:32)
net3 <- c(151:155)
net4 <- c(396:400)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#last, remove all irrelevant variables,
#excluding the last column, as this is respnr
excl <- 1356:length(df)
excl <- excl[-length(excl)] #minus last columns (respnr)

#and time records for question groups
excl <- excl[ ! excl %in% which(grepl("groupTime", names(df)))]

df <- df[, -excl]

fsave(df, "wave4_public.RDa")
``` 
 
#### Cohort II: wave 3
```{r, eval=F}
#copy dataframe
df <- data5

#attach respnr
df <- merge(df,merge,by="email")

#reorder by respr
df <- df[order(df$respnr),]
rownames(df) <- 1:nrow(df)

#delete first columns
#but keep start/submit date
#head(names(df),12)
df <- df[,-c(1,2,4:7,9:12)]

#again, replace named persons with unique alter_id p_i
net1 <- c(3:7)
net2 <- c(8:12)
net3 <- c(98:102)
net4 <- c(253:257)
names_colnumber <- c(net1,net2,net3,net4)

for ( i in 1:length(unique(names_colnumber))) { # for (potentially listed) alters 1:20
  # get the index, corresponding to the column number where particular alter was listed
  ind <- names_colnumber[i]
  # if the corresponding entry in the name-list is NOT empty
  df[,ind] <- ifelse(!df[,ind] == "", 
                       # replace it by "p" (for person), followed by the increment i,
                     paste0("p", i), df[,ind])
}

#also, anonymize the 'other' reasons given for not renaming an alter.
open <- c(1166:1185,1206:1225)
df[,open][df[,open]==""] <- NA
df[,open][!is.na(df[,open])] <- 1 #just put "1" instead...

#last, remove all irrelevant variables,
#excluding the last column, as this is respnr

#tail(names(df),308)


excl <- 1226:length(df)
excl <- excl[-length(excl)] #minus last columns (respnr)

#and time records for question groups
excl <- excl[ ! excl %in% which(grepl("groupTime", names(df)))]

df <- df[, -excl]
#fix(df)

fsave(df, "wave5_public.RDa")
```


### {-}


</div>

----

<br>

# Import data

Load the downloaded data. First, we clean our environment (but we keep our functions; we need them later on).

```{r,eval=F}
#clean environment, but keep functions
rm(list = setdiff(ls(), lsf.str()))

#load public data

#Cohort II
data3 <- fload("./data/wave3_public.RDa")
data4 <- fload("./data/wave4_public.RDa")
data5 <- fload("./data/wave5_public.RDa")
```


<br>

----


# tie_maintenance.RDa 

In the following script, I construct a long data-frame (based on Cohort II, waves 1-2-3), with alters nested in egos, to investigate which alters are maintained in the personal networks of ego over the academic year. To each ego, relevant individual-level, contextual-level, and network-level attributes are assigned; to each alter, relevant alter and dyadic features are assigned. The data-set is saved.

## wave 1 --> wave 2 (maintained vs dropped)

```{r, eval=FALSE}
# subset respondents that filled out both surveys w1 and w2;
# by matching on `respnr`
data <- data3[which(data3$respnr %in% unique(data4$respnr)),]
row.names(data) <- 1:nrow(data)
nrow(data) #N=516

# make a dataframe of w1-alters, with potential duplicates...
# recall that all alter names were replaced with a unique anonymized alter-id
df_names <- data.frame(
  p1 = data$egonet1.SQ001.,
  p2 = data$egonet1.SQ002.,
  p3 = data$egonet1.SQ003.,
  p4 = data$egonet1.SQ004.,
  p5 = data$egonet1.SQ005.,
  p6 = data$egonet2.SQ001.,
  p7 = data$egonet2.SQ002.,
  p8 = data$egonet2.SQ003.,
  p9 = data$egonet2.SQ004.,
  p10= data$egonet2.SQ005.,
  p11= data$egonet3.SQ001.,
  p12= data$egonet3.SQ002.,
  p13= data$egonet3.SQ003.,
  p14= data$egonet3.SQ004.,
  p15= data$egonet3.SQ005.,
  p16= data$egonet4.SQ001.,
  p17= data$egonet4.SQ002.,
  p18= data$egonet4.SQ003.,
  p19= data$egonet4.SQ004.,
  p20= data$egonet4.SQ005.)

#"pre-allocate" empty list of length equals number of egos
alterL <- vector("list", nrow(df_names))

# loop over all egos
for ( i in 1:length(alterL)) {
    alterL[[i]] <- data.frame(
      ego_gender = NA, ego_age = NA, ego_educ = NA,
      alterid = 1:20, name1 = NA, name2 = NA, name3 = NA, name4 = NA,
      alter_gender = NA, alter_age = NA, alter_educ=NA,
      same_gender = NA, dif_age = NA, sim_educ = NA,
      cdn_embed.t1 = NA, study_embed.t1 = NA, bff_embed.t1 = NA, csn_embed.t1 = NA)
}

# fill the names based on the names data-frame and replace empty strings with <NA>
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=FALSE)
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
}

# a matching matrix allowed ego to match names that referred to the same person.
# in limesurvey, i could not let the number of columns condition
# on the number of alters named in the latest egonet, so i made 5 separate matrices,
# conditional on netsize.
# i calculate netsize for each net
{
  net1 <- cbind(data$egonet1.SQ001.,data$egonet1.SQ002., data$egonet1.SQ003.,data$egonet1.SQ004., data$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data$egonet2.SQ001.,data$egonet2.SQ002., data$egonet2.SQ003.,data$egonet2.SQ004., data$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data$egonet3.SQ001.,data$egonet3.SQ002., data$egonet3.SQ003.,data$egonet3.SQ004., data$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data$egonet4.SQ001.,data$egonet4.SQ002., data$egonet4.SQ003.,data$egonet4.SQ004., data$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}

# i construct the matching matrices, 1 - 5 (conditional on the ns of the latest net)
# put these in a list and list these again, for each respondent (so a list of lists...)
matchingList <- list()
for (i in 1:length(alterL)) { # for ego i
  matchingL <- list()
  # make 5 seperate matching matrices. naturally, only 1 is relevant... but i will select that later.
  matchingL[[1]] <- cbind(data$matching1N1.SQ001_SQ001.[i],data$matching1N1.SQ002_SQ001.[i],data$matching1N1.SQ003_SQ001.[i],data$matching1N1.SQ004_SQ001.[i],data$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data$matching1N2.SQ001_SQ001.[i], data$matching1N2.SQ002_SQ001.[i], data$matching1N2.SQ003_SQ001.[i], data$matching1N2.SQ004_SQ001.[i], data$matching1N2.SQ005_SQ001.[i]),
    cbind(data$matching1N2.SQ001_SQ002.[i], data$matching1N2.SQ002_SQ002.[i], data$matching1N2.SQ003_SQ002.[i], data$matching1N2.SQ004_SQ002.[i], data$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data$matching1N3.SQ001_SQ001.[i], data$matching1N3.SQ002_SQ001.[i], data$matching1N3.SQ003_SQ001.[i], data$matching1N3.SQ004_SQ001.[i], data$matching1N3.SQ005_SQ001.[i]),
    cbind(data$matching1N3.SQ001_SQ002.[i], data$matching1N3.SQ002_SQ002.[i], data$matching1N3.SQ003_SQ002.[i], data$matching1N3.SQ004_SQ002.[i], data$matching1N3.SQ005_SQ002.[i]),
    cbind(data$matching1N3.SQ001_SQ003.[i], data$matching1N3.SQ002_SQ003.[i], data$matching1N3.SQ003_SQ003.[i], data$matching1N3.SQ004_SQ003.[i], data$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data$matching1N4.SQ001_SQ001.[i], data$matching1N4.SQ002_SQ001.[i], data$matching1N4.SQ003_SQ001.[i], data$matching1N4.SQ004_SQ001.[i], data$matching1N4.SQ005_SQ001.[i]),
    cbind(data$matching1N4.SQ001_SQ002.[i], data$matching1N4.SQ002_SQ002.[i], data$matching1N4.SQ003_SQ002.[i], data$matching1N4.SQ004_SQ002.[i], data$matching1N4.SQ005_SQ002.[i]),
    cbind(data$matching1N4.SQ001_SQ003.[i], data$matching1N4.SQ002_SQ003.[i], data$matching1N4.SQ003_SQ003.[i], data$matching1N4.SQ004_SQ003.[i], data$matching1N4.SQ005_SQ003.[i]),
    cbind(data$matching1N4.SQ001_SQ004.[i], data$matching1N4.SQ002_SQ004.[i], data$matching1N4.SQ003_SQ004.[i], data$matching1N4.SQ004_SQ004.[i], data$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data$matching1N5.SQ001_SQ001.[i], data$matching1N5.SQ002_SQ001.[i], data$matching1N5.SQ003_SQ001.[i], data$matching1N5.SQ004_SQ001.[i], data$matching1N5.SQ005_SQ001.[i]),
    cbind(data$matching1N5.SQ001_SQ002.[i], data$matching1N5.SQ002_SQ002.[i], data$matching1N5.SQ003_SQ002.[i], data$matching1N5.SQ004_SQ002.[i], data$matching1N5.SQ005_SQ002.[i]),
    cbind(data$matching1N5.SQ001_SQ003.[i], data$matching1N5.SQ002_SQ003.[i], data$matching1N5.SQ003_SQ003.[i], data$matching1N5.SQ004_SQ003.[i], data$matching1N5.SQ005_SQ003.[i]),
    cbind(data$matching1N5.SQ001_SQ004.[i], data$matching1N5.SQ002_SQ004.[i], data$matching1N5.SQ003_SQ004.[i], data$matching1N5.SQ004_SQ004.[i], data$matching1N5.SQ005_SQ004.[i]),
    cbind(data$matching1N5.SQ001_SQ005.[i], data$matching1N5.SQ002_SQ005.[i], data$matching1N5.SQ003_SQ005.[i], data$matching1N5.SQ004_SQ005.[i], data$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
} # so... matchingL[[1]][[5]] is matchingmatrix 5 (i.e., 5 alters named in egonet2) for ego 1

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
} #ignore warning

#repeat for the second matching matrices
# (i.e., matching egonet3 alters to prev. alters);
matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data$matching2N1.SQ001_SQ001.[i],data$matching2N1.SQ002_SQ001.[i],data$matching2N1.SQ003_SQ001.[i],data$matching2N1.SQ004_SQ001.[i],data$matching2N1.SQ005_SQ001.[i],data$matching2N1.SQ006_SQ001.[i],data$matching2N1.SQ007_SQ001.[i],data$matching2N1.SQ008_SQ001.[i],data$matching2N1.SQ009_SQ001.[i],data$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data$matching2N2.SQ001_SQ001.[i],data$matching2N2.SQ002_SQ001.[i],data$matching2N2.SQ003_SQ001.[i],data$matching2N2.SQ004_SQ001.[i],data$matching2N2.SQ005_SQ001.[i],data$matching2N2.SQ006_SQ001.[i],data$matching2N2.SQ007_SQ001.[i],data$matching2N2.SQ008_SQ001.[i],data$matching2N2.SQ009_SQ001.[i],data$matching2N2.SQ010_SQ001.[i]),
    cbind(data$matching2N2.SQ001_SQ002.[i],data$matching2N2.SQ002_SQ002.[i],data$matching2N2.SQ003_SQ002.[i],data$matching2N2.SQ004_SQ002.[i],data$matching2N2.SQ005_SQ002.[i],data$matching2N2.SQ006_SQ002.[i],data$matching2N2.SQ007_SQ002.[i],data$matching2N2.SQ008_SQ002.[i],data$matching2N2.SQ009_SQ002.[i],data$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data$matching2N3.SQ001_SQ001.[i],data$matching2N3.SQ002_SQ001.[i],data$matching2N3.SQ003_SQ001.[i],data$matching2N3.SQ004_SQ001.[i],data$matching2N3.SQ005_SQ001.[i],data$matching2N3.SQ006_SQ001.[i],data$matching2N3.SQ007_SQ001.[i],data$matching2N3.SQ008_SQ001.[i],data$matching2N3.SQ009_SQ001.[i],data$matching2N3.SQ010_SQ001.[i]),
    cbind(data$matching2N3.SQ001_SQ002.[i],data$matching2N3.SQ002_SQ002.[i],data$matching2N3.SQ003_SQ002.[i],data$matching2N3.SQ004_SQ002.[i],data$matching2N3.SQ005_SQ002.[i],data$matching2N3.SQ006_SQ002.[i],data$matching2N3.SQ007_SQ002.[i],data$matching2N3.SQ008_SQ002.[i],data$matching2N3.SQ009_SQ002.[i],data$matching2N3.SQ010_SQ002.[i]),
    cbind(data$matching2N3.SQ001_SQ003.[i],data$matching2N3.SQ002_SQ003.[i],data$matching2N3.SQ003_SQ003.[i],data$matching2N3.SQ004_SQ003.[i],data$matching2N3.SQ005_SQ003.[i],data$matching2N3.SQ006_SQ003.[i],data$matching2N3.SQ007_SQ003.[i],data$matching2N3.SQ008_SQ003.[i],data$matching2N3.SQ009_SQ003.[i],data$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data$matching2N4.SQ001_SQ001.[i],data$matching2N4.SQ002_SQ001.[i],data$matching2N4.SQ003_SQ001.[i],data$matching2N4.SQ004_SQ001.[i],data$matching2N4.SQ005_SQ001.[i],data$matching2N4.SQ006_SQ001.[i],data$matching2N4.SQ007_SQ001.[i],data$matching2N4.SQ008_SQ001.[i],data$matching2N4.SQ009_SQ001.[i],data$matching2N4.SQ010_SQ001.[i]),
    cbind(data$matching2N4.SQ001_SQ002.[i],data$matching2N4.SQ002_SQ002.[i],data$matching2N4.SQ003_SQ002.[i],data$matching2N4.SQ004_SQ002.[i],data$matching2N4.SQ005_SQ002.[i],data$matching2N4.SQ006_SQ002.[i],data$matching2N4.SQ007_SQ002.[i],data$matching2N4.SQ008_SQ002.[i],data$matching2N4.SQ009_SQ002.[i],data$matching2N4.SQ010_SQ002.[i]),
    cbind(data$matching2N4.SQ001_SQ003.[i],data$matching2N4.SQ002_SQ003.[i],data$matching2N4.SQ003_SQ003.[i],data$matching2N4.SQ004_SQ003.[i],data$matching2N4.SQ005_SQ003.[i],data$matching2N4.SQ006_SQ003.[i],data$matching2N4.SQ007_SQ003.[i],data$matching2N4.SQ008_SQ003.[i],data$matching2N4.SQ009_SQ003.[i],data$matching2N4.SQ010_SQ003.[i]),
    cbind(data$matching2N4.SQ001_SQ004.[i],data$matching2N4.SQ002_SQ004.[i],data$matching2N4.SQ003_SQ004.[i],data$matching2N4.SQ004_SQ004.[i],data$matching2N4.SQ005_SQ004.[i],data$matching2N4.SQ006_SQ004.[i],data$matching2N4.SQ007_SQ004.[i],data$matching2N4.SQ008_SQ004.[i],data$matching2N4.SQ009_SQ004.[i],data$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data$matching2N5.SQ001_SQ001.[i],data$matching2N5.SQ002_SQ001.[i],data$matching2N5.SQ003_SQ001.[i],data$matching2N5.SQ004_SQ001.[i],data$matching2N5.SQ005_SQ001.[i],data$matching2N5.SQ006_SQ001.[i],data$matching2N5.SQ007_SQ001.[i],data$matching2N5.SQ008_SQ001.[i],data$matching2N5.SQ009_SQ001.[i],data$matching2N5.SQ010_SQ001.[i]),
    cbind(data$matching2N5.SQ001_SQ002.[i],data$matching2N5.SQ002_SQ002.[i],data$matching2N5.SQ003_SQ002.[i],data$matching2N5.SQ004_SQ002.[i],data$matching2N5.SQ005_SQ002.[i],data$matching2N5.SQ006_SQ002.[i],data$matching2N5.SQ007_SQ002.[i],data$matching2N5.SQ008_SQ002.[i],data$matching2N5.SQ009_SQ002.[i],data$matching2N5.SQ010_SQ002.[i]),
    cbind(data$matching2N5.SQ001_SQ003.[i],data$matching2N5.SQ002_SQ003.[i],data$matching2N5.SQ003_SQ003.[i],data$matching2N5.SQ004_SQ003.[i],data$matching2N5.SQ005_SQ003.[i],data$matching2N5.SQ006_SQ003.[i],data$matching2N5.SQ007_SQ003.[i],data$matching2N5.SQ008_SQ003.[i],data$matching2N5.SQ009_SQ003.[i],data$matching2N5.SQ010_SQ003.[i]),
    cbind(data$matching2N5.SQ001_SQ004.[i],data$matching2N5.SQ002_SQ004.[i],data$matching2N5.SQ003_SQ004.[i],data$matching2N5.SQ004_SQ004.[i],data$matching2N5.SQ005_SQ004.[i],data$matching2N5.SQ006_SQ004.[i],data$matching2N5.SQ007_SQ004.[i],data$matching2N5.SQ008_SQ004.[i],data$matching2N5.SQ009_SQ004.[i],data$matching2N5.SQ010_SQ004.[i]),
    cbind(data$matching2N5.SQ001_SQ005.[i],data$matching2N5.SQ002_SQ005.[i],data$matching2N5.SQ003_SQ005.[i],data$matching2N5.SQ004_SQ005.[i],data$matching2N5.SQ005_SQ005.[i],data$matching2N5.SQ006_SQ005.[i],data$matching2N5.SQ007_SQ005.[i],data$matching2N5.SQ008_SQ005.[i],data$matching2N5.SQ009_SQ005.[i],data$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

#and for matching matrix 3 (i.e., matching egonet4 with egonets 1-3)
matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data$matching3N1.SQ001_SQ001.[i],data$matching3N1.SQ002_SQ001.[i],data$matching3N1.SQ003_SQ001.[i],data$matching3N1.SQ004_SQ001.[i],data$matching3N1.SQ005_SQ001.[i],data$matching3N1.SQ006_SQ001.[i],data$matching3N1.SQ007_SQ001.[i],data$matching3N1.SQ008_SQ001.[i],data$matching3N1.SQ009_SQ001.[i],data$matching3N1.SQ010_SQ001.[i], data$matching3N1.SQ011_SQ001.[i], data$matching3N1.SQ012_SQ001.[i], data$matching3N1.SQ013_SQ001.[i], data$matching3N1.SQ014_SQ001.[i], data$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data$matching3N2.SQ001_SQ001.[i],data$matching3N2.SQ002_SQ001.[i],data$matching3N2.SQ003_SQ001.[i],data$matching3N2.SQ004_SQ001.[i],data$matching3N2.SQ005_SQ001.[i],data$matching3N2.SQ006_SQ001.[i],data$matching3N2.SQ007_SQ001.[i],data$matching3N2.SQ008_SQ001.[i],data$matching3N2.SQ009_SQ001.[i],data$matching3N2.SQ010_SQ001.[i], data$matching3N2.SQ011_SQ001.[i], data$matching3N2.SQ012_SQ001.[i], data$matching3N2.SQ013_SQ001.[i], data$matching3N2.SQ014_SQ001.[i], data$matching3N2.SQ015_SQ001.[i]),
    cbind(data$matching3N2.SQ001_SQ002.[i],data$matching3N2.SQ002_SQ002.[i],data$matching3N2.SQ003_SQ002.[i],data$matching3N2.SQ004_SQ002.[i],data$matching3N2.SQ005_SQ002.[i],data$matching3N2.SQ006_SQ002.[i],data$matching3N2.SQ007_SQ002.[i],data$matching3N2.SQ008_SQ002.[i],data$matching3N2.SQ009_SQ002.[i],data$matching3N2.SQ010_SQ002.[i], data$matching3N2.SQ011_SQ002.[i], data$matching3N2.SQ012_SQ002.[i], data$matching3N2.SQ013_SQ002.[i], data$matching3N2.SQ014_SQ002.[i], data$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data$matching3N3.SQ001_SQ001.[i],data$matching3N3.SQ002_SQ001.[i],data$matching3N3.SQ003_SQ001.[i],data$matching3N3.SQ004_SQ001.[i],data$matching3N3.SQ005_SQ001.[i],data$matching3N3.SQ006_SQ001.[i],data$matching3N3.SQ007_SQ001.[i],data$matching3N3.SQ008_SQ001.[i],data$matching3N3.SQ009_SQ001.[i],data$matching3N3.SQ010_SQ001.[i], data$matching3N3.SQ011_SQ001.[i], data$matching3N3.SQ012_SQ001.[i], data$matching3N3.SQ013_SQ001.[i], data$matching3N3.SQ014_SQ001.[i], data$matching3N3.SQ015_SQ001.[i]),
    cbind(data$matching3N3.SQ001_SQ002.[i],data$matching3N3.SQ002_SQ002.[i],data$matching3N3.SQ003_SQ002.[i],data$matching3N3.SQ004_SQ002.[i],data$matching3N3.SQ005_SQ002.[i],data$matching3N3.SQ006_SQ002.[i],data$matching3N3.SQ007_SQ002.[i],data$matching3N3.SQ008_SQ002.[i],data$matching3N3.SQ009_SQ002.[i],data$matching3N3.SQ010_SQ002.[i], data$matching3N3.SQ011_SQ002.[i], data$matching3N3.SQ012_SQ002.[i], data$matching3N3.SQ013_SQ002.[i], data$matching3N3.SQ014_SQ002.[i], data$matching3N3.SQ015_SQ002.[i]),
    cbind(data$matching3N3.SQ001_SQ003.[i],data$matching3N3.SQ002_SQ003.[i],data$matching3N3.SQ003_SQ003.[i],data$matching3N3.SQ004_SQ003.[i],data$matching3N3.SQ005_SQ003.[i],data$matching3N3.SQ006_SQ003.[i],data$matching3N3.SQ007_SQ003.[i],data$matching3N3.SQ008_SQ003.[i],data$matching3N3.SQ009_SQ003.[i],data$matching3N3.SQ010_SQ003.[i], data$matching3N3.SQ011_SQ003.[i], data$matching3N3.SQ012_SQ003.[i], data$matching3N3.SQ013_SQ003.[i], data$matching3N3.SQ014_SQ003.[i], data$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data$matching3N4.SQ001_SQ001.[i],data$matching3N4.SQ002_SQ001.[i],data$matching3N4.SQ003_SQ001.[i],data$matching3N4.SQ004_SQ001.[i],data$matching3N4.SQ005_SQ001.[i],data$matching3N4.SQ006_SQ001.[i],data$matching3N4.SQ007_SQ001.[i],data$matching3N4.SQ008_SQ001.[i],data$matching3N4.SQ009_SQ001.[i],data$matching3N4.SQ010_SQ001.[i], data$matching3N4.SQ011_SQ001.[i], data$matching3N4.SQ012_SQ001.[i], data$matching3N4.SQ013_SQ001.[i], data$matching3N4.SQ014_SQ001.[i], data$matching3N4.SQ015_SQ001.[i]),
    cbind(data$matching3N4.SQ001_SQ002.[i],data$matching3N4.SQ002_SQ002.[i],data$matching3N4.SQ003_SQ002.[i],data$matching3N4.SQ004_SQ002.[i],data$matching3N4.SQ005_SQ002.[i],data$matching3N4.SQ006_SQ002.[i],data$matching3N4.SQ007_SQ002.[i],data$matching3N4.SQ008_SQ002.[i],data$matching3N4.SQ009_SQ002.[i],data$matching3N4.SQ010_SQ002.[i], data$matching3N4.SQ011_SQ002.[i], data$matching3N4.SQ012_SQ002.[i], data$matching3N4.SQ013_SQ002.[i], data$matching3N4.SQ014_SQ002.[i], data$matching3N4.SQ015_SQ002.[i]),
    cbind(data$matching3N4.SQ001_SQ003.[i],data$matching3N4.SQ002_SQ003.[i],data$matching3N4.SQ003_SQ003.[i],data$matching3N4.SQ004_SQ003.[i],data$matching3N4.SQ005_SQ003.[i],data$matching3N4.SQ006_SQ003.[i],data$matching3N4.SQ007_SQ003.[i],data$matching3N4.SQ008_SQ003.[i],data$matching3N4.SQ009_SQ003.[i],data$matching3N4.SQ010_SQ003.[i], data$matching3N4.SQ011_SQ003.[i], data$matching3N4.SQ012_SQ003.[i], data$matching3N4.SQ013_SQ003.[i], data$matching3N4.SQ014_SQ003.[i], data$matching3N4.SQ015_SQ003.[i]),
    cbind(data$matching3N4.SQ001_SQ003.[i],data$matching3N4.SQ002_SQ003.[i],data$matching3N4.SQ003_SQ003.[i],data$matching3N4.SQ004_SQ003.[i],data$matching3N4.SQ005_SQ003.[i],data$matching3N4.SQ006_SQ003.[i],data$matching3N4.SQ007_SQ003.[i],data$matching3N4.SQ008_SQ003.[i],data$matching3N4.SQ009_SQ003.[i],data$matching3N4.SQ010_SQ003.[i], data$matching3N4.SQ011_SQ003.[i], data$matching3N4.SQ012_SQ003.[i], data$matching3N4.SQ013_SQ003.[i], data$matching3N4.SQ014_SQ003.[i], data$matching3N4.SQ015_SQ003.[i]),
    cbind(data$matching3N4.SQ001_SQ004.[i],data$matching3N4.SQ002_SQ004.[i],data$matching3N4.SQ003_SQ004.[i],data$matching3N4.SQ004_SQ004.[i],data$matching3N4.SQ005_SQ004.[i],data$matching3N4.SQ006_SQ004.[i],data$matching3N4.SQ007_SQ004.[i],data$matching3N4.SQ008_SQ004.[i],data$matching3N4.SQ009_SQ004.[i],data$matching3N4.SQ010_SQ004.[i], data$matching3N4.SQ011_SQ004.[i], data$matching3N4.SQ012_SQ004.[i], data$matching3N4.SQ013_SQ004.[i], data$matching3N4.SQ014_SQ004.[i], data$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data$matching3N5.SQ001_SQ001.[i],data$matching3N5.SQ002_SQ001.[i],data$matching3N5.SQ003_SQ001.[i],data$matching3N5.SQ004_SQ001.[i],data$matching3N5.SQ005_SQ001.[i],data$matching3N5.SQ006_SQ001.[i],data$matching3N5.SQ007_SQ001.[i],data$matching3N5.SQ008_SQ001.[i],data$matching3N5.SQ009_SQ001.[i],data$matching3N5.SQ010_SQ001.[i], data$matching3N5.SQ011_SQ001.[i], data$matching3N5.SQ012_SQ001.[i], data$matching3N5.SQ013_SQ001.[i], data$matching3N5.SQ014_SQ001.[i], data$matching3N5.SQ015_SQ001.[i]),
    cbind(data$matching3N5.SQ001_SQ002.[i],data$matching3N5.SQ002_SQ002.[i],data$matching3N5.SQ003_SQ002.[i],data$matching3N5.SQ004_SQ002.[i],data$matching3N5.SQ005_SQ002.[i],data$matching3N5.SQ006_SQ002.[i],data$matching3N5.SQ007_SQ002.[i],data$matching3N5.SQ008_SQ002.[i],data$matching3N5.SQ009_SQ002.[i],data$matching3N5.SQ010_SQ002.[i], data$matching3N5.SQ011_SQ002.[i], data$matching3N5.SQ012_SQ002.[i], data$matching3N5.SQ013_SQ002.[i], data$matching3N5.SQ014_SQ002.[i], data$matching3N5.SQ015_SQ002.[i]),
    cbind(data$matching3N5.SQ001_SQ003.[i],data$matching3N5.SQ002_SQ003.[i],data$matching3N5.SQ003_SQ003.[i],data$matching3N5.SQ004_SQ003.[i],data$matching3N5.SQ005_SQ003.[i],data$matching3N5.SQ006_SQ003.[i],data$matching3N5.SQ007_SQ003.[i],data$matching3N5.SQ008_SQ003.[i],data$matching3N5.SQ009_SQ003.[i],data$matching3N5.SQ010_SQ003.[i], data$matching3N5.SQ011_SQ003.[i], data$matching3N5.SQ012_SQ003.[i], data$matching3N5.SQ013_SQ003.[i], data$matching3N5.SQ014_SQ003.[i], data$matching3N5.SQ015_SQ003.[i]),
    cbind(data$matching3N5.SQ001_SQ003.[i],data$matching3N5.SQ002_SQ003.[i],data$matching3N5.SQ003_SQ003.[i],data$matching3N5.SQ004_SQ003.[i],data$matching3N5.SQ005_SQ003.[i],data$matching3N5.SQ006_SQ003.[i],data$matching3N5.SQ007_SQ003.[i],data$matching3N5.SQ008_SQ003.[i],data$matching3N5.SQ009_SQ003.[i],data$matching3N5.SQ010_SQ003.[i], data$matching3N5.SQ011_SQ003.[i], data$matching3N5.SQ012_SQ003.[i], data$matching3N5.SQ013_SQ003.[i], data$matching3N5.SQ014_SQ003.[i], data$matching3N5.SQ015_SQ003.[i]),
    cbind(data$matching3N5.SQ001_SQ005.[i],data$matching3N5.SQ002_SQ005.[i],data$matching3N5.SQ003_SQ005.[i],data$matching3N5.SQ004_SQ005.[i],data$matching3N5.SQ005_SQ005.[i],data$matching3N5.SQ006_SQ005.[i],data$matching3N5.SQ007_SQ005.[i],data$matching3N5.SQ008_SQ005.[i],data$matching3N5.SQ009_SQ005.[i],data$matching3N5.SQ010_SQ005.[i], data$matching3N5.SQ011_SQ005.[i], data$matching3N5.SQ012_SQ005.[i], data$matching3N5.SQ013_SQ005.[i], data$matching3N5.SQ014_SQ005.[i], data$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

#i also calculate structural embeddedness (alter-level) per ego-net.

#first i fix some irregularities in the data
#1. a mistake was made in labeling the adjacency matrix variables for the friends network, so i first relabel them, so the same script below for constructing density scores can be used.
#(I started with N1 instead of N2 (N denotes the number of named alters; which should start at 2, because only at netsizes > 1, I asked for alter-ties) )
data <- data %>%
  rename (adj3N2a.SQ001. = adj3N1a.SQ001., adj3N3a.SQ001. = adj3N2a.SQ001.,adj3N3a.SQ002. = adj3N2a.SQ002.,adj3N3b.SQ001. = adj3N2b.SQ001.,adj3N4a.SQ001. = adj3N3a.SQ001.,adj3N4a.SQ002. = adj3N3a.SQ002.,adj3N4a.SQ003. = adj3N3a.SQ003.,adj3N4b.SQ001. = adj3N3b.SQ001.,adj3N4b.SQ002. = adj3N3b.SQ002.,   adj3N4c.SQ001. = adj3N3c.SQ001.,adj3N5a.SQ001. = adj3N4a.SQ001.,adj3N5a.SQ002. = adj3N4a.SQ002.,adj3N5a.SQ003. = adj3N4a.SQ003.,adj3N5a.SQ004. = adj3N4a.SQ004.,adj3N5b.SQ001. = adj3N4b.SQ001.,adj3N5b.SQ002. = adj3N4b.SQ002.,adj3N5b.SQ003. = adj3N4b.SQ003.,adj3N5c.SQ001. = adj3N4c.SQ001.,adj3N5c.SQ002. = adj3N4c.SQ002.,adj3N5d.SQ001. = adj3N4d.SQ001.)

#2. a labeling mistake: for the adjacency matrix for sports ties, with ns=3, the question about relation between alter 2 and 3 was coded wrongly. 
data$adj4N3b.SQ001. <- data$adj4N3b.SQ002.
##

for (i in 1:length(alterL)) { # for ego i

  #get number of names listed in the egonets
  nnames_cdn <- length(which(data[i,c("egonet1.SQ001.","egonet1.SQ002.","egonet1.SQ003.","egonet1.SQ004.","egonet1.SQ005.")]!=""))
  nnames_study <- length(which(data[i,c("egonet2.SQ001.","egonet2.SQ002.","egonet2.SQ003.","egonet2.SQ004.","egonet2.SQ005.")]!=""))
  nnames_bff <- length(which(data[i,c("egonet3.SQ001.","egonet3.SQ002.","egonet3.SQ003.","egonet3.SQ004.","egonet3.SQ005.")]!=""))
  nnames_csn <- length(which(data[i,c("egonet4.SQ001.","egonet4.SQ002.","egonet4.SQ003.","egonet4.SQ004.","egonet4.SQ005.")]!=""))
  
  #CDN
  
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj1N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj1N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj1N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj1N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj1N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj1N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj1N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj1N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj1N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj1N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj1N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj1N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj1N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj1N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj1N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj1N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj1N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj1N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj1N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj1N5d.SQ001.[i]
  }

  if(nnames_cdn>1) { #we only know alter-alter relations for nnames>1
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_cdn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
  
  #CDN: 1:nnames_CDN
  alterL[[i]]$cdn_embed.t1[1:nnames_cdn] <- embed 
  }
  
  #STUDY

  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj2N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj2N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj2N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj2N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj2N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj2N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj2N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj2N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj2N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj2N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj2N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj2N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj2N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj2N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj2N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj2N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj2N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj2N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj2N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj2N5d.SQ001.[i]
  }

  if(nnames_study>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_study]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$study_embed.t1[6:(5+nnames_study)] <- embed 
  }
  
  # BEST FRIENDS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj3N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj3N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj3N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj3N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj3N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj3N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj3N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj3N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj3N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj3N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj3N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj3N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj3N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj3N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj3N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj3N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj3N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj3N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj3N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj3N5d.SQ001.[i]
  }

  if(nnames_bff>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_bff]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$bff_embed.t1[11:(10+nnames_bff)] <- embed 
  }
  
  # SPORTS PARTNERS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data$adj4N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data$adj4N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data$adj4N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data$adj4N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data$adj4N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data$adj4N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data$adj4N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data$adj4N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data$adj4N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data$adj4N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data$adj4N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data$adj4N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data$adj4N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data$adj4N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data$adj4N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data$adj4N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data$adj4N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data$adj4N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data$adj4N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data$adj4N5d.SQ001.[i]
  }

  if(nnames_csn>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_csn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$csn_embed.t1[16:(15+nnames_csn)] <- embed 
  }
}

#alters could be named in multiple name generators. each element of alterL contains rows corresponding to alters that may be 'duplicates'.
#i match the structural embeddedness measures

for (i in 1:length(alterL)) { #for ego i 
  for (j in 1:nrow(alterL[[i]])) { #for alter j
        
    #if name2 is not empty, alter j was matched to the study network; and more precisely, to the study partner with name1: alterL[[i]]$name2[j]
    alterL[[i]]$study_embed.t1[j] <- ifelse(!is.na(alterL[[i]]$name2[j]), 
                                            alterL[[i]]$study_embed.t1[which(alterL[[i]]$name1==alterL[[i]]$name2[j] & !is.na(alterL[[i]]$study_embed.t1))],
                                            alterL[[i]]$study_embed.t1[j])
    
    #if name3 is not empty, alter j was matched to the friends network; and more precisely, to the friend with name1: alterL[[i]]$name3[j]
    alterL[[i]]$bff_embed.t1[j] <- ifelse(!is.na(alterL[[i]]$name3[j]), 
                                          alterL[[i]]$bff_embed.t1[which(alterL[[i]]$name1==alterL[[i]]$name3[j] & !is.na(alterL[[i]]$bff_embed.t1))],
                                          alterL[[i]]$bff_embed.t1[j])
    
    #if name4 is not empty, alter j was matched to the sports network; and more precisely, to the sports partner with name1: alterL[[i]]$name4[j]
    alterL[[i]]$csn_embed.t1[j] <- ifelse(!is.na(alterL[[i]]$name4[j]), 
                                          alterL[[i]]$csn_embed.t1[which(alterL[[i]]$name1==alterL[[i]]$name4[j] & !is.na(alterL[[i]]$csn_embed.t1))],     
                                          alterL[[i]]$csn_embed.t1[j] )
  }
}

# i use the name interpreter data to add (stable) alter characteristics (e.g., gender)
# in name-interpreter questions, only unique (i.e., non-matched) alters were listed... thus, alter ids without a gender assigned to them are non-uniques, and can later be filtered out.
df_gender <- data.frame(p1 = data$gender.SQ001.,p2 = data$gender.SQ002.,p3 = data$gender.SQ003.,p4 = data$gender.SQ004.,p5 = data$gender.SQ005.,p6 = data$gender.SQ006.,p7 = data$gender.SQ007.,p8 = data$gender.SQ008.,p9 = data$gender.SQ009.,p10= data$gender.SQ010.,p11= data$gender.SQ011.,p12= data$gender.SQ012.,p13= data$gender.SQ013.,p14= data$gender.SQ014.,p15= data$gender.SQ015.,p16= data$gender.SQ016.,p17= data$gender.SQ017.,p18= data$gender.SQ018.,p19= data$gender.SQ019.,p20= data$gender.SQ020.)

# same for age...
df_age <- data.frame(p1 = data$age.SQ001.,p2 = data$age.SQ002.,p3 = data$age.SQ003.,p4 = data$age.SQ004.,p5 = data$age.SQ005.,p6 = data$age.SQ006.,p7 = data$age.SQ007.,p8 = data$age.SQ008.,p9 = data$age.SQ009.,p10= data$age.SQ010.,p11= data$age.SQ011.,p12= data$age.SQ012.,p13= data$age.SQ013.,p14= data$age.SQ014.,p15= data$age.SQ015.,p16= data$age.SQ016.,p17= data$age.SQ017.,p18= data$age.SQ018.,p19= data$age.SQ019.,p20= data$age.SQ020.)

# kin
df_kin <- data.frame( p1 = data$kin.SQ001.,p2 = data$kin.SQ002.,p3 = data$kin.SQ003.,p4 = data$kin.SQ004.,p5 = data$kin.SQ005.,p6 = data$kin.SQ006.,p7 = data$kin.SQ007.,p8 = data$kin.SQ008.,p9 = data$kin.SQ009.,p10= data$kin.SQ010.,p11= data$kin.SQ011.,p12= data$kin.SQ012.,p13= data$kin.SQ013.,p14= data$kin.SQ014.,p15= data$kin.SQ015.,p16= data$kin.SQ016.,p17= data$kin.SQ017.,p18= data$kin.SQ018.,p19= data$kin.SQ019.,p20= data$kin.SQ020.)

#education
df_educ <- data.frame(p1 = data$educ.SQ001.,p2 = data$educ.SQ002.,p3 = data$educ.SQ003.,p4 = data$educ.SQ004.,p5 = data$educ.SQ005.,p6 = data$educ.SQ006.,p7 = data$educ.SQ007.,p8 = data$educ.SQ008., p9 = data$educ.SQ009.,p10= data$educ.SQ010., p11= data$educ.SQ011., p12= data$educ.SQ012., p13= data$educ.SQ013., p14= data$educ.SQ014., p15= data$educ.SQ015., p16= data$educ.SQ016., p17= data$educ.SQ017., p18= data$educ.SQ018., p19= data$educ.SQ019., p20= data$educ.SQ020.)

#also dynamic characteristics
#communication frequency
df_freq <- data.frame(p1 = data$freq.SQ001.,p2 = data$freq.SQ002.,p3 = data$freq.SQ003.,p4 = data$freq.SQ004.,p5 = data$freq.SQ005.,p6 = data$freq.SQ006.,p7 = data$freq.SQ007.,p8 = data$freq.SQ008., p9 = data$freq.SQ009.,p10= data$freq.SQ010., p11= data$freq.SQ011., p12= data$freq.SQ012., p13= data$freq.SQ013., p14= data$freq.SQ014., p15= data$freq.SQ015., p16= data$freq.SQ016., p17= data$freq.SQ017., p18= data$freq.SQ018., p19= data$freq.SQ019., p20= data$freq.SQ020.)

#closeness
df_close <- data.frame(p1 = data$close.SQ001.,p2 = data$close.SQ002.,p3 = data$close.SQ003.,p4 = data$close.SQ004.,p5 = data$close.SQ005.,p6 = data$close.SQ006.,p7 = data$close.SQ007.,p8 = data$close.SQ008., p9 = data$close.SQ009.,p10= data$close.SQ010., p11= data$close.SQ011., p12= data$close.SQ012., p13= data$close.SQ013., p14= data$close.SQ014., p15= data$close.SQ015., p16= data$close.SQ016., p17= data$close.SQ017., p18= data$close.SQ018., p19= data$close.SQ019., p20= data$close.SQ020.)

#and other dyadic featuers
# duration;
df_duration <- data.frame(p1 = data$duur.SQ001.,p2 = data$duur.SQ002.,p3 = data$duur.SQ003.,p4 = data$duur.SQ004.,p5 = data$duur.SQ005.,p6 = data$duur.SQ006.,p7 = data$duur.SQ007.,p8 = data$duur.SQ008., p9 = data$duur.SQ009.,p10= data$duur.SQ010., p11= data$duur.SQ011., p12= data$duur.SQ012., p13= data$duur.SQ013., p14= data$duur.SQ014., p15= data$duur.SQ015., p16= data$duur.SQ016., p17= data$duur.SQ017., p18= data$duur.SQ018., p19= data$duur.SQ019., p20= data$duur.SQ020.)

#geographical proximity
df_proximity <- data.frame(p1 = data$prox.SQ001.,p2 = data$prox.SQ002.,p3 = data$prox.SQ003.,p4 = data$prox.SQ004.,p5 = data$prox.SQ005.,p6 = data$prox.SQ006.,p7 = data$prox.SQ007.,p8 = data$prox.SQ008., p9 = data$prox.SQ009.,p10= data$prox.SQ010., p11= data$prox.SQ011., p12= data$prox.SQ012., p13= data$prox.SQ013., p14= data$prox.SQ014., p15= data$prox.SQ015., p16= data$prox.SQ016., p17= data$prox.SQ017., p18= data$prox.SQ018., p19= data$prox.SQ019., p20= data$prox.SQ020.)

for (i in 1:length(alterL)) {
  
  #gender
  alterL[[i]]$alter_gender <- 
    ifelse(unlist(df_gender[i,], use.names=F)=="Man", 0,  # male = ref.
           ifelse(unlist(df_gender[i,], use.names=F)=="Vrouw", 1,
                  ifelse(unlist(df_gender[i,], use.names=F)=="Anders", 2, "")))
  #kin  
  alterL[[i]]$kin <- 
    ifelse(unlist(df_kin[i,], use.names=F)=="Ja", 1,
           ifelse(unlist(df_kin[i,], use.names=F)=="Nee", 0, ""))
  #age  
  alterL[[i]]$alter_age <- 
    ifelse(unlist(df_age[i,], use.names=F)=="Jonger dan 18 jaar", 16, #???
           ifelse(unlist(df_age[i,], use.names=F)=="18 tot 21 jaar", 20,
                  ifelse(unlist(df_age[i,], use.names=F)=="22 tot 25 jaar", 23,
                         ifelse(unlist(df_age[i,], use.names=F)=="26 tot 30 jaar", 28,
                                ifelse(unlist(df_age[i,], use.names=F)=="31 tot 40 jaar", 35,
                                       ifelse(unlist(df_age[i,], use.names=F)=="Ouder dan 40 jaar", 45, #???
                                              ifelse(unlist(df_age[i,], use.names=F)=="Weet ik niet", NA, # i don't know = missing
                                                     unlist(df_age[i,], use.names=F))))))))
  #education
  alterL[[i]]$alter_educ <- ifelse(unlist(df_educ[i,],use.names=F)=="lagere school", 1,
                               ifelse(unlist(df_educ[i,],use.names=F)=="vmbo, mavo", 2,
                                      ifelse(unlist(df_educ[i,],use.names=F)=="mbo", 3,
                                         ifelse(unlist(df_educ[i,],use.names=F)=="havo", 4,
                                            ifelse(unlist(df_educ[i,],use.names=F)=="vwo / gymnasium", 5,
                                                 ifelse(unlist(df_educ[i,],use.names=F)=="hbo", 6,
                                                     ifelse(unlist(df_educ[i,],use.names=F)=="universiteit", 7, NA))))))) 
  #frequency
  alterL[[i]]$frequency.t1 <- ifelse(unlist(df_freq[i,],use.names=F)=="(Bijna) elke dag", 7,
                               ifelse(unlist(df_freq[i,],use.names=F)=="1-2 keer per week",6,
                                      ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per maand",5, 
                                         ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per maand",4, 
                                            ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per jaar",3,
                                                 ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per jaar",2,
                                                     ifelse(unlist(df_freq[i,],use.names=F)=="Nooit",1, NA )))))))
  #closeness
  alterL[[i]]$closeness.t1 <- ifelse(unlist(df_close[i,],use.names=F)=="Niet hecht", 1, 
                               ifelse(unlist(df_close[i,],use.names=F)=="Enigszins hecht", 2,
                                      ifelse(unlist(df_close[i,],use.names=F)=="Hecht", 3,
                                             ifelse(unlist(df_close[i,],use.names=F)=="Heel erg hecht", 4, NA ))))
  #proximity
  alterL[[i]]$proximity <- ifelse(unlist(df_proximity[i,], use.names=FALSE) == "In hetzelfde huis", "roommate",
                                  ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In dezelfde buurt" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde straat" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde gemeente", "close",
                                    ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In hetzelfde land" |
                                             unlist(df_proximity[i,], use.names = FALSE) == "In een ander land","far", NA)))

  
   #duration:  take midpoint on scale.
   alterL[[i]]$duration <- ifelse(unlist(df_duration[i,],use.names=F)=="Minder dan 1 jaar", 0, 
                               ifelse(unlist(df_duration[i,],use.names=F)=="1 tot 3 jaar", 2,
                                      ifelse(unlist(df_duration[i,],use.names=F)=="4 tot 8 jaar", 6,
                                             ifelse(unlist(df_duration[i,],use.names=F)=="9 tot 15 jaar", 12,
                                                    ifelse(unlist(df_duration[i,],use.names=F)=="Meer dan 15 jaar", 15,NA )))))
}

# before we can construct similarity indices, we must add ego-attributes
for (i in 1:length(alterL)) {
  #  gender
  alterL[[i]]$ego_gender <- ifelse(data$A1[i] == "Man", 0,ifelse(data$A1[i] == "Vrouw", 1,ifelse(data$A1[i] == "Overige", 2,NA)))
  
  # education 
  # these score should correspond to scores given to alter educ...
  alterL[[i]]$ego_educ <- ifelse(data$T1[i] == "aan de Hogeschool van Arnhem en Nijmegen (HAN)", "6", "7" )
 
  # age
  # calculated as the difference between submission date and birthdate
  # in years.
  birth_date <- as.Date(data$A2[i])
  #x_date <- as.Date(data$submitdate) #submission date was not deposited; so, take 1-jan 2023
  x_date <- as.Date("2023-01-01")
  alterL[[i]]$ego_age <- trunc((birth_date %--% x_date) / years(1))
}

# i filter out the unique, non-duplicate alters;
# by excluding alters with empty strings for gender attribute
for ( i in 1:length(alterL)) {
  alterL[[i]] <- alterL[[i]][which(alterL[[i]]$alter_gender!=""),]
  # and replace the alter id: 1 : no. unique alters
  alterL[[i]]$alterid <- 1:nrow(alterL[[i]])
}

#dyadic similarity measures
for (i in 1:length(alterL))  { # for ego i
  # get attributes of ego
  agei <- alterL[[i]]$ego_age[1]
  genderi <- alterL[[i]]$ego_gender[1]
  edi <- alterL[[i]]$ego_educ[1]
  
  for (j in 1:max(alterL[[i]]$alterid)) { # for alter j
    # calculate "same gender" (0/1)
    genderj <- as.numeric(alterL[[i]]$alter_gender[j]) # get alter j gender
    same <- ifelse(genderi==genderj, 1, 0)
    alterL[[i]]$same_gender[which(alterL[[i]]$alterid==j)] <- same
    
    #same education
    eduj <- alterL[[i]]$alter_educ[j]
    same <- ifelse(eduj==edi, 1, 0)
    alterL[[i]]$sim_educ[which(alterL[[i]]$alterid==j)] <- same
    
    # calculate similarity for age as the absolute difference between alter and ego value
    #get alter j attributes
    agej <- as.numeric(alterL[[i]]$alter_age[j])
    #difference score
    difage <- abs(agei - agej)
    
    # so higher values represent *less* similarity
    if( !is.na (agej) ) { # age similarity only calculable if z_j is known!
        alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- difage
      } else { alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- NA}
  }
}

#based on this, i make a long dataframe with alters nested in ego.
#first, add an ego_id
for (i in 1:length(alterL)) {
  alterL[[i]]$ego <- i
  alterL[[i]]$respnr <- data$respnr[i]
}

#combine using rbind
df <- do.call(rbind,alterL)

#other ego variables:
#transitions
# we need info of waves 1 and 2, so i merge them
d <- merge(data3, data4, by="respnr")
d <- data.frame(d[order(d$respnr, decreasing = FALSE), ]) # and sort by respnr

#1. residential transitions
df$housing.t1a <- NA #wave 1, pre-transition
df$housing.t1b <- NA #wave 1, post-transition
df$housing.t2 <- NA #wave 2
df$housing.transition <- NA #no. of changes 

#make sure descriptions of living situation matches over waves
d$A7 <- ifelse(d$A7 == "inwonend bij ouder(s)/verzorger(s)", "inwonend bij je ouder(s)/verzorger(s)", d$A7)
d$A4d <- ifelse(d$A4d == "inwonend bij ouder(s)/verzorger(s)", "inwonend bij je ouder(s)/verzorger(s)", d$A4d)

for (i in unique(df$ego)) {
  df$housing.t1a[which(df$ego == i)] <- d$A4.x[i] # w1, half year prior to transition
    
  # if respondent had moved at the time of w1, get current living situation
  df$housing.t1b[which(df$ego==i)] <- ifelse(d$A6[i] == "Nee", d$A7[i], df$housing.t1a[which(df$ego==i)])
  
  #if they moved, this indicates a transition
  transition <- ifelse(d$A6[i]=="Nee", 1, 0)
  
  # if respondent has moved since then (in w2), ...
  df$housing.t2[which(df$ego==i)] <- ifelse(d$A4.y[i] == "Ja" | d$A4b[i] == "Ja" | d$A4c[i] == "Ja", d$A4d[i], df$housing.t1b[which(df$ego==i)]) 
  
  transition <- ifelse(d$A4.y[i] == "Ja" | d$A4b[i] == "Ja" | d$A4c[i] == "Ja", transition + 1, transition)
  
  # no. of housing transition
  df$housing.transition[which(df$ego==i)] <- transition
}

#also make it binary (yes/no)
df$housing.transition_bin <- ifelse(df$housing.transition > 0, 1, 0)

#2. transition to university
df$occupation.t1 <- NA
df$occupation.t2 <- "higher" #everybody is in school...
df$study.year <- NA
df$occupation.transition <- NA 

for (i in unique(df$ego)) {
  df$occupation.t1[which(df$ego==i)] <- ifelse(data$E1[i]=="Ik zat op de middelbare school", "secondary",
                             ifelse(data$E1[i]=="Ik deed een MBO-, HBO- of WO-opleiding","higher",
                                    ifelse(data$E1[i]=="Ik werkte","work",
                                             ifelse(data$E1[i]=="Ik had een tussenjaar","gap",
                                                    ifelse(data$E1[i]=="Overige","other",NA)))))

  df$study.year[which(df$ego==i)] <- ifelse(data$T3[i]=="in het eerste jaar", 1,
                                            ifelse(data$T3[i]=="in het tweede jaar", 2,
                                                   ifelse(data$T3[i]=="in het derde jaar of hoger", 3, NA)))
  
  #transition = 1 for those who are first years and who came from secondary school/gap year
  transition <- ifelse( data$T3[i]=="in het eerste jaar" & (data$E1[i]=="Ik zat op de middelbare school" | data$E2[i]=="Ik had een tussenjaar"), 1, 0)
  
  # did ego drop out/start a new study?
  transition <- ifelse(data4$G07Q90[i]== "Nee", transition + 1, transition)
  df$occupation.transition[which(df$ego==i)] <- transition
}

#also make binary
df$occupation.transition_bin <- ifelse(df$occupation.transition > 0, 1, 0)

#romantic relationship 
df$romantic <- NA
for (i in unique(df$ego)) {
  df$romantic[which(df$ego==i)] <- ifelse(data$R3[i] == "Ja", 1,0)
}

#psychological/other variables (wave 2)
#loneliness  
l <- cbind(d$stellingen.SQ001., d$stellingen.SQ002.)
l <- ifelse(l=="Helemaal oneens", 1, ifelse(l=="Mee oneens", 2,
                   # do not know = neutral... this category was sparsely populated 
                   ifelse(l=="Neutraal" | l=="Ik weet het niet", 3, 
                          ifelse(l=="Mee eens", 4, ifelse(l=="Helemaal eens", 5, l)))))
# turn the first indicator, so that higher scores indicate "more" loneliness.
l[,1] <- 6-as.numeric(l[,1])

#extraversion
e <- cbind(d$stellingen.SQ003., d$stellingen.SQ006.)
e <- ifelse(e=="Helemaal oneens", 1,
            ifelse(e=="Mee oneens", 2,
                   ifelse(e=="Neutraal" | e=="Ik weet het niet", 3, 
                          ifelse(e=="Mee eens", 4,
                                 ifelse(e=="Helemaal eens", 5, e)))))
# turn the second indicator
e[,2] <- 6-as.numeric(e[,2])

# calculate averages.
df$loneliness <- NA
df$extraversion <- NA
for (i in unique(df$ego)) {
  l_scores <- as.numeric(l[i,])
  e_scores <- as.numeric(e[i,])
  df$loneliness[which(df$ego==i)] <- mean(l_scores)
  df$extraversion[which(df$ego==i)] <- mean(e_scores)
}

# financial restrictions (for social interaction)
df$fin_restr <- NA
# 0. never; 1. sometimes; 2. often; 3. always
for (i in unique(df$ego)) {
  df$fin_restr[which(df$ego==i)] <- d$G01Q111[i]
}
df$fin_restr <- ifelse(df$fin_restr=="Nooit",0,
                       ifelse(df$fin_restr=="Soms",1,
                              ifelse(df$fin_restr=="Vaak",2,
                                     ifelse(df$fin_restr=="Altijd",3, NA))))

# parental SES / educational level:
for (i in unique(df$ego)) {
  df$educ_father[which(df$ego == i)] <- data4$G08Q56.SQ001.[i]
  df$educ_mother[which(df$ego == i)] <- data4$G08Q56.SQ002.[i]
}

# add network variables
# 1. size (no. of unique alters)
df$netsize <- NA
# also per egonet
df$cdn.size <- df$study.size <- df$csn.size <- df$bff.size <- 0

for (i in unique(df$ego)) {
  #no. of unique alters is simply the number of rows per ego
  df$netsize[which(df$ego==i)] <- nrow(df[which(df$ego==i),])
  
  #size of each egonet is the number of alters named in the name generators
  #thus, including 'duplicates'
  df$cdn.size[which(df$ego==i)] <- length(which(!df_names[i,c(1:5)]==""))
  df$study.size[which(df$ego==i)] <- length(which(!df_names[i,c(6:10)]==""))
  df$bff.size[which(df$ego==i)] <- length(which(!df_names[i,c(11:15)]==""))
  df$csn.size[which(df$ego==i)] <- length(which(!df_names[i,c(16:20)]==""))
}

# 2. density.
# this can only be calculated per egonet, as only between-alter ties within egonet are asked.
df$cdn.density <- df$study.density <- df$csn.density <- df$bff.density <- NA

for (i in unique(df$ego)) {
  #a. CDN
  size <- df$cdn.size[which(df$ego==i)][1]
  #no. possible ties between alters
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  #get no. of observed ties
  #only if ego named at least 2 alters in this name generator
  #because only then i can calculate density
  if(size>1) {
    #get alter-alter adjacency matrices of this particular egonet
    adjacency <- data[i,c(13:32)]
   
    #i want the correct ones, given the no. of listed alters in this name generator
    #i list the columns numbers of the variables corresponding to each matching matrix set,
    #and subset corresponding columns in `adjacency`
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$cdn.density[which(df$ego==i)] <- obs/pos  }
  
  #b. study
  size <- df$study.size[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data[i,c(53:72)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
  
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$study.density[which(df$ego==i)] <- obs/pos }
  
  #c. friends
  size <- df$bff.size[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data[i,c(169:188)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$bff.density[which(df$ego==i)] <- obs/pos }

  #d. sports partners
  size <- df$csn.size[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data[i,c(444:463)]

    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$csn.density[which(df$ego==i)] <- obs/pos }
}

#set NAs to 0.
df$cdn.density[is.na(df$cdn.density)] <- 0
df$study.density[is.na(df$study.density)] <- 0
df$bff.density[is.na(df$bff.density)] <- 0
df$csn.density[is.na(df$csn.density)] <- 0

#binary attributes indicating whether alters appeared in each of the ego-nets at w1.
for (i in unique(df$ego)) {  
    for (j in 1:nrow(df[which(df$ego==i),])) { # for alters nested in ego

    # find out if alter_id denoting alter j appear in the 4 egonets
    alter <- unlist(df[which(df$ego==i & df$alterid==j),][5:8], use.names = FALSE) # get name(s) of alter j
    alter <- alter[!is.na(alter)] # exclude NAs
    
    cdn <- alter %in% net1[i,]
    study <- alter %in% net2[i,]
    bff <- alter %in% net3[i,]
    csn <- alter %in% net4[i,]
    
    # and if so, give alter j score 1 on indicators; 0 otherwise
    df$cdn1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% cdn, 1, 0)
    df$study1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% study, 1, 0)
    df$bff1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% bff, 1, 0)
    df$csn1[which(df$ego==i & df$alterid==j)] <- ifelse("TRUE" %in% csn, 1, 0)
  }
}

#calculate multiplexity as an alter/tie attribute (i.e., number of *additional* networks j appeared in)
df$multiplex <- rowSums(df[,c(51:54)]) - 1 #minus one to reach a meaningful intercept
#psych::describe(df$multiplex) #M=.51; SD=.76

#now, find out whether w1-alters were maintainted in w2.
df$survive <- NA

# In wave 2, another matching procedure was used; in rows, alters named in w2 are listed; columns denote (unique) alters of w1, so the same one as our data-frame.
# again, i made multiple matrices, conditional on the no. of unique alters of w1 (15, to be precise)
# if, in one of these matrices, column j was marked as corresponding to a named alter in w2, then alter j did, indeed, survive

# subset w1-w2 matching matrices; and respnr
w1w2 <- data4[,c(756:1155,length(data4))]

for (i in unique(df$ego)) {  # for ego 
  matchingL <- vector("list", 20) #pre-allocate empty list of length 20, to store matching matrices

  {
    matchingL[[1]] <- w1w2[i,1:20]
    matchingL[[2]] <- w1w2[i,21:40]
    matchingL[[3]] <- w1w2[i,41:60]
    matchingL[[4]] <- w1w2[i,61:80]
    matchingL[[5]] <- w1w2[i,81:100]
    matchingL[[6]] <- w1w2[i,101:120]
    matchingL[[7]] <- w1w2[i,121:140]
    matchingL[[8]] <- w1w2[i,141:160]
    matchingL[[9]] <- w1w2[i,161:180]
    matchingL[[10]] <- w1w2[i,181:200]
    matchingL[[11]] <- w1w2[i,201:220]
    matchingL[[12]] <- w1w2[i,221:240]
    matchingL[[13]] <- w1w2[i,241:260]
    matchingL[[14]] <- w1w2[i,261:280]
    matchingL[[15]] <- w1w2[i,281:300]
    matchingL[[16]] <- w1w2[i,301:320]
    matchingL[[17]] <- w1w2[i,321:340]
    matchingL[[18]] <- w1w2[i,341:360]
    matchingL[[19]] <- w1w2[i,361:380]
    matchingL[[20]] <- w1w2[i,381:400]
  }
  
  #find the 'right' matching matrix in this list, by taking the one that contains answers
  ind <- NULL
  for (j in seq_along(matchingL)) {
    #check along the sequence of elements j in the matching matrix list if they are non-empty and not NA
    check <- FALSE
    for (col in matchingL[[j]]) {
      if (!all(is.na(col)) && any(nchar(col) > 0)) {
        check <- TRUE
        break  
      }
      }
    if (check) {
      ind <- j
      break  
    }
  }
  if(length(ind)>0) {
    # get the  corresponding matrix
    mm <- matchingL[[ind]]
    
    # unlist the answers given
    ans <- unlist(mm, use.names = FALSE)
    
    # use 'stringr' to extract numbers in these answers,
    # they refer to the id of "survived" alters
    id <- unlist(stringr::str_extract_all(ans,"\\(?[0-9,.]+\\)?"))
    id <- id[!is.na(id)] # exclude NA
    
    # if alter id of ego i corresponds to this id, they did indeed survive:
    df$survive[which(df$ego==i)] <- ifelse(df$alterid[which(df$ego==i)] %in% id, 1, 0)
  }
}

# in which egonet did they (re)appear?
df$cdn2 <- NA
df$study2 <- NA
df$bff2 <- NA
df$csn2 <- NA

# i use the matching matrices of w2; where columns refer to unique w1 alters (those of our constructed dataframe) and rows referring to unique w2 alters
# so, if w1-alter/column j is matched to w2-alter/row i, then w1-alter did survive (but we already knew that); but more importantly, j was the i^th alter mentioned
# and this allows me to infer in what network j (re-)appeared
# since cdn = 1-5; study = 6-10; bff = 11-15; csn = 16-20.
# however, since the rows only include non-duplicate w2-alters;
# if alter j at t1 was named more than once at t2, i only know the first
# egonet in which j was named...
# thus, before i proceed, i use the matching matrices that allowed ego to match alters from different egonets to make alter ids for w2 alters; and figure out to which egonets they belonged

# i make a dataframe of w2-alters, with potential duplicates...
df_names <- data.frame(p1 = data4$egonet1.SQ001.,p2 = data4$egonet1.SQ002.,p3 = data4$egonet1.SQ003.,p4 = data4$egonet1.SQ004.,p5 = data4$egonet1.SQ005.,p6 = data4$egonet2.SQ001.,p7 = data4$egonet2.SQ002.,p8 = data4$egonet2.SQ003.,p9 = data4$egonet2.SQ004.,p10= data4$egonet2.SQ005.,p11= data4$egonet3.SQ001.,p12= data4$egonet3.SQ002.,p13= data4$egonet3.SQ003.,p14= data4$egonet3.SQ004.,p15= data4$egonet3.SQ005.,p16= data4$egonet4.SQ001., p17= data4$egonet4.SQ002., p18= data4$egonet4.SQ003.,p19= data4$egonet4.SQ004.,p20= data4$egonet4.SQ005.)

#males
df_gender <- data.frame(p1 = data4$males.SQ001.,p2 = data4$males.SQ002.,p3 = data4$males.SQ003.,p4 = data4$males.SQ004.,p5 = data4$males.SQ005.,p6 = data4$males.SQ006.,p7 = data4$males.SQ007.,p8 = data4$males.SQ008.,p9 = data4$males.SQ009.,p10= data4$males.SQ010.,p11= data4$males.SQ011.,p12= data4$males.SQ012.,p13= data4$males.SQ013.,p14= data4$males.SQ014.,p15= data4$males.SQ015.,p16= data4$males.SQ016.,p17= data4$males.SQ017.,p18= data4$males.SQ018.,p19= data4$males.SQ019.,p20= data4$males.SQ020.)

# list of dataframes per ego with rows reflecting alters
# and columns indicating the name(s) of the particular alters
alterL <- list()
# loop over all egos
for ( i in 1:nrow(df_names)) {alterL[[i]] <- data.frame(
    alterid = 1:20,name1 = NA,name2 = NA, name3 = NA,name4 = NA, gender=NA,
    cdn_embed.t2 = NA, study_embed.t2 = NA, bff_embed.t2 = NA, csn_embed.t2 = NA)}

# fill the names based on the names data-frame
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=F)
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
  
  alterL[[i]]$gender <- unlist(df_gender[i,], use.names=F)
  alterL[[i]]$gender <- ifelse(alterL[[i]]$gender=="Ja", 0, #male = ref
                               ifelse(alterL[[i]]$gender=="Nee", 1, NA))
  
  }

# calculate netsize for each net to get the correct matching matrix
{
  net1 <- cbind(data4$egonet1.SQ001.,data4$egonet1.SQ002., data4$egonet1.SQ003.,data4$egonet1.SQ004., data4$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data4$egonet2.SQ001.,data4$egonet2.SQ002., data4$egonet2.SQ003.,data4$egonet2.SQ004., data4$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data4$egonet3.SQ001.,data4$egonet3.SQ002., data4$egonet3.SQ003.,data4$egonet3.SQ004., data4$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data4$egonet4.SQ001.,data4$egonet4.SQ002., data4$egonet4.SQ003.,data4$egonet4.SQ004., data4$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}

# construct the matching matrices list for egonet1-2
matchingList <- list()
for (i in 1:length(alterL)) {
  matchingL <- list()
  matchingL[[1]] <- cbind(data4$matching1N1.SQ001_SQ001.[i],data4$matching1N1.SQ002_SQ001.[i],data4$matching1N1.SQ003_SQ001.[i],data4$matching1N1.SQ004_SQ001.[i],data4$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data4$matching1N2.SQ001_SQ001.[i], data4$matching1N2.SQ002_SQ001.[i], data4$matching1N2.SQ003_SQ001.[i], data4$matching1N2.SQ004_SQ001.[i], data4$matching1N2.SQ005_SQ001.[i]),
    cbind(data4$matching1N2.SQ001_SQ002.[i], data4$matching1N2.SQ002_SQ002.[i], data4$matching1N2.SQ003_SQ002.[i], data4$matching1N2.SQ004_SQ002.[i], data4$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data4$matching1N3.SQ001_SQ001.[i], data4$matching1N3.SQ002_SQ001.[i], data4$matching1N3.SQ003_SQ001.[i], data4$matching1N3.SQ004_SQ001.[i], data4$matching1N3.SQ005_SQ001.[i]),
    cbind(data4$matching1N3.SQ001_SQ002.[i], data4$matching1N3.SQ002_SQ002.[i], data4$matching1N3.SQ003_SQ002.[i], data4$matching1N3.SQ004_SQ002.[i], data4$matching1N3.SQ005_SQ002.[i]),
    cbind(data4$matching1N3.SQ001_SQ003.[i], data4$matching1N3.SQ002_SQ003.[i], data4$matching1N3.SQ003_SQ003.[i], data4$matching1N3.SQ004_SQ003.[i], data4$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data4$matching1N4.SQ001_SQ001.[i], data4$matching1N4.SQ002_SQ001.[i], data4$matching1N4.SQ003_SQ001.[i], data4$matching1N4.SQ004_SQ001.[i], data4$matching1N4.SQ005_SQ001.[i]),
    cbind(data4$matching1N4.SQ001_SQ002.[i], data4$matching1N4.SQ002_SQ002.[i], data4$matching1N4.SQ003_SQ002.[i], data4$matching1N4.SQ004_SQ002.[i], data4$matching1N4.SQ005_SQ002.[i]),
    cbind(data4$matching1N4.SQ001_SQ003.[i], data4$matching1N4.SQ002_SQ003.[i], data4$matching1N4.SQ003_SQ003.[i], data4$matching1N4.SQ004_SQ003.[i], data4$matching1N4.SQ005_SQ003.[i]),
    cbind(data4$matching1N4.SQ001_SQ004.[i], data4$matching1N4.SQ002_SQ004.[i], data4$matching1N4.SQ003_SQ004.[i], data4$matching1N4.SQ004_SQ004.[i], data4$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data4$matching1N5.SQ001_SQ001.[i], data4$matching1N5.SQ002_SQ001.[i], data4$matching1N5.SQ003_SQ001.[i], data4$matching1N5.SQ004_SQ001.[i], data4$matching1N5.SQ005_SQ001.[i]),
    cbind(data4$matching1N5.SQ001_SQ002.[i], data4$matching1N5.SQ002_SQ002.[i], data4$matching1N5.SQ003_SQ002.[i], data4$matching1N5.SQ004_SQ002.[i], data4$matching1N5.SQ005_SQ002.[i]),
    cbind(data4$matching1N5.SQ001_SQ003.[i], data4$matching1N5.SQ002_SQ003.[i], data4$matching1N5.SQ003_SQ003.[i], data4$matching1N5.SQ004_SQ003.[i], data4$matching1N5.SQ005_SQ003.[i]),
    cbind(data4$matching1N5.SQ001_SQ004.[i], data4$matching1N5.SQ002_SQ004.[i], data4$matching1N5.SQ003_SQ004.[i], data4$matching1N5.SQ004_SQ004.[i], data4$matching1N5.SQ005_SQ004.[i]),
    cbind(data4$matching1N5.SQ001_SQ005.[i], data4$matching1N5.SQ002_SQ005.[i], data4$matching1N5.SQ003_SQ005.[i], data4$matching1N5.SQ004_SQ005.[i], data4$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
} # so... matchingL[[1]][[5]] is matchingmatrix 5 (i.e., 5 alters named in egonet2) for ego 1

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
}#ignore warning

# again, make a matching list for the second matching matrices
# (i.e., matching egonet3 alters to prev. alters);
matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data4$matching2N1.SQ001_SQ001.[i],data4$matching2N1.SQ002_SQ001.[i],data4$matching2N1.SQ003_SQ001.[i],data4$matching2N1.SQ004_SQ001.[i],data4$matching2N1.SQ005_SQ001.[i],data4$matching2N1.SQ006_SQ001.[i],data4$matching2N1.SQ007_SQ001.[i],data4$matching2N1.SQ008_SQ001.[i],data4$matching2N1.SQ009_SQ001.[i],data4$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data4$matching2N2.SQ001_SQ001.[i],data4$matching2N2.SQ002_SQ001.[i],data4$matching2N2.SQ003_SQ001.[i],data4$matching2N2.SQ004_SQ001.[i],data4$matching2N2.SQ005_SQ001.[i],data4$matching2N2.SQ006_SQ001.[i],data4$matching2N2.SQ007_SQ001.[i],data4$matching2N2.SQ008_SQ001.[i],data4$matching2N2.SQ009_SQ001.[i],data4$matching2N2.SQ010_SQ001.[i]),
    cbind(data4$matching2N2.SQ001_SQ002.[i],data4$matching2N2.SQ002_SQ002.[i],data4$matching2N2.SQ003_SQ002.[i],data4$matching2N2.SQ004_SQ002.[i],data4$matching2N2.SQ005_SQ002.[i],data4$matching2N2.SQ006_SQ002.[i],data4$matching2N2.SQ007_SQ002.[i],data4$matching2N2.SQ008_SQ002.[i],data4$matching2N2.SQ009_SQ002.[i],data4$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data4$matching2N3.SQ001_SQ001.[i],data4$matching2N3.SQ002_SQ001.[i],data4$matching2N3.SQ003_SQ001.[i],data4$matching2N3.SQ004_SQ001.[i],data4$matching2N3.SQ005_SQ001.[i],data4$matching2N3.SQ006_SQ001.[i],data4$matching2N3.SQ007_SQ001.[i],data4$matching2N3.SQ008_SQ001.[i],data4$matching2N3.SQ009_SQ001.[i],data4$matching2N3.SQ010_SQ001.[i]),
    cbind(data4$matching2N3.SQ001_SQ002.[i],data4$matching2N3.SQ002_SQ002.[i],data4$matching2N3.SQ003_SQ002.[i],data4$matching2N3.SQ004_SQ002.[i],data4$matching2N3.SQ005_SQ002.[i],data4$matching2N3.SQ006_SQ002.[i],data4$matching2N3.SQ007_SQ002.[i],data4$matching2N3.SQ008_SQ002.[i],data4$matching2N3.SQ009_SQ002.[i],data4$matching2N3.SQ010_SQ002.[i]),
    cbind(data4$matching2N3.SQ001_SQ003.[i],data4$matching2N3.SQ002_SQ003.[i],data4$matching2N3.SQ003_SQ003.[i],data4$matching2N3.SQ004_SQ003.[i],data4$matching2N3.SQ005_SQ003.[i],data4$matching2N3.SQ006_SQ003.[i],data4$matching2N3.SQ007_SQ003.[i],data4$matching2N3.SQ008_SQ003.[i],data4$matching2N3.SQ009_SQ003.[i],data4$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data4$matching2N4.SQ001_SQ001.[i],data4$matching2N4.SQ002_SQ001.[i],data4$matching2N4.SQ003_SQ001.[i],data4$matching2N4.SQ004_SQ001.[i],data4$matching2N4.SQ005_SQ001.[i],data4$matching2N4.SQ006_SQ001.[i],data4$matching2N4.SQ007_SQ001.[i],data4$matching2N4.SQ008_SQ001.[i],data4$matching2N4.SQ009_SQ001.[i],data4$matching2N4.SQ010_SQ001.[i]),
    cbind(data4$matching2N4.SQ001_SQ002.[i],data4$matching2N4.SQ002_SQ002.[i],data4$matching2N4.SQ003_SQ002.[i],data4$matching2N4.SQ004_SQ002.[i],data4$matching2N4.SQ005_SQ002.[i],data4$matching2N4.SQ006_SQ002.[i],data4$matching2N4.SQ007_SQ002.[i],data4$matching2N4.SQ008_SQ002.[i],data4$matching2N4.SQ009_SQ002.[i],data4$matching2N4.SQ010_SQ002.[i]),
    cbind(data4$matching2N4.SQ001_SQ003.[i],data4$matching2N4.SQ002_SQ003.[i],data4$matching2N4.SQ003_SQ003.[i],data4$matching2N4.SQ004_SQ003.[i],data4$matching2N4.SQ005_SQ003.[i],data4$matching2N4.SQ006_SQ003.[i],data4$matching2N4.SQ007_SQ003.[i],data4$matching2N4.SQ008_SQ003.[i],data4$matching2N4.SQ009_SQ003.[i],data4$matching2N4.SQ010_SQ003.[i]),
    cbind(data4$matching2N4.SQ001_SQ004.[i],data4$matching2N4.SQ002_SQ004.[i],data4$matching2N4.SQ003_SQ004.[i],data4$matching2N4.SQ004_SQ004.[i],data4$matching2N4.SQ005_SQ004.[i],data4$matching2N4.SQ006_SQ004.[i],data4$matching2N4.SQ007_SQ004.[i],data4$matching2N4.SQ008_SQ004.[i],data4$matching2N4.SQ009_SQ004.[i],data4$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data4$matching2N5.SQ001_SQ001.[i],data4$matching2N5.SQ002_SQ001.[i],data4$matching2N5.SQ003_SQ001.[i],data4$matching2N5.SQ004_SQ001.[i],data4$matching2N5.SQ005_SQ001.[i],data4$matching2N5.SQ006_SQ001.[i],data4$matching2N5.SQ007_SQ001.[i],data4$matching2N5.SQ008_SQ001.[i],data4$matching2N5.SQ009_SQ001.[i],data4$matching2N5.SQ010_SQ001.[i]),
    cbind(data4$matching2N5.SQ001_SQ002.[i],data4$matching2N5.SQ002_SQ002.[i],data4$matching2N5.SQ003_SQ002.[i],data4$matching2N5.SQ004_SQ002.[i],data4$matching2N5.SQ005_SQ002.[i],data4$matching2N5.SQ006_SQ002.[i],data4$matching2N5.SQ007_SQ002.[i],data4$matching2N5.SQ008_SQ002.[i],data4$matching2N5.SQ009_SQ002.[i],data4$matching2N5.SQ010_SQ002.[i]),
    cbind(data4$matching2N5.SQ001_SQ003.[i],data4$matching2N5.SQ002_SQ003.[i],data4$matching2N5.SQ003_SQ003.[i],data4$matching2N5.SQ004_SQ003.[i],data4$matching2N5.SQ005_SQ003.[i],data4$matching2N5.SQ006_SQ003.[i],data4$matching2N5.SQ007_SQ003.[i],data4$matching2N5.SQ008_SQ003.[i],data4$matching2N5.SQ009_SQ003.[i],data4$matching2N5.SQ010_SQ003.[i]),
    cbind(data4$matching2N5.SQ001_SQ004.[i],data4$matching2N5.SQ002_SQ004.[i],data4$matching2N5.SQ003_SQ004.[i],data4$matching2N5.SQ004_SQ004.[i],data4$matching2N5.SQ005_SQ004.[i],data4$matching2N5.SQ006_SQ004.[i],data4$matching2N5.SQ007_SQ004.[i],data4$matching2N5.SQ008_SQ004.[i],data4$matching2N5.SQ009_SQ004.[i],data4$matching2N5.SQ010_SQ004.[i]),
    cbind(data4$matching2N5.SQ001_SQ005.[i],data4$matching2N5.SQ002_SQ005.[i],data4$matching2N5.SQ003_SQ005.[i],data4$matching2N5.SQ004_SQ005.[i],data4$matching2N5.SQ005_SQ005.[i],data4$matching2N5.SQ006_SQ005.[i],data4$matching2N5.SQ007_SQ005.[i],data4$matching2N5.SQ008_SQ005.[i],data4$matching2N5.SQ009_SQ005.[i],data4$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

# same for matching 3 (i.e., egonet4 with egonets 1-3)
matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data4$matching3N1.SQ001_SQ001.[i],data4$matching3N1.SQ002_SQ001.[i],data4$matching3N1.SQ003_SQ001.[i],data4$matching3N1.SQ004_SQ001.[i],data4$matching3N1.SQ005_SQ001.[i],data4$matching3N1.SQ006_SQ001.[i],data4$matching3N1.SQ007_SQ001.[i],data4$matching3N1.SQ008_SQ001.[i],data4$matching3N1.SQ009_SQ001.[i],data4$matching3N1.SQ010_SQ001.[i], data4$matching3N1.SQ011_SQ001.[i], data4$matching3N1.SQ012_SQ001.[i], data4$matching3N1.SQ013_SQ001.[i], data4$matching3N1.SQ014_SQ001.[i], data4$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data4$matching3N2.SQ001_SQ001.[i],data4$matching3N2.SQ002_SQ001.[i],data4$matching3N2.SQ003_SQ001.[i],data4$matching3N2.SQ004_SQ001.[i],data4$matching3N2.SQ005_SQ001.[i],data4$matching3N2.SQ006_SQ001.[i],data4$matching3N2.SQ007_SQ001.[i],data4$matching3N2.SQ008_SQ001.[i],data4$matching3N2.SQ009_SQ001.[i],data4$matching3N2.SQ010_SQ001.[i], data4$matching3N2.SQ011_SQ001.[i], data4$matching3N2.SQ012_SQ001.[i], data4$matching3N2.SQ013_SQ001.[i], data4$matching3N2.SQ014_SQ001.[i], data4$matching3N2.SQ015_SQ001.[i]),
    cbind(data4$matching3N2.SQ001_SQ002.[i],data4$matching3N2.SQ002_SQ002.[i],data4$matching3N2.SQ003_SQ002.[i],data4$matching3N2.SQ004_SQ002.[i],data4$matching3N2.SQ005_SQ002.[i],data4$matching3N2.SQ006_SQ002.[i],data4$matching3N2.SQ007_SQ002.[i],data4$matching3N2.SQ008_SQ002.[i],data4$matching3N2.SQ009_SQ002.[i],data4$matching3N2.SQ010_SQ002.[i], data4$matching3N2.SQ011_SQ002.[i], data4$matching3N2.SQ012_SQ002.[i], data4$matching3N2.SQ013_SQ002.[i], data4$matching3N2.SQ014_SQ002.[i], data4$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data4$matching3N3.SQ001_SQ001.[i],data4$matching3N3.SQ002_SQ001.[i],data4$matching3N3.SQ003_SQ001.[i],data4$matching3N3.SQ004_SQ001.[i],data4$matching3N3.SQ005_SQ001.[i],data4$matching3N3.SQ006_SQ001.[i],data4$matching3N3.SQ007_SQ001.[i],data4$matching3N3.SQ008_SQ001.[i],data4$matching3N3.SQ009_SQ001.[i],data4$matching3N3.SQ010_SQ001.[i], data4$matching3N3.SQ011_SQ001.[i], data4$matching3N3.SQ012_SQ001.[i], data4$matching3N3.SQ013_SQ001.[i], data4$matching3N3.SQ014_SQ001.[i], data4$matching3N3.SQ015_SQ001.[i]),
    cbind(data4$matching3N3.SQ001_SQ002.[i],data4$matching3N3.SQ002_SQ002.[i],data4$matching3N3.SQ003_SQ002.[i],data4$matching3N3.SQ004_SQ002.[i],data4$matching3N3.SQ005_SQ002.[i],data4$matching3N3.SQ006_SQ002.[i],data4$matching3N3.SQ007_SQ002.[i],data4$matching3N3.SQ008_SQ002.[i],data4$matching3N3.SQ009_SQ002.[i],data4$matching3N3.SQ010_SQ002.[i], data4$matching3N3.SQ011_SQ002.[i], data4$matching3N3.SQ012_SQ002.[i], data4$matching3N3.SQ013_SQ002.[i], data4$matching3N3.SQ014_SQ002.[i], data4$matching3N3.SQ015_SQ002.[i]),
    cbind(data4$matching3N3.SQ001_SQ003.[i],data4$matching3N3.SQ002_SQ003.[i],data4$matching3N3.SQ003_SQ003.[i],data4$matching3N3.SQ004_SQ003.[i],data4$matching3N3.SQ005_SQ003.[i],data4$matching3N3.SQ006_SQ003.[i],data4$matching3N3.SQ007_SQ003.[i],data4$matching3N3.SQ008_SQ003.[i],data4$matching3N3.SQ009_SQ003.[i],data4$matching3N3.SQ010_SQ003.[i], data4$matching3N3.SQ011_SQ003.[i], data4$matching3N3.SQ012_SQ003.[i], data4$matching3N3.SQ013_SQ003.[i], data4$matching3N3.SQ014_SQ003.[i], data4$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data4$matching3N4.SQ001_SQ001.[i],data4$matching3N4.SQ002_SQ001.[i],data4$matching3N4.SQ003_SQ001.[i],data4$matching3N4.SQ004_SQ001.[i],data4$matching3N4.SQ005_SQ001.[i],data4$matching3N4.SQ006_SQ001.[i],data4$matching3N4.SQ007_SQ001.[i],data4$matching3N4.SQ008_SQ001.[i],data4$matching3N4.SQ009_SQ001.[i],data4$matching3N4.SQ010_SQ001.[i], data4$matching3N4.SQ011_SQ001.[i], data4$matching3N4.SQ012_SQ001.[i], data4$matching3N4.SQ013_SQ001.[i], data4$matching3N4.SQ014_SQ001.[i], data4$matching3N4.SQ015_SQ001.[i]),
    cbind(data4$matching3N4.SQ001_SQ002.[i],data4$matching3N4.SQ002_SQ002.[i],data4$matching3N4.SQ003_SQ002.[i],data4$matching3N4.SQ004_SQ002.[i],data4$matching3N4.SQ005_SQ002.[i],data4$matching3N4.SQ006_SQ002.[i],data4$matching3N4.SQ007_SQ002.[i],data4$matching3N4.SQ008_SQ002.[i],data4$matching3N4.SQ009_SQ002.[i],data4$matching3N4.SQ010_SQ002.[i], data4$matching3N4.SQ011_SQ002.[i], data4$matching3N4.SQ012_SQ002.[i], data4$matching3N4.SQ013_SQ002.[i], data4$matching3N4.SQ014_SQ002.[i], data4$matching3N4.SQ015_SQ002.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ004.[i],data4$matching3N4.SQ002_SQ004.[i],data4$matching3N4.SQ003_SQ004.[i],data4$matching3N4.SQ004_SQ004.[i],data4$matching3N4.SQ005_SQ004.[i],data4$matching3N4.SQ006_SQ004.[i],data4$matching3N4.SQ007_SQ004.[i],data4$matching3N4.SQ008_SQ004.[i],data4$matching3N4.SQ009_SQ004.[i],data4$matching3N4.SQ010_SQ004.[i], data4$matching3N4.SQ011_SQ004.[i], data4$matching3N4.SQ012_SQ004.[i], data4$matching3N4.SQ013_SQ004.[i], data4$matching3N4.SQ014_SQ004.[i], data4$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data4$matching3N5.SQ001_SQ001.[i],data4$matching3N5.SQ002_SQ001.[i],data4$matching3N5.SQ003_SQ001.[i],data4$matching3N5.SQ004_SQ001.[i],data4$matching3N5.SQ005_SQ001.[i],data4$matching3N5.SQ006_SQ001.[i],data4$matching3N5.SQ007_SQ001.[i],data4$matching3N5.SQ008_SQ001.[i],data4$matching3N5.SQ009_SQ001.[i],data4$matching3N5.SQ010_SQ001.[i], data4$matching3N5.SQ011_SQ001.[i], data4$matching3N5.SQ012_SQ001.[i], data4$matching3N5.SQ013_SQ001.[i], data4$matching3N5.SQ014_SQ001.[i], data4$matching3N5.SQ015_SQ001.[i]),
    cbind(data4$matching3N5.SQ001_SQ002.[i],data4$matching3N5.SQ002_SQ002.[i],data4$matching3N5.SQ003_SQ002.[i],data4$matching3N5.SQ004_SQ002.[i],data4$matching3N5.SQ005_SQ002.[i],data4$matching3N5.SQ006_SQ002.[i],data4$matching3N5.SQ007_SQ002.[i],data4$matching3N5.SQ008_SQ002.[i],data4$matching3N5.SQ009_SQ002.[i],data4$matching3N5.SQ010_SQ002.[i], data4$matching3N5.SQ011_SQ002.[i], data4$matching3N5.SQ012_SQ002.[i], data4$matching3N5.SQ013_SQ002.[i], data4$matching3N5.SQ014_SQ002.[i], data4$matching3N5.SQ015_SQ002.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ005.[i],data4$matching3N5.SQ002_SQ005.[i],data4$matching3N5.SQ003_SQ005.[i],data4$matching3N5.SQ004_SQ005.[i],data4$matching3N5.SQ005_SQ005.[i],data4$matching3N5.SQ006_SQ005.[i],data4$matching3N5.SQ007_SQ005.[i],data4$matching3N5.SQ008_SQ005.[i],data4$matching3N5.SQ009_SQ005.[i],data4$matching3N5.SQ010_SQ005.[i], data4$matching3N5.SQ011_SQ005.[i], data4$matching3N5.SQ012_SQ005.[i], data4$matching3N5.SQ013_SQ005.[i], data4$matching3N5.SQ014_SQ005.[i], data4$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

#let us here also attach structural embeddedness. because the 'network environment' for maintained alters may change in t2

#again, fix the irregularities in the data.
data4 <- data4 %>%
  rename (adj3N2a.SQ001. = adj3N1a.SQ001., adj3N3a.SQ001. = adj3N2a.SQ001.,adj3N3a.SQ002. = adj3N2a.SQ002.,adj3N3b.SQ001. = adj3N2b.SQ001.,adj3N4a.SQ001. = adj3N3a.SQ001.,adj3N4a.SQ002. = adj3N3a.SQ002.,adj3N4a.SQ003. = adj3N3a.SQ003.,adj3N4b.SQ001. = adj3N3b.SQ001.,adj3N4b.SQ002. = adj3N3b.SQ002.,   adj3N4c.SQ001. = adj3N3c.SQ001.,adj3N5a.SQ001. = adj3N4a.SQ001.,adj3N5a.SQ002. = adj3N4a.SQ002.,adj3N5a.SQ003. = adj3N4a.SQ003.,adj3N5a.SQ004. = adj3N4a.SQ004.,adj3N5b.SQ001. = adj3N4b.SQ001.,adj3N5b.SQ002. = adj3N4b.SQ002.,adj3N5b.SQ003. = adj3N4b.SQ003.,adj3N5c.SQ001. = adj3N4c.SQ001.,adj3N5c.SQ002. = adj3N4c.SQ002.,adj3N5d.SQ001. = adj3N4d.SQ001.)

data4$adj4N3b.SQ001. <- data4$adj4N3b.SQ002.

for (i in 1:length(alterL)) { # for ego i

  #get number of names listed in the egonets
  nnames_cdn <- length(which(data4[i,c("egonet1.SQ001.","egonet1.SQ002.","egonet1.SQ003.","egonet1.SQ004.","egonet1.SQ005.")]!=""))
  nnames_study <- length(which(data4[i,c("egonet2.SQ001.","egonet2.SQ002.","egonet2.SQ003.","egonet2.SQ004.","egonet2.SQ005.")]!=""))
  nnames_bff <- length(which(data4[i,c("egonet3.SQ001.","egonet3.SQ002.","egonet3.SQ003.","egonet3.SQ004.","egonet3.SQ005.")]!=""))
  nnames_csn <- length(which(data4[i,c("egonet4.SQ001.","egonet4.SQ002.","egonet4.SQ003.","egonet4.SQ004.","egonet4.SQ005.")]!=""))
  
  #CDN
  
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj1N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj1N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj1N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj1N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj1N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj1N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj1N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj1N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj1N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj1N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj1N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj1N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj1N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj1N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj1N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj1N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj1N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj1N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj1N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj1N5d.SQ001.[i]
  }

  if(nnames_cdn>1) { #we only know alter-alter relations for nnames>1
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_cdn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
  
  #CDN: 1:nnames_CDN
  alterL[[i]]$cdn_embed.t2[1:nnames_cdn] <- embed 
  }
  
  #STUDY

  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj2N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj2N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj2N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj2N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj2N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj2N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj2N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj2N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj2N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj2N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj2N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj2N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj2N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj2N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj2N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj2N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj2N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj2N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj2N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj2N5d.SQ001.[i]
  }

  if(nnames_study>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_study]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$study_embed.t2[6:(5+nnames_study)] <- embed 
  }
  
  # BEST FRIENDS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj3N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj3N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj3N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj3N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj3N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj3N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj3N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj3N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj3N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj3N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj3N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj3N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj3N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj3N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj3N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj3N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj3N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj3N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj3N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj3N5d.SQ001.[i]
  }

  if(nnames_bff>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_bff]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$bff_embed.t2[11:(10+nnames_bff)] <- embed 
  }
  
  # SPORTS PARTNERS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj4N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj4N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj4N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj4N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj4N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj4N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj4N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj4N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj4N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj4N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj4N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj4N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj4N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj4N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj4N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj4N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj4N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj4N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj4N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj4N5d.SQ001.[i]
  }

  if(nnames_csn>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_csn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$csn_embed.t2[16:(15+nnames_csn)] <- embed 
  }
}

#alters could be named in multiple name generators. each element of alterL contains rows corresponding to alters that may be 'duplicates'.
#i match the structural embeddedness measures

for (i in 1:length(alterL)) { #for ego i 
  for (j in 1:nrow(alterL[[i]])) { #for alter j
        
    #if name2 is not empty, alter j was matched to the study network; and more precisely, to the study partner with name1: alterL[[i]]$name2[j]
    alterL[[i]]$study_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name2[j]), 
                                            alterL[[i]]$study_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name2[j] & !is.na(alterL[[i]]$study_embed.t2))],
                                            alterL[[i]]$study_embed.t2[j])
    
    #if name3 is not empty, alter j was matched to the friends network; and more precisely, to the friend with name1: alterL[[i]]$name3[j]
    alterL[[i]]$bff_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name3[j]), 
                                          alterL[[i]]$bff_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name3[j] & !is.na(alterL[[i]]$bff_embed.t2))],
                                          alterL[[i]]$bff_embed.t2[j])
    
    #if name4 is not empty, alter j was matched to the sports network; and more precisely, to the sports partner with name1: alterL[[i]]$name4[j]
    alterL[[i]]$csn_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name4[j]), 
                                          alterL[[i]]$csn_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name4[j] & !is.na(alterL[[i]]$csn_embed.t2))],     
                                          alterL[[i]]$csn_embed.t2[j] )
  }
}

# make long df with alters in ego,
# and indicators for 4 egonets;
df2 <- data.frame(
  ego=rep(1:length(alterL),each=20),alter=rep(1:20),cdn=NA, study=NA,bff=NA,csn=NA,
  cdn_embed.t2=NA, study_embed.t2=NA, bff_embed.t2=NA, csn_embed.t2=NA)

for (i in unique(df2$ego)) {  
      for (j in 1:20) { # for alters nested in ego
                 
        # find out if names denoting alter j appear in the 4 egonets
        alter <- unlist(alterL[[i]][j,c(2:5)], use.names = FALSE)
        alter <- alter[!is.na(alter)] # exclude NAs
        
        cdn <- alter %in% net1[i,]
        study <- alter %in% net2[i,]
        bff <- alter %in% net3[i,]
        csn <- alter %in% net4[i,]
        
        # and if so, give alter j score 1 on indicators; 0 otherwise
        df2$cdn[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% cdn, 1, 0)
        df2$study[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% study, 1, 0)
        df2$bff[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% bff, 1, 0)
        df2$csn[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% csn, 1, 0)
        
        #and attach embeddedness t2 scores
        df2$cdn_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$cdn_embed.t2[j]
        df2$study_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$study_embed.t2[j]
        df2$bff_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$bff_embed.t2[j]
        df2$csn_embed.t2[which(df2$ego==i & df2$alter==j)] <- alterL[[i]]$csn_embed.t2[j]
      }
}

# now that i have, for each alter of ego (including duplicates) at t2,
# the nets to which they belong..
# i continue with the w1-w2 matching matrices

# we already subsetted the matching matrices (in `w1w2`)
for (i in unique(df$ego)) {  # for ego i
  
  matchingL <- vector("list", 20) #pre-allocate empty list of length 15, to store matching matrices
  {
    matchingL[[1]] <- w1w2[i,1:20]
    matchingL[[2]] <- w1w2[i,21:40]
    matchingL[[3]] <- w1w2[i,41:60]
    matchingL[[4]] <- w1w2[i,61:80]
    matchingL[[5]] <- w1w2[i,81:100]
    matchingL[[6]] <- w1w2[i,101:120]
    matchingL[[7]] <- w1w2[i,121:140]
    matchingL[[8]] <- w1w2[i,141:160]
    matchingL[[9]] <- w1w2[i,161:180]
    matchingL[[10]] <- w1w2[i,181:200]
    matchingL[[11]] <- w1w2[i,201:220]
    matchingL[[12]] <- w1w2[i,221:240]
    matchingL[[13]] <- w1w2[i,241:260]
    matchingL[[14]] <- w1w2[i,261:280]
    matchingL[[15]] <- w1w2[i,281:300]
    matchingL[[16]] <- w1w2[i,301:320]
    matchingL[[17]] <- w1w2[i,321:340]
    matchingL[[18]] <- w1w2[i,341:360]
    matchingL[[19]] <- w1w2[i,361:380]
    matchingL[[20]] <- w1w2[i,381:400]
  }
  
  #find the 'right' matching matrix in this list, by taking the one that contains answers
  ind <- NULL
  for (j in seq_along(matchingL)) {
    #check along the sequence of elements j in the matching matrix list if they are non-empty and not NA
    check <- FALSE
    for (col in matchingL[[j]]) {
      if (!all(is.na(col)) && any(nchar(col) > 0)) {
        check <- TRUE
        break  
      }
      }
    if (check) {
      ind <- j
      break  
    }
  }
  
  if(length(ind)>0) {
    # get the  corresponding matrix
    mm <- matchingL[[ind]]
    
    # extract alter ids
    ans <- stringr::str_extract_all(mm,"\\(?[0-9,.]+\\)?") 
    # here, the object refers to the w1-alter j;
    # and the element indicator in the list refers to the w2-alter 
    
    for (j in unique(df$alterid[which(df$ego==i)])) { # for w1-alter j,
      
      # to which row/w2-alter was alter j matched?
      match <- which(ans==j)
      # and in which networks did this w2-alter belong?
      nets <- df2[which(df2$ego==i & df2$alter==match[1]),] #add [1] to match, for the few cases where multiple w2-alters were matched to the same w1-alter (i.e., when ego did not correclty perform match w2 alters to each other.)
    
    if(length(match)>0) { # if j was matched to w2-alters...
      # assign to alter j the networks in which j reappeared
      df$cdn2[which(df$ego==i & df$alterid==j)] <- nets$cdn[1]
      df$study2[which(df$ego==i & df$alterid==j)] <- nets$study[1]
      df$bff2[which(df$ego==i & df$alterid==j)] <- nets$bff[1]
      df$csn2[which(df$ego==i & df$alterid==j)] <- nets$csn[1]
      
      #and attach their t2 embeddednes score
      df$cdn_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$cdn_embed.t2[1]
      df$study_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$study_embed.t2[1]
      df$bff_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$bff_embed.t2[1]
      df$csn_embed.t2[which(df$ego==i & df$alterid==j)] <- nets$csn_embed.t2[1]
    }
  }
  }
}

# currently; a lot of NAs in the attributes indicating whether alter j was member of the respective networks.
# those are alters that were not renamed, and thus have no 'chance' to re-appear. so, 0s are alters that were once part of a specific network, but re-appeared not in that particular network.
# naturally, for analyses, NAs can be set to 0.

# as a second tie change indicator, i look at whether (un-listed) alters and ego were still in touch
df$frequency.t2 <- NA
df$closeness.t2 <- NA

# subset w2 name interpreters on contact freq. of w1 alters (sq021 - sq040!)
freq <- data4[,c(1176:1195)]
# also closeness
close <- data4[,c(1216:1235)]

# recode into numeric values;
freq <- ifelse(freq=="(Bijna) elke dag",7,
                       ifelse(freq=="1-2 keer per week",6,
                              ifelse(freq=="Aantal keer per maand",5, 
                                     ifelse(freq=="Ong. 1 keer per maand",4, 
                                            ifelse(freq=="Aantal keer per jaar",3,
                                                   ifelse(freq=="Ong. 1 keer per jaar",2,
                                                          ifelse(freq=="Nooit",1,
                                                                      NA   )))))))
close <- ifelse(close=="Heel erg hecht", 4, ifelse(close=="Hecht", 3, ifelse(close=="Enigszins hecht",2, ifelse(close=="Niet hecht", 1, NA))))

for (i in unique(df$ego)) {
  for (j in unique(df$alterid[which(df$ego==i)])) {
    df$frequency.t2[which(df$ego==i & df$alterid==j)] <- freq[i,][!is.na(freq[i,])][j]
    df$closeness.t2[which(df$ego==i & df$alterid==j)] <- close[i,][!is.na(close[i,])][j]
  }
}

# size of each egonet at t2
df$cdn.size2 <- NA
df$study.size2 <- NA
df$bff.size2 <- NA
df$csn.size2 <- NA

for (i in unique(df$ego)) {
  df$cdn.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet1.SQ001.[i],data4$egonet1.SQ002.[i],data4$egonet1.SQ003.[i],data4$egonet1.SQ004.[i],data4$egonet1.SQ005.[i])!="")
    df$study.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet2.SQ001.[i],data4$egonet2.SQ002.[i],data4$egonet2.SQ003.[i],data4$egonet2.SQ004.[i],data4$egonet2.SQ005.[i])!="")
    df$bff.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet3.SQ001.[i],data4$egonet3.SQ002.[i],data4$egonet3.SQ003.[i],data4$egonet3.SQ004.[i],data4$egonet3.SQ005.[i])!="")
    df$csn.size2[which(df$ego==i)] <- rowSums(cbind(data4$egonet4.SQ001.[i],data4$egonet4.SQ002.[i],data4$egonet4.SQ003.[i],data4$egonet4.SQ004.[i],data4$egonet4.SQ005.[i])!="") }

#and density at t2.
df$cdn.density2 <- df$study.density2 <- df$csn.density2 <- df$bff.density2 <- NA

for (i in unique(df$ego)) {
  #a. CDN
  size <- df$cdn.size2[which(df$ego==i)][1]
  #no. possible ties between alters
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  #get no. of observed ties
  #only if ego named at least 2 alters in this name generator
  #because only then i can calculate density
  if(size>1) {
    #get alter-alter adjacency matrices of this particular egonet
    adjacency <- data4[i,c(8:27)]

    #i want the correct ones, given the no. of listed alters in this name generator
    #i list the columns numbers of the variables corresponding to each matching matrix set,
    #and subset corresponding columns in `adjacency`
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
   
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$cdn.density2[which(df$ego==i)] <- obs/pos  }
  
  #b. study
  size <- df$study.size2[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data4[i,c(33:52)]
    #reorder columns.... messy!
    adjacency <- adjacency[,c("adj2N2a.SQ001.", "adj2N3a.SQ001.", "adj2N3a.SQ002.", "adj2N3b.SQ001.", "adj2N4a.SQ001.", "adj2N4a.SQ002.",
                       "adj2N4a.SQ003.", "adj2N4b.SQ001.", "adj2N4b.SQ002.", "adj2N4c.SQ001.", "adj2N5a.SQ001.", "adj2N5a.SQ002.",
                       "adj2N5a.SQ003.", "adj2N5a.SQ004.", "adj2N5b.SQ001.", "adj2N5b.SQ002.", "adj2N5b.SQ003.", "adj2N5c.SQ001.",
                       "adj2N5c.SQ002.", "adj2N5d.SQ001.")]
    
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
  
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$study.density2[which(df$ego==i)] <- obs/pos }
  
  #c. friends
  size <- df$bff.size2[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data4[i,c(156:175)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
    
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$bff.density2[which(df$ego==i)] <- obs/pos }

  #d. sports partners
  size <- df$csn.size2[which(df$ego==i)][1]
  pos <- ifelse(size==2, 1, ifelse(size==3, 3, ifelse(size==4, 6, ifelse(size==5, 10, 0))))
  if(size>1) {
    adjacency <- data4[i,c(401:420)]
    adj <- adjacency[,list(0,1,c(2:4),c(5:10),c(11:20))[size][[1]]]
  
    #count number of ties that ego described as (very) close (vs. not close but no stranger; strangers)
    obs <- length(which(adj=="Hecht"|adj=="Erg hecht"))
    df$csn.density2[which(df$ego==i)] <- obs/pos }
}

#set NAs to 0.
df$cdn.density2[is.na(df$cdn.density2)] <- 0
df$study.density2[is.na(df$study.density2)] <- 0
df$bff.density2[is.na(df$bff.density2)] <- 0
df$csn.density2[is.na(df$csn.density2)] <- 0

#add Wave 2 status
df$statusW2 <- NA
df$statusW2 <- ifelse(df$survive==1, "Maintained", "Dropped")

## methodological controls

#include whether max number of names was reached at t2
df$cdn.max2 <- ifelse(df$cdn.size2==5,1,0)
df$study.max2 <- ifelse(df$study.size2==5,1,0)
df$bff.max2 <- ifelse(df$bff.size2==5,1,0)
df$csn.max2 <- ifelse(df$csn.size2==5,1,0)

#survey time 
df$survey2time <- NA

#time of day
df$beforenine2 <- NA
df$ninetofive2 <- NA
df$afterfive2 <- NA

#filled out on the weekend (sa/su)?
df$weekend2 <- NA

#time per name generator
df$cdn.time2 <- NA
df$study.time2 <- NA
df$bff.time2 <- NA
df$csn.time2 <- NA

#time till name generator
df$cdn.timetill2 <- NA
df$study.timetill2 <- NA
df$bff.timetill2 <- NA
df$csn.timetill2 <- NA

for (i in unique(df$ego)) {
  
  #survey time at wave 2
  #convert start and submitdate to POSIXct and calculate difference
  #in minutes; extract number.
  df$survey2time[which(df$ego==i)] <- stringr::str_extract_all(trunc(as.POSIXct(data4$submitdate[i]) - as.POSIXct(data4$startdate[i]), "mins"),"\\(?[0-9,.]+\\)?")[[1]]
  
  #time of day of survey.
  #table(substr(data.table::as.ITime(as.POSIXct(data4$startdate)),1,2)) #distribution
  df$beforenine2[which(df$ego==i)] <- ifelse ( data.table::as.ITime(as.POSIXct(data4$startdate[i])) < data.table::as.ITime("9:00:00"), 1, 0)
  df$ninetofive2[which(df$ego==i)] <- ifelse ( data.table::as.ITime(as.POSIXct(data4$startdate[i])) > data.table::as.ITime("9:00:00") &
                                                 data.table::as.ITime(as.POSIXct(data4$startdate[i])) < data.table::as.ITime("17:00:00"), 1, 0)
  df$afterfive2[which(df$ego==i)] <- ifelse ( data.table::as.ITime(as.POSIXct(data4$startdate[i])) > data.table::as.ITime("17:00:00"), 1, 0)
  
  #time per name generator question, in minutes
 
  df$cdn.time2[which(df$ego==i)] <- data4$groupTime718[i]/60
  df$study.time2[which(df$ego==i)] <- data4$groupTime722[i]/60
  df$bff.time2[which(df$ego==i)] <- data4$groupTime730[i]/60
  df$csn.time2[which(df$ego==i)] <- data4$groupTime737[i]/60
  
  #also get weekday/weekend
  #counting friday as weekend, after 5pm
  
  #table(wday(ymd_hms(as.POSIXct(data4$submitdate)),label = TRUE)) #distribution
  #extract weekday
  weekday <- wday(ymd_hms(as.POSIXct(data4$submitdate[i])),label = TRUE)
  #and time
  time <- hour(as.POSIXct(data4$submitdate[i]))
  df$weekend2[which(df$ego==i)] <- ifelse( (weekday %in% c("za", "zo")) | (weekday == "vr" & time >= 17), 1, 0)
  
  #time till name generator question, in minutes
  # wave 2: question groups 717 CDN718 720 STUDY722 723 724 725 748 793 726 749 729 740 BFF730 731 732 734 735 736 743 CSN737
  df$cdn.timetill2[which(df$ego==i)] <- sum(data4[,c("groupTime717")][i])/60
  df$study.timetill2[which(df$ego==i)] <- rowSums(data4[,c("groupTime717", "groupTime718", "groupTime720")][i,], na.rm = TRUE)/60
  df$bff.timetill2[which(df$ego==i)] <- rowSums(data4[,c("groupTime717", "groupTime718", "groupTime720", "groupTime722", "groupTime723", "groupTime724", "groupTime725", "groupTime748", "groupTime793", "groupTime726", "groupTime749","groupTime729", "groupTime740")][i,], na.rm = TRUE)/60
  df$csn.timetill2[which(df$ego==i)] <- rowSums(data4[,c("groupTime717", "groupTime718", "groupTime720", "groupTime722", "groupTime723", "groupTime724", "groupTime725", "groupTime748", "groupTime793", "groupTime726", "groupTime749","groupTime729", "groupTime740","groupTime730", "groupTime731", "groupTime732", "groupTime734", "groupTime735", "groupTime736", "groupTime743")][i,], na.rm = TRUE)/60
  
}

df_maintained <- df
#fix(df_maintained)
#str(df_maintained)
#last make gender numeric..
df_maintained$alter_gender <- as.numeric(df_maintained$alter_gender)
```

<br>

## wave 2 created

```{r, eval = FALSE}
#create a new alter List
alterL <- vector("list", nrow(data))

# loop over all egos
for ( i in 1:length(alterL)) {
    alterL[[i]] <- data.frame(
      ego_gender = NA, ego_age = NA, ego_educ = NA,
      alterid = 1:20, name1 = NA, name2 = NA, name3 = NA, name4 = NA,
      alter_gender = NA, alter_age = NA, alter_educ=NA,
      same_gender = NA, dif_age = NA, sim_educ = NA,
      cdn_embed.t2 = NA, study_embed.t2 = NA, bff_embed.t2 = NA, csn_embed.t2 = NA)
}

#fill in names based on names_df
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=FALSE)
}

# replace empty strings with <NA>
for (i in 1:length(alterL)) {
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
}

#matching
{
  net1 <- cbind(data4$egonet1.SQ001.,data4$egonet1.SQ002., data4$egonet1.SQ003.,data4$egonet1.SQ004., data4$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data4$egonet2.SQ001.,data4$egonet2.SQ002., data4$egonet2.SQ003.,data4$egonet2.SQ004., data4$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data4$egonet3.SQ001.,data4$egonet3.SQ002., data4$egonet3.SQ003.,data4$egonet3.SQ004., data4$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data4$egonet4.SQ001.,data4$egonet4.SQ002., data4$egonet4.SQ003.,data4$egonet4.SQ004., data4$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}

matchingList <- list()
for (i in 1:length(alterL)) { # for ego i
  matchingL <- list()
  # make 5 seperate matching matrices. naturally, only 1 is relevant... but i will select that later.
  matchingL[[1]] <- cbind(data4$matching1N1.SQ001_SQ001.[i],data4$matching1N1.SQ002_SQ001.[i],data4$matching1N1.SQ003_SQ001.[i],data4$matching1N1.SQ004_SQ001.[i],data4$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data4$matching1N2.SQ001_SQ001.[i], data4$matching1N2.SQ002_SQ001.[i], data4$matching1N2.SQ003_SQ001.[i], data4$matching1N2.SQ004_SQ001.[i], data4$matching1N2.SQ005_SQ001.[i]),
    cbind(data4$matching1N2.SQ001_SQ002.[i], data4$matching1N2.SQ002_SQ002.[i], data4$matching1N2.SQ003_SQ002.[i], data4$matching1N2.SQ004_SQ002.[i], data4$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data4$matching1N3.SQ001_SQ001.[i], data4$matching1N3.SQ002_SQ001.[i], data4$matching1N3.SQ003_SQ001.[i], data4$matching1N3.SQ004_SQ001.[i], data4$matching1N3.SQ005_SQ001.[i]),
    cbind(data4$matching1N3.SQ001_SQ002.[i], data4$matching1N3.SQ002_SQ002.[i], data4$matching1N3.SQ003_SQ002.[i], data4$matching1N3.SQ004_SQ002.[i], data4$matching1N3.SQ005_SQ002.[i]),
    cbind(data4$matching1N3.SQ001_SQ003.[i], data4$matching1N3.SQ002_SQ003.[i], data4$matching1N3.SQ003_SQ003.[i], data4$matching1N3.SQ004_SQ003.[i], data4$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data4$matching1N4.SQ001_SQ001.[i], data4$matching1N4.SQ002_SQ001.[i], data4$matching1N4.SQ003_SQ001.[i], data4$matching1N4.SQ004_SQ001.[i], data4$matching1N4.SQ005_SQ001.[i]),
    cbind(data4$matching1N4.SQ001_SQ002.[i], data4$matching1N4.SQ002_SQ002.[i], data4$matching1N4.SQ003_SQ002.[i], data4$matching1N4.SQ004_SQ002.[i], data4$matching1N4.SQ005_SQ002.[i]),
    cbind(data4$matching1N4.SQ001_SQ003.[i], data4$matching1N4.SQ002_SQ003.[i], data4$matching1N4.SQ003_SQ003.[i], data4$matching1N4.SQ004_SQ003.[i], data4$matching1N4.SQ005_SQ003.[i]),
    cbind(data4$matching1N4.SQ001_SQ004.[i], data4$matching1N4.SQ002_SQ004.[i], data4$matching1N4.SQ003_SQ004.[i], data4$matching1N4.SQ004_SQ004.[i], data4$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data4$matching1N5.SQ001_SQ001.[i], data4$matching1N5.SQ002_SQ001.[i], data4$matching1N5.SQ003_SQ001.[i], data4$matching1N5.SQ004_SQ001.[i], data4$matching1N5.SQ005_SQ001.[i]),
    cbind(data4$matching1N5.SQ001_SQ002.[i], data4$matching1N5.SQ002_SQ002.[i], data4$matching1N5.SQ003_SQ002.[i], data4$matching1N5.SQ004_SQ002.[i], data4$matching1N5.SQ005_SQ002.[i]),
    cbind(data4$matching1N5.SQ001_SQ003.[i], data4$matching1N5.SQ002_SQ003.[i], data4$matching1N5.SQ003_SQ003.[i], data4$matching1N5.SQ004_SQ003.[i], data4$matching1N5.SQ005_SQ003.[i]),
    cbind(data4$matching1N5.SQ001_SQ004.[i], data4$matching1N5.SQ002_SQ004.[i], data4$matching1N5.SQ003_SQ004.[i], data4$matching1N5.SQ004_SQ004.[i], data4$matching1N5.SQ005_SQ004.[i]),
    cbind(data4$matching1N5.SQ001_SQ005.[i], data4$matching1N5.SQ002_SQ005.[i], data4$matching1N5.SQ003_SQ005.[i], data4$matching1N5.SQ004_SQ005.[i], data4$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
}

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
}

matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data4$matching2N1.SQ001_SQ001.[i],data4$matching2N1.SQ002_SQ001.[i],data4$matching2N1.SQ003_SQ001.[i],data4$matching2N1.SQ004_SQ001.[i],data4$matching2N1.SQ005_SQ001.[i],data4$matching2N1.SQ006_SQ001.[i],data4$matching2N1.SQ007_SQ001.[i],data4$matching2N1.SQ008_SQ001.[i],data4$matching2N1.SQ009_SQ001.[i],data4$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data4$matching2N2.SQ001_SQ001.[i],data4$matching2N2.SQ002_SQ001.[i],data4$matching2N2.SQ003_SQ001.[i],data4$matching2N2.SQ004_SQ001.[i],data4$matching2N2.SQ005_SQ001.[i],data4$matching2N2.SQ006_SQ001.[i],data4$matching2N2.SQ007_SQ001.[i],data4$matching2N2.SQ008_SQ001.[i],data4$matching2N2.SQ009_SQ001.[i],data4$matching2N2.SQ010_SQ001.[i]),
    cbind(data4$matching2N2.SQ001_SQ002.[i],data4$matching2N2.SQ002_SQ002.[i],data4$matching2N2.SQ003_SQ002.[i],data4$matching2N2.SQ004_SQ002.[i],data4$matching2N2.SQ005_SQ002.[i],data4$matching2N2.SQ006_SQ002.[i],data4$matching2N2.SQ007_SQ002.[i],data4$matching2N2.SQ008_SQ002.[i],data4$matching2N2.SQ009_SQ002.[i],data4$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data4$matching2N3.SQ001_SQ001.[i],data4$matching2N3.SQ002_SQ001.[i],data4$matching2N3.SQ003_SQ001.[i],data4$matching2N3.SQ004_SQ001.[i],data4$matching2N3.SQ005_SQ001.[i],data4$matching2N3.SQ006_SQ001.[i],data4$matching2N3.SQ007_SQ001.[i],data4$matching2N3.SQ008_SQ001.[i],data4$matching2N3.SQ009_SQ001.[i],data4$matching2N3.SQ010_SQ001.[i]),
    cbind(data4$matching2N3.SQ001_SQ002.[i],data4$matching2N3.SQ002_SQ002.[i],data4$matching2N3.SQ003_SQ002.[i],data4$matching2N3.SQ004_SQ002.[i],data4$matching2N3.SQ005_SQ002.[i],data4$matching2N3.SQ006_SQ002.[i],data4$matching2N3.SQ007_SQ002.[i],data4$matching2N3.SQ008_SQ002.[i],data4$matching2N3.SQ009_SQ002.[i],data4$matching2N3.SQ010_SQ002.[i]),
    cbind(data4$matching2N3.SQ001_SQ003.[i],data4$matching2N3.SQ002_SQ003.[i],data4$matching2N3.SQ003_SQ003.[i],data4$matching2N3.SQ004_SQ003.[i],data4$matching2N3.SQ005_SQ003.[i],data4$matching2N3.SQ006_SQ003.[i],data4$matching2N3.SQ007_SQ003.[i],data4$matching2N3.SQ008_SQ003.[i],data4$matching2N3.SQ009_SQ003.[i],data4$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data4$matching2N4.SQ001_SQ001.[i],data4$matching2N4.SQ002_SQ001.[i],data4$matching2N4.SQ003_SQ001.[i],data4$matching2N4.SQ004_SQ001.[i],data4$matching2N4.SQ005_SQ001.[i],data4$matching2N4.SQ006_SQ001.[i],data4$matching2N4.SQ007_SQ001.[i],data4$matching2N4.SQ008_SQ001.[i],data4$matching2N4.SQ009_SQ001.[i],data4$matching2N4.SQ010_SQ001.[i]),
    cbind(data4$matching2N4.SQ001_SQ002.[i],data4$matching2N4.SQ002_SQ002.[i],data4$matching2N4.SQ003_SQ002.[i],data4$matching2N4.SQ004_SQ002.[i],data4$matching2N4.SQ005_SQ002.[i],data4$matching2N4.SQ006_SQ002.[i],data4$matching2N4.SQ007_SQ002.[i],data4$matching2N4.SQ008_SQ002.[i],data4$matching2N4.SQ009_SQ002.[i],data4$matching2N4.SQ010_SQ002.[i]),
    cbind(data4$matching2N4.SQ001_SQ003.[i],data4$matching2N4.SQ002_SQ003.[i],data4$matching2N4.SQ003_SQ003.[i],data4$matching2N4.SQ004_SQ003.[i],data4$matching2N4.SQ005_SQ003.[i],data4$matching2N4.SQ006_SQ003.[i],data4$matching2N4.SQ007_SQ003.[i],data4$matching2N4.SQ008_SQ003.[i],data4$matching2N4.SQ009_SQ003.[i],data4$matching2N4.SQ010_SQ003.[i]),
    cbind(data4$matching2N4.SQ001_SQ004.[i],data4$matching2N4.SQ002_SQ004.[i],data4$matching2N4.SQ003_SQ004.[i],data4$matching2N4.SQ004_SQ004.[i],data4$matching2N4.SQ005_SQ004.[i],data4$matching2N4.SQ006_SQ004.[i],data4$matching2N4.SQ007_SQ004.[i],data4$matching2N4.SQ008_SQ004.[i],data4$matching2N4.SQ009_SQ004.[i],data4$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data4$matching2N5.SQ001_SQ001.[i],data4$matching2N5.SQ002_SQ001.[i],data4$matching2N5.SQ003_SQ001.[i],data4$matching2N5.SQ004_SQ001.[i],data4$matching2N5.SQ005_SQ001.[i],data4$matching2N5.SQ006_SQ001.[i],data4$matching2N5.SQ007_SQ001.[i],data4$matching2N5.SQ008_SQ001.[i],data4$matching2N5.SQ009_SQ001.[i],data4$matching2N5.SQ010_SQ001.[i]),
    cbind(data4$matching2N5.SQ001_SQ002.[i],data4$matching2N5.SQ002_SQ002.[i],data4$matching2N5.SQ003_SQ002.[i],data4$matching2N5.SQ004_SQ002.[i],data4$matching2N5.SQ005_SQ002.[i],data4$matching2N5.SQ006_SQ002.[i],data4$matching2N5.SQ007_SQ002.[i],data4$matching2N5.SQ008_SQ002.[i],data4$matching2N5.SQ009_SQ002.[i],data4$matching2N5.SQ010_SQ002.[i]),
    cbind(data4$matching2N5.SQ001_SQ003.[i],data4$matching2N5.SQ002_SQ003.[i],data4$matching2N5.SQ003_SQ003.[i],data4$matching2N5.SQ004_SQ003.[i],data4$matching2N5.SQ005_SQ003.[i],data4$matching2N5.SQ006_SQ003.[i],data4$matching2N5.SQ007_SQ003.[i],data4$matching2N5.SQ008_SQ003.[i],data4$matching2N5.SQ009_SQ003.[i],data4$matching2N5.SQ010_SQ003.[i]),
    cbind(data4$matching2N5.SQ001_SQ004.[i],data4$matching2N5.SQ002_SQ004.[i],data4$matching2N5.SQ003_SQ004.[i],data4$matching2N5.SQ004_SQ004.[i],data4$matching2N5.SQ005_SQ004.[i],data4$matching2N5.SQ006_SQ004.[i],data4$matching2N5.SQ007_SQ004.[i],data4$matching2N5.SQ008_SQ004.[i],data4$matching2N5.SQ009_SQ004.[i],data4$matching2N5.SQ010_SQ004.[i]),
    cbind(data4$matching2N5.SQ001_SQ005.[i],data4$matching2N5.SQ002_SQ005.[i],data4$matching2N5.SQ003_SQ005.[i],data4$matching2N5.SQ004_SQ005.[i],data4$matching2N5.SQ005_SQ005.[i],data4$matching2N5.SQ006_SQ005.[i],data4$matching2N5.SQ007_SQ005.[i],data4$matching2N5.SQ008_SQ005.[i],data4$matching2N5.SQ009_SQ005.[i],data4$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data4$matching3N1.SQ001_SQ001.[i],data4$matching3N1.SQ002_SQ001.[i],data4$matching3N1.SQ003_SQ001.[i],data4$matching3N1.SQ004_SQ001.[i],data4$matching3N1.SQ005_SQ001.[i],data4$matching3N1.SQ006_SQ001.[i],data4$matching3N1.SQ007_SQ001.[i],data4$matching3N1.SQ008_SQ001.[i],data4$matching3N1.SQ009_SQ001.[i],data4$matching3N1.SQ010_SQ001.[i], data4$matching3N1.SQ011_SQ001.[i], data4$matching3N1.SQ012_SQ001.[i], data4$matching3N1.SQ013_SQ001.[i], data4$matching3N1.SQ014_SQ001.[i], data4$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data4$matching3N2.SQ001_SQ001.[i],data4$matching3N2.SQ002_SQ001.[i],data4$matching3N2.SQ003_SQ001.[i],data4$matching3N2.SQ004_SQ001.[i],data4$matching3N2.SQ005_SQ001.[i],data4$matching3N2.SQ006_SQ001.[i],data4$matching3N2.SQ007_SQ001.[i],data4$matching3N2.SQ008_SQ001.[i],data4$matching3N2.SQ009_SQ001.[i],data4$matching3N2.SQ010_SQ001.[i], data4$matching3N2.SQ011_SQ001.[i], data4$matching3N2.SQ012_SQ001.[i], data4$matching3N2.SQ013_SQ001.[i], data4$matching3N2.SQ014_SQ001.[i], data4$matching3N2.SQ015_SQ001.[i]),
    cbind(data4$matching3N2.SQ001_SQ002.[i],data4$matching3N2.SQ002_SQ002.[i],data4$matching3N2.SQ003_SQ002.[i],data4$matching3N2.SQ004_SQ002.[i],data4$matching3N2.SQ005_SQ002.[i],data4$matching3N2.SQ006_SQ002.[i],data4$matching3N2.SQ007_SQ002.[i],data4$matching3N2.SQ008_SQ002.[i],data4$matching3N2.SQ009_SQ002.[i],data4$matching3N2.SQ010_SQ002.[i], data4$matching3N2.SQ011_SQ002.[i], data4$matching3N2.SQ012_SQ002.[i], data4$matching3N2.SQ013_SQ002.[i], data4$matching3N2.SQ014_SQ002.[i], data4$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data4$matching3N3.SQ001_SQ001.[i],data4$matching3N3.SQ002_SQ001.[i],data4$matching3N3.SQ003_SQ001.[i],data4$matching3N3.SQ004_SQ001.[i],data4$matching3N3.SQ005_SQ001.[i],data4$matching3N3.SQ006_SQ001.[i],data4$matching3N3.SQ007_SQ001.[i],data4$matching3N3.SQ008_SQ001.[i],data4$matching3N3.SQ009_SQ001.[i],data4$matching3N3.SQ010_SQ001.[i], data4$matching3N3.SQ011_SQ001.[i], data4$matching3N3.SQ012_SQ001.[i], data4$matching3N3.SQ013_SQ001.[i], data4$matching3N3.SQ014_SQ001.[i], data4$matching3N3.SQ015_SQ001.[i]),
    cbind(data4$matching3N3.SQ001_SQ002.[i],data4$matching3N3.SQ002_SQ002.[i],data4$matching3N3.SQ003_SQ002.[i],data4$matching3N3.SQ004_SQ002.[i],data4$matching3N3.SQ005_SQ002.[i],data4$matching3N3.SQ006_SQ002.[i],data4$matching3N3.SQ007_SQ002.[i],data4$matching3N3.SQ008_SQ002.[i],data4$matching3N3.SQ009_SQ002.[i],data4$matching3N3.SQ010_SQ002.[i], data4$matching3N3.SQ011_SQ002.[i], data4$matching3N3.SQ012_SQ002.[i], data4$matching3N3.SQ013_SQ002.[i], data4$matching3N3.SQ014_SQ002.[i], data4$matching3N3.SQ015_SQ002.[i]),
    cbind(data4$matching3N3.SQ001_SQ003.[i],data4$matching3N3.SQ002_SQ003.[i],data4$matching3N3.SQ003_SQ003.[i],data4$matching3N3.SQ004_SQ003.[i],data4$matching3N3.SQ005_SQ003.[i],data4$matching3N3.SQ006_SQ003.[i],data4$matching3N3.SQ007_SQ003.[i],data4$matching3N3.SQ008_SQ003.[i],data4$matching3N3.SQ009_SQ003.[i],data4$matching3N3.SQ010_SQ003.[i], data4$matching3N3.SQ011_SQ003.[i], data4$matching3N3.SQ012_SQ003.[i], data4$matching3N3.SQ013_SQ003.[i], data4$matching3N3.SQ014_SQ003.[i], data4$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data4$matching3N4.SQ001_SQ001.[i],data4$matching3N4.SQ002_SQ001.[i],data4$matching3N4.SQ003_SQ001.[i],data4$matching3N4.SQ004_SQ001.[i],data4$matching3N4.SQ005_SQ001.[i],data4$matching3N4.SQ006_SQ001.[i],data4$matching3N4.SQ007_SQ001.[i],data4$matching3N4.SQ008_SQ001.[i],data4$matching3N4.SQ009_SQ001.[i],data4$matching3N4.SQ010_SQ001.[i], data4$matching3N4.SQ011_SQ001.[i], data4$matching3N4.SQ012_SQ001.[i], data4$matching3N4.SQ013_SQ001.[i], data4$matching3N4.SQ014_SQ001.[i], data4$matching3N4.SQ015_SQ001.[i]),
    cbind(data4$matching3N4.SQ001_SQ002.[i],data4$matching3N4.SQ002_SQ002.[i],data4$matching3N4.SQ003_SQ002.[i],data4$matching3N4.SQ004_SQ002.[i],data4$matching3N4.SQ005_SQ002.[i],data4$matching3N4.SQ006_SQ002.[i],data4$matching3N4.SQ007_SQ002.[i],data4$matching3N4.SQ008_SQ002.[i],data4$matching3N4.SQ009_SQ002.[i],data4$matching3N4.SQ010_SQ002.[i], data4$matching3N4.SQ011_SQ002.[i], data4$matching3N4.SQ012_SQ002.[i], data4$matching3N4.SQ013_SQ002.[i], data4$matching3N4.SQ014_SQ002.[i], data4$matching3N4.SQ015_SQ002.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ003.[i],data4$matching3N4.SQ002_SQ003.[i],data4$matching3N4.SQ003_SQ003.[i],data4$matching3N4.SQ004_SQ003.[i],data4$matching3N4.SQ005_SQ003.[i],data4$matching3N4.SQ006_SQ003.[i],data4$matching3N4.SQ007_SQ003.[i],data4$matching3N4.SQ008_SQ003.[i],data4$matching3N4.SQ009_SQ003.[i],data4$matching3N4.SQ010_SQ003.[i], data4$matching3N4.SQ011_SQ003.[i], data4$matching3N4.SQ012_SQ003.[i], data4$matching3N4.SQ013_SQ003.[i], data4$matching3N4.SQ014_SQ003.[i], data4$matching3N4.SQ015_SQ003.[i]),
    cbind(data4$matching3N4.SQ001_SQ004.[i],data4$matching3N4.SQ002_SQ004.[i],data4$matching3N4.SQ003_SQ004.[i],data4$matching3N4.SQ004_SQ004.[i],data4$matching3N4.SQ005_SQ004.[i],data4$matching3N4.SQ006_SQ004.[i],data4$matching3N4.SQ007_SQ004.[i],data4$matching3N4.SQ008_SQ004.[i],data4$matching3N4.SQ009_SQ004.[i],data4$matching3N4.SQ010_SQ004.[i], data4$matching3N4.SQ011_SQ004.[i], data4$matching3N4.SQ012_SQ004.[i], data4$matching3N4.SQ013_SQ004.[i], data4$matching3N4.SQ014_SQ004.[i], data4$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data4$matching3N5.SQ001_SQ001.[i],data4$matching3N5.SQ002_SQ001.[i],data4$matching3N5.SQ003_SQ001.[i],data4$matching3N5.SQ004_SQ001.[i],data4$matching3N5.SQ005_SQ001.[i],data4$matching3N5.SQ006_SQ001.[i],data4$matching3N5.SQ007_SQ001.[i],data4$matching3N5.SQ008_SQ001.[i],data4$matching3N5.SQ009_SQ001.[i],data4$matching3N5.SQ010_SQ001.[i], data4$matching3N5.SQ011_SQ001.[i], data4$matching3N5.SQ012_SQ001.[i], data4$matching3N5.SQ013_SQ001.[i], data4$matching3N5.SQ014_SQ001.[i], data4$matching3N5.SQ015_SQ001.[i]),
    cbind(data4$matching3N5.SQ001_SQ002.[i],data4$matching3N5.SQ002_SQ002.[i],data4$matching3N5.SQ003_SQ002.[i],data4$matching3N5.SQ004_SQ002.[i],data4$matching3N5.SQ005_SQ002.[i],data4$matching3N5.SQ006_SQ002.[i],data4$matching3N5.SQ007_SQ002.[i],data4$matching3N5.SQ008_SQ002.[i],data4$matching3N5.SQ009_SQ002.[i],data4$matching3N5.SQ010_SQ002.[i], data4$matching3N5.SQ011_SQ002.[i], data4$matching3N5.SQ012_SQ002.[i], data4$matching3N5.SQ013_SQ002.[i], data4$matching3N5.SQ014_SQ002.[i], data4$matching3N5.SQ015_SQ002.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ003.[i],data4$matching3N5.SQ002_SQ003.[i],data4$matching3N5.SQ003_SQ003.[i],data4$matching3N5.SQ004_SQ003.[i],data4$matching3N5.SQ005_SQ003.[i],data4$matching3N5.SQ006_SQ003.[i],data4$matching3N5.SQ007_SQ003.[i],data4$matching3N5.SQ008_SQ003.[i],data4$matching3N5.SQ009_SQ003.[i],data4$matching3N5.SQ010_SQ003.[i], data4$matching3N5.SQ011_SQ003.[i], data4$matching3N5.SQ012_SQ003.[i], data4$matching3N5.SQ013_SQ003.[i], data4$matching3N5.SQ014_SQ003.[i], data4$matching3N5.SQ015_SQ003.[i]),
    cbind(data4$matching3N5.SQ001_SQ005.[i],data4$matching3N5.SQ002_SQ005.[i],data4$matching3N5.SQ003_SQ005.[i],data4$matching3N5.SQ004_SQ005.[i],data4$matching3N5.SQ005_SQ005.[i],data4$matching3N5.SQ006_SQ005.[i],data4$matching3N5.SQ007_SQ005.[i],data4$matching3N5.SQ008_SQ005.[i],data4$matching3N5.SQ009_SQ005.[i],data4$matching3N5.SQ010_SQ005.[i], data4$matching3N5.SQ011_SQ005.[i], data4$matching3N5.SQ012_SQ005.[i], data4$matching3N5.SQ013_SQ005.[i], data4$matching3N5.SQ014_SQ005.[i], data4$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

#i also calculate structural embeddedness (alter-level) per ego-net.
for (i in 1:length(alterL)) { # for ego i

  #get number of names listed in the egonets
  nnames_cdn <- length(which(data4[i,c("egonet1.SQ001.","egonet1.SQ002.","egonet1.SQ003.","egonet1.SQ004.","egonet1.SQ005.")]!=""))
  nnames_study <- length(which(data4[i,c("egonet2.SQ001.","egonet2.SQ002.","egonet2.SQ003.","egonet2.SQ004.","egonet2.SQ005.")]!=""))
  nnames_bff <- length(which(data4[i,c("egonet3.SQ001.","egonet3.SQ002.","egonet3.SQ003.","egonet3.SQ004.","egonet3.SQ005.")]!=""))
  nnames_csn <- length(which(data4[i,c("egonet4.SQ001.","egonet4.SQ002.","egonet4.SQ003.","egonet4.SQ004.","egonet4.SQ005.")]!=""))
  
  #CDN
  
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj1N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj1N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj1N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj1N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj1N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj1N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj1N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj1N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj1N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj1N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj1N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj1N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj1N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj1N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj1N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj1N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj1N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj1N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj1N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj1N5d.SQ001.[i]
  }

  if(nnames_cdn>1) { #we only know alter-alter relations for nnames>1
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_cdn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
  
  #CDN: 1:nnames_CDN
  alterL[[i]]$cdn_embed.t2[1:nnames_cdn] <- embed 
  }
  
  #STUDY

  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj2N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj2N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj2N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj2N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj2N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj2N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj2N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj2N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj2N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj2N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj2N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj2N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj2N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj2N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj2N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj2N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj2N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj2N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj2N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj2N5d.SQ001.[i]
  }

  if(nnames_study>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_study]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$study_embed.t2[6:(5+nnames_study)] <- embed 
  }
  
  # BEST FRIENDS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj3N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj3N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj3N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj3N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj3N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj3N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj3N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj3N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj3N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj3N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj3N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj3N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj3N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj3N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj3N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj3N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj3N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj3N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj3N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj3N5d.SQ001.[i]
  }

  if(nnames_bff>1) { #we only know alter-alter relations for nnames
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_bff]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$bff_embed.t2[11:(10+nnames_bff)] <- embed 
  }

  # SPORTS PARTNERS
  #make list of adjacency matrices
  adjL <- list()
  {
  #2 alters
  adjL[[2]] <- matrix(NA,ncol=2,nrow=2)
  adjL[[2]][1,2] <- adjL[[2]][2,1] <- data4$adj4N2a.SQ001.[i]

  #3 alters
  adjL[[3]] <- matrix(NA,ncol=3,nrow=3)
  adjL[[3]][1,2] <- adjL[[3]][2,1] <- data4$adj4N3a.SQ001.[i]
  adjL[[3]][1,3] <- adjL[[3]][3,1] <- data4$adj4N3a.SQ002.[i]
  adjL[[3]][2,3] <- adjL[[3]][3,2] <- data4$adj4N3b.SQ001.[i]

  #4 alters
  adjL[[4]] <- matrix(NA,ncol=4,nrow=4)
  adjL[[4]][1,2] <- adjL[[4]][2,1] <- data4$adj4N4a.SQ001.[i]
  adjL[[4]][1,3] <- adjL[[4]][3,1] <- data4$adj4N4a.SQ002.[i]
  adjL[[4]][1,4] <- adjL[[4]][4,1] <- data4$adj4N4a.SQ003.[i]
  adjL[[4]][2,3] <- adjL[[4]][3,2] <- data4$adj4N4b.SQ001.[i]
  adjL[[4]][2,4] <- adjL[[4]][4,2] <- data4$adj4N4b.SQ002.[i]
  adjL[[4]][3,4] <- adjL[[4]][4,3] <- data4$adj4N4c.SQ001.[i]

  #5 alters
  adjL[[5]] <- matrix(NA,ncol=5,nrow=5)
  adjL[[5]][1,2] <- adjL[[5]][2,1] <- data4$adj4N5a.SQ001.[i]
  adjL[[5]][1,3] <- adjL[[5]][3,1] <- data4$adj4N5a.SQ002.[i]
  adjL[[5]][1,4] <- adjL[[5]][4,1] <- data4$adj4N5a.SQ003.[i]
  adjL[[5]][1,5] <- adjL[[5]][5,1] <- data4$adj4N5a.SQ004.[i]
  adjL[[5]][2,3] <- adjL[[5]][3,2] <- data4$adj4N5b.SQ001.[i]
  adjL[[5]][2,4] <- adjL[[5]][4,2] <- data4$adj4N5b.SQ002.[i]
  adjL[[5]][2,5] <- adjL[[5]][5,2] <- data4$adj4N5b.SQ003.[i]
  adjL[[5]][3,4] <- adjL[[5]][4,3] <- data4$adj4N5c.SQ001.[i]
  adjL[[5]][3,5] <- adjL[[5]][5,3] <- data4$adj4N5c.SQ002.[i]
  adjL[[5]][4,5] <- adjL[[5]][5,4] <- data4$adj4N5d.SQ001.[i]
  }

  if(nnames_csn>1) { #we only know alter-alter relations for nnames>2
    #take the matrix corresponding the nnames
    mat <- adjL[[nnames_csn]]
    
    #ties that are (very) close are 1
    mat[!is.na(mat)] <- ifelse(mat[!is.na(mat)]=="Erg hecht" | mat[!is.na(mat)]=="Hecht", 1,0)

    #calculate embeddedness for each alter/row:
    embed <- apply(mat, 1, function(row) {
    n_cols <- sum(!is.na(row))
    n_ones <- sum(row == "1", na.rm = TRUE)
    ratio <- n_ones / n_cols
    return(ratio)
    })
    
    alterL[[i]]$csn_embed.t2[16:(15+nnames_csn)] <- embed 
  }
}

#alters could be named in multiple name generators. each element of alterL contains rows corresponding to alters that may be 'duplicates'.
#i match the structural embeddedness measures

for (i in 1:length(alterL)) { #for ego i 
  for (j in 1:nrow(alterL[[i]])) { #for alter j
 
    #if name2 is not empty, alter j was matched to the study network; and more precisely, to the study partner with name1: alterL[[i]]$name2[j]
    alterL[[i]]$study_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name2[j]), 
                                            alterL[[i]]$study_embed.t2[which(alterL[[i]]$name1==alterL[[i]]$name2[j] & !is.na(alterL[[i]]$study_embed.t2))], 
                                            alterL[[i]]$study_embed.t2[j])
    
    #if name3 is not empty, alter j was matched to the friends network; and more precisely, to the friend with name1: alterL[[i]]$name3[j]
    alterL[[i]]$bff_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name3[j]), 
                                          alterL[[i]]$bff_embed.t2[which(alterL[[i]]$name1 == alterL[[i]]$name3[j] & !is.na(alterL[[i]]$bff_embed.t2))], 
                                          alterL[[i]]$bff_embed.t2[j])
    
    #if name4 is not empty, alter j was matched to the sports network; and more precisely, to the sports partner with name1: alterL[[i]]$name4[j]
    alterL[[i]]$csn_embed.t2[j] <- ifelse(!is.na(alterL[[i]]$name4[j]), 
                                          alterL[[i]]$csn_embed.t2[which(alterL[[i]]$name1 == alterL[[i]]$name4[j] & !is.na(alterL[[i]]$csn_embed.t2))],
                                          alterL[[i]]$csn_embed.t2[j] )

  }
}

#while in wave 1, only the unique alters were listed in the gender-name interpreters
#in wave 2, only the NEW alters were listed (since it is a stable characteristic)
#so only the non-matched alters that did not appear in w1
#i asked respondents to 'check' males.

#gender
df_males <- data.frame(p1 = data4$males.SQ001.,p2 = data4$males.SQ002.,p3 = data4$males.SQ003.,p4 = data4$males.SQ004.,p5 = data4$males.SQ005.,p6 = data4$males.SQ006.,p7 = data4$males.SQ007.,p8 = data4$males.SQ008.,p9 = data4$males.SQ009.,p10= data4$males.SQ010.,p11= data4$males.SQ011.,p12= data4$males.SQ012.,p13= data4$males.SQ013.,p14= data4$males.SQ014.,p15= data4$males.SQ015.,p16= data4$males.SQ016.,p17= data4$males.SQ017.,p18= data4$males.SQ018.,p19= data4$males.SQ019.,p20= data4$males.SQ020.)

#kin 
df_kin <- data.frame(p1 = data4$kin.SQ001.,p2 = data4$kin.SQ002.,p3 = data4$kin.SQ003.,p4 = data4$kin.SQ004.,p5 = data4$kin.SQ005.,p6 = data4$kin.SQ006.,p7 = data4$kin.SQ007.,p8 = data4$kin.SQ008.,p9 = data4$kin.SQ009., p10= data4$kin.SQ010., p11= data4$kin.SQ011., p12= data4$kin.SQ012., p13= data4$kin.SQ013., p14= data4$kin.SQ014.,p15= data4$kin.SQ015., p16= data4$kin.SQ016., p17= data4$kin.SQ017., p18= data4$kin.SQ018., p19= data4$kin.SQ019., p20= data4$kin.SQ020.)

#age 
df_age <- data.frame(p1 = data4$age.SQ001.,p2 = data4$age.SQ002.,p3 = data4$age.SQ003.,p4 = data4$age.SQ004.,p5 = data4$age.SQ005.,p6 = data4$age.SQ006.,p7 = data4$age.SQ007.,p8 = data4$age.SQ008.,p9 = data4$age.SQ009., p10= data4$age.SQ010., p11= data4$age.SQ011., p12= data4$age.SQ012., p13= data4$age.SQ013., p14= data4$age.SQ014.,p15= data4$age.SQ015., p16= data4$age.SQ016., p17= data4$age.SQ017., p18= data4$age.SQ018., p19= data4$age.SQ019., p20= data4$age.SQ020.)

#educ 
df_educ <- data.frame(p1 = data4$educ.SQ001.,p2 = data4$educ.SQ002.,p3 = data4$educ.SQ003.,p4 = data4$educ.SQ004.,p5 = data4$educ.SQ005.,p6 = data4$educ.SQ006.,p7 = data4$educ.SQ007.,p8 = data4$educ.SQ008.,p9 = data4$educ.SQ009., p10= data4$educ.SQ010., p11= data4$educ.SQ011., p12= data4$educ.SQ012., p13= data4$educ.SQ013., p14= data4$educ.SQ014.,p15= data4$educ.SQ015., p16= data4$educ.SQ016., p17= data4$educ.SQ017., p18= data4$educ.SQ018., p19= data4$educ.SQ019., p20= data4$educ.SQ020.)

#duration 
df_duration <- data.frame(p1 = data4$duur.SQ001.,p2 = data4$duur.SQ002.,p3 = data4$duur.SQ003.,p4 = data4$duur.SQ004.,p5 = data4$duur.SQ005.,p6 = data4$duur.SQ006.,p7 = data4$duur.SQ007.,p8 = data4$duur.SQ008.,p9 = data4$duur.SQ009.,p10= data4$duur.SQ010.,p11= data4$duur.SQ011.,p12= data4$duur.SQ012.,p13= data4$duur.SQ013.,p14= data4$duur.SQ014.,p15= data4$duur.SQ015.,p16= data4$duur.SQ016.,p17= data4$duur.SQ017.,p18= data4$duur.SQ018.,p19= data4$duur.SQ019.,p20= data4$duur.SQ020.)

#proximity
df_proximity <- data.frame(p1 = data4$prox.SQ001.,p2 = data4$prox.SQ002.,p3 = data4$prox.SQ003.,p4 = data4$prox.SQ004.,p5 = data4$prox.SQ005.,p6 = data4$prox.SQ006.,p7 = data4$prox.SQ007.,p8 = data4$prox.SQ008., p9 = data4$prox.SQ009.,p10= data4$prox.SQ010., p11= data4$prox.SQ011., p12= data4$prox.SQ012., p13= data4$prox.SQ013., p14= data4$prox.SQ014., p15= data4$prox.SQ015., p16= data4$prox.SQ016., p17= data4$prox.SQ017., p18= data4$prox.SQ018., p19= data4$prox.SQ019., p20= data4$prox.SQ020.)

#and dynamic
#communication frequency
df_freq <- data.frame(p1 = data4$freq.SQ001.,p2 = data4$freq.SQ002.,p3 = data4$freq.SQ003.,p4 = data4$freq.SQ004.,p5 = data4$freq.SQ005.,p6 = data4$freq.SQ006.,p7 = data4$freq.SQ007.,p8 = data4$freq.SQ008., p9 = data4$freq.SQ009.,p10= data4$freq.SQ010., p11= data4$freq.SQ011., p12= data4$freq.SQ012., p13= data4$freq.SQ013., p14= data4$freq.SQ014., p15= data4$freq.SQ015., p16= data4$freq.SQ016., p17= data4$freq.SQ017., p18= data4$freq.SQ018., p19= data4$freq.SQ019., p20= data4$freq.SQ020.)

#closeness
df_close <- data.frame(p1 = data4$close.SQ001.,p2 = data4$close.SQ002.,p3 = data4$close.SQ003.,p4 = data4$close.SQ004.,p5 = data4$close.SQ005.,p6 = data4$close.SQ006.,p7 = data4$close.SQ007.,p8 = data4$close.SQ008., p9 = data4$close.SQ009.,p10= data4$close.SQ010., p11= data4$close.SQ011., p12= data4$close.SQ012., p13= data4$close.SQ013., p14= data4$close.SQ014., p15= data4$close.SQ015., p16= data4$close.SQ016., p17= data4$close.SQ017., p18= data4$close.SQ018., p19= data4$close.SQ019., p20= data4$close.SQ020.)

for (i in 1:length(alterL)) {
  
  alterL[[i]]$alter_gender <- ifelse(unlist(df_males[i,], use.names=F)=="Ja", 0, ifelse(unlist(df_males[i,], use.names=F)=="Nee", 1, NA))
  
  alterL[[i]]$alter_age <- 
    ifelse(unlist(df_age[i,], use.names=F)=="Jonger dan 18 jaar", 16, #???
           ifelse(unlist(df_age[i,], use.names=F)=="18 tot 21 jaar", 20,
                  ifelse(unlist(df_age[i,], use.names=F)=="22 tot 25 jaar", 23,
                         ifelse(unlist(df_age[i,], use.names=F)=="26 tot 30 jaar", 28,
                                ifelse(unlist(df_age[i,], use.names=F)=="31 tot 40 jaar", 35,
                                       ifelse(unlist(df_age[i,], use.names=F)=="Ouder dan 40 jaar", 45, #???
                                              ifelse(unlist(df_age[i,], use.names=F)=="Weet ik niet", NA, # i don't know = missing
                                                     unlist(df_age[i,], use.names=F))))))))
  alterL[[i]]$kin <- ifelse(unlist(df_kin[i,], use.names=F)=="Ja", 1, ifelse(unlist(df_kin[i,], use.names=F)=="Nee", 0, ""))
  
  alterL[[i]]$alter_educ <- ifelse(unlist(df_educ[i,],use.names=F)=="lagere school", 1,
                               ifelse(unlist(df_educ[i,],use.names=F)=="vmbo, mavo", 2,
                                      ifelse(unlist(df_educ[i,],use.names=F)=="mbo", 3,
                                         ifelse(unlist(df_educ[i,],use.names=F)=="havo", 4,
                                            ifelse(unlist(df_educ[i,],use.names=F)=="vwo / gymnasium", 5,
                                                 ifelse(unlist(df_educ[i,],use.names=F)=="hbo", 6,
                                                     ifelse(unlist(df_educ[i,],use.names=F)=="universiteit", 7, NA))))))) 
  
  alterL[[i]]$proximity <- ifelse(unlist(df_proximity[i,], use.names=FALSE) == "In hetzelfde huis", "roommate",
                                  ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In dezelfde buurt" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde straat" |
                                    unlist(df_proximity[i,],use.names=F)=="In dezelfde gemeente", "close",
                                    ifelse(unlist(df_proximity[i,], use.names = FALSE) == "In hetzelfde land" |
                                             unlist(df_proximity[i,], use.names = FALSE) == "In een ander land","far", NA)))
  
  
  
  alterL[[i]]$duration <- ifelse(unlist(df_duration[i,],use.names=F)=="Minder dan 1 jaar", 0, 
                               ifelse(unlist(df_duration[i,],use.names=F)=="1 tot 3 jaar", 2,
                                      ifelse(unlist(df_duration[i,],use.names=F)=="4 tot 8 jaar", 6,
                                             ifelse(unlist(df_duration[i,],use.names=F)=="9 tot 15 jaar", 12,
                                                    ifelse(unlist(df_duration[i,],use.names=F)=="Meer dan 15 jaar", 15,NA )))))
  
  alterL[[i]]$frequency.t2 <- ifelse(unlist(df_freq[i,],use.names=F)=="(Bijna) elke dag", 7,
                               ifelse(unlist(df_freq[i,],use.names=F)=="1-2 keer per week",6,
                                      ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per maand",5, 
                                         ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per maand",4, 
                                            ifelse(unlist(df_freq[i,],use.names=F)=="Aantal keer per jaar",3,
                                                 ifelse(unlist(df_freq[i,],use.names=F)=="Ong. 1 keer per jaar",2,
                                                     ifelse(unlist(df_freq[i,],use.names=F)=="Nooit",1, NA )))))))
  alterL[[i]]$closeness.t2 <- ifelse(unlist(df_close[i,],use.names=F)=="Niet hecht", 1, 
                               ifelse(unlist(df_close[i,],use.names=F)=="Enigszins hecht", 2,
                                      ifelse(unlist(df_close[i,],use.names=F)=="Hecht", 3,
                                             ifelse(unlist(df_close[i,],use.names=F)=="Heel erg hecht", 4, NA ))))
}

#match ego-covars (based on 'df'.) 
for (i in 1:length(alterL)) {
  alterL[[i]]$ego_gender <- df$ego_gender[which(df$ego==1)][1]
  alterL[[i]]$ego_age <- df$ego_age[which(df$ego==1)][1]
  alterL[[i]]$ego_educ <- df$ego_educ[which(df$ego==1)][1]
}

#sameness
for (i in 1:length(alterL))  { # for ego i
  # get attributes of ego
  agei <- alterL[[i]]$ego_age[1]
  genderi <- alterL[[i]]$ego_gender[1]
  edi <- alterL[[i]]$ego_educ[1]
  
  for (j in 1:max(alterL[[i]]$alterid)) { # for alter j
    # calculate "same gender" (0/1)
    genderj <- as.numeric(alterL[[i]]$alter_gender[j]) # get alter j gender
    same <- ifelse(genderi==genderj, 1, 0)
    alterL[[i]]$same_gender[which(alterL[[i]]$alterid==j)] <- same
    
    #same education
    eduj <- alterL[[i]]$alter_educ[j]
    same <- ifelse(eduj==edi, 1, 0)
    alterL[[i]]$sim_educ[which(alterL[[i]]$alterid==j)] <- same
    
    # calculate similarity for age as the absolute difference between alter and ego value
    #get alter j attributes
    agej <- as.numeric(alterL[[i]]$alter_age[j])
    #difference score
    difage <- abs(agei - agej)
    
    # so higher values represent *less* similarity
    if( !is.na (agej) ) { # age similarity only calculable if z_j is known!
        alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- difage
      } else { alterL[[i]]$dif_age[which(alterL[[i]]$alterid==j)] <- NA}
  }
}
  
#in which egonets did altesr appear at t2?
for (i in 1:length(alterL))  { 
  for (j in 1:nrow(alterL[[i]])) {
    
    #get name(s) of j
    alter <- unlist(alterL[[i]][j,c(5:8)], use.names=FALSE)
    alter <- alter[!is.na(alter)] # exclude NAs
    
    #find out alter j appears in the 4 egonets
    cdn <- alter %in% net1[i,]
    study <- alter %in% net2[i,]
    bff <- alter %in% net3[i,]
    csn <- alter %in% net4[i,]
    
    # and if so, give alter j score 1 on indicators; 0 otherwise
    alterL[[i]]$cdn2[j] <- ifelse("TRUE" %in% cdn, 1, 0)
    alterL[[i]]$study2[j] <- ifelse("TRUE" %in% study, 1, 0)
    alterL[[i]]$bff2[j] <- ifelse("TRUE" %in% bff, 1, 0)
    alterL[[i]]$csn2[j] <- ifelse("TRUE" %in% csn, 1, 0)
  }
}

#calculate multiplexity as an alter/tie attribute (i.e., number of *additional* networks j appeared in)
for (i in 1:length(alterL))  {
  for (j in 1:nrow(alterL[[i]])) {
    alterL[[i]]$multiplex.t2[j] <- rowSums(alterL[[i]][j,c(24:27)]) - 1 #minus one to reach a meaningful intercept
  }
}

# i filter out the maintained alters
# by excluding alters with empty strings for gender attribute
for ( i in 1:length(alterL)) {
  alterL[[i]] <- alterL[[i]][which(!is.na(alterL[[i]]$alter_gender) & alterL[[i]]$alter_gender!=""),]
  
    if (nrow(alterL[[i]])>0){ #if no. of created alters > 0
      # replace the alter id: 1 : no. unique alters
      alterL[[i]]$alterid <- 1:nrow(alterL[[i]])
  }
}

#based on this, i make a long dataframe with alters nested in ego.
#first, add an ego_id
for (i in 1:length(alterL)) {
  if(nrow(alterL[[i]]>0)) {
    alterL[[i]]$ego <- i
    alterL[[i]]$respnr <- data$respnr[i]
    }
}

#combine using rbind
df <- do.call(rbind,alterL)

#add network variables; match based on df_maintained
df$cdn.size2 <- df$study.size2 <- df$bff.size2 <- df$csn.size2 <- NA
df$cdn.density2 <- df$study.density2 <- df$bff.density2 <- df$csn.density2 <- NA

for ( i in unique(df$ego)) { 
  df$cdn.size2[which(df$ego == i)] <- df_maintained$cdn.size2[which(df_maintained$ego == i)][1]
  df$study.size2[which(df$ego == i)] <- df_maintained$study.size2[which(df_maintained$ego == i)][1]
  df$bff.size2[which(df$ego == i)] <- df_maintained$bff.size2[which(df_maintained$ego == i)][1]
  df$csn.size2[which(df$ego == i)] <- df_maintained$csn.size2[which(df_maintained$ego == i)][1]
  
  df$cdn.density2[which(df$ego == i)] <- df_maintained$cdn.density2[which(df_maintained$ego == i)][1]
  df$study.density2[which(df$ego == i)] <- df_maintained$study.density2[which(df_maintained$ego == i)][1]
  df$bff.density2[which(df$ego == i)] <- df_maintained$bff.density2[which(df_maintained$ego == i)][1]
  df$csn.density2[which(df$ego == i)] <- df_maintained$csn.density2[which(df_maintained$ego == i)][1]
}

#methodological controls??

df$statusW2 <- "Created"
df_created <- df

#some empty name generator entries were saved, due to a bug.
#empties <- which(is.na(df$name1)) #done using the data with names instead of alter_ids
#save(empties,file="./data/empty_entries_indicators.Rda")
empties <- fload("./data_shared/empty_entries_indicators.Rda")

df_created <- df_created[-empties,]
```


```{r, eval=FALSE}
#bind them together
df_alters <- dplyr::bind_rows(df_maintained,df_created)

#arrange, by ego, and status
df_alters <- df_alters %>% arrange(ego, factor(statusW2, levels=c("Maintained", "Created", "Dropped")))
row.names(df_alters) <- 1:nrow(df_alters)
```


<br>

## wave 2 --> wave 3

Let's see whether wave 2 alters were maintained into wave 3. 

```{r, eval = FALSE}
#first subset from `df_alters` w2-maintained or w2-created alters
test <- df_alters[which(df_alters$statusW2=="Maintained" | df_alters$statusW2=="Created"),]

#make variable indicating whether alter was (re-)named in w3
test$surviveW3 <- NA

#some alters had no 'chance' to reappear, because ego did fill out the last survey
test$w3participation <- NA
for (i in unique(test$respnr)) {
  test$w3participation[which(test$respnr==i)] <- ifelse(i %in% data5$respnr,1,0)
}

#fix alter ids, 1:N
for (i in unique(test$respnr)) {
  test$alterid[which(test$respnr==i)] <- 1:length(test$alterid[which(test$respnr==i)])
}

#subset wave 5 matching matrix (matching wave 3 alters to wave 2 alters - either created or maintained )
#also add respnr
w1w2 <- data5[,c(486:885, ncol(data5))]

for ( i in unique(test$respnr)) { # for all egos (both those that participated in w3, and those that didnt)
  
    matchingL <- vector("list", 20) #pre-allocate empty list of length 20, to store matching matrices
  {
    matchingL[[1]] <- w1w2[which(w1w2$respnr==i), 1:20]
    matchingL[[2]] <- w1w2[which(w1w2$respnr==i),21:40]
    matchingL[[3]] <- w1w2[which(w1w2$respnr==i),41:60]
    matchingL[[4]] <- w1w2[which(w1w2$respnr==i),61:80]
    matchingL[[5]] <- w1w2[which(w1w2$respnr==i),81:100]
    matchingL[[6]] <- w1w2[which(w1w2$respnr==i),101:120]
    matchingL[[7]] <- w1w2[which(w1w2$respnr==i),121:140]
    matchingL[[8]] <- w1w2[which(w1w2$respnr==i),141:160]
    matchingL[[9]] <- w1w2[which(w1w2$respnr==i),161:180]
    matchingL[[10]] <- w1w2[which(w1w2$respnr==i),181:200]
    matchingL[[11]] <- w1w2[which(w1w2$respnr==i),201:220]
    matchingL[[12]] <- w1w2[which(w1w2$respnr==i),221:240]
    matchingL[[13]] <- w1w2[which(w1w2$respnr==i),241:260]
    matchingL[[14]] <- w1w2[which(w1w2$respnr==i),261:280]
    matchingL[[15]] <- w1w2[which(w1w2$respnr==i),281:300]
    matchingL[[16]] <- w1w2[which(w1w2$respnr==i),301:320]
    matchingL[[17]] <- w1w2[which(w1w2$respnr==i),321:340]
    matchingL[[18]] <- w1w2[which(w1w2$respnr==i),341:360]
    matchingL[[19]] <- w1w2[which(w1w2$respnr==i),361:380]
    matchingL[[20]] <- w1w2[which(w1w2$respnr==i),381:400]
   }
   #find the 'right' matching matrix in this list, by taking the one that contains answers
  ind <- NULL
  for (j in seq_along(matchingL)) {
    #check along the sequence of elements j in the matching matrix list if they are non-empty and not NA
    check <- FALSE
    for (col in matchingL[[j]]) {
      if (!all(is.na(col)) && any(nchar(col) > 0)) {
        check <- TRUE
        break  
      }
      }
    if (check) {
      ind <- j
      break  
    }
  }
  if(length(ind)>0) {
    # get the  matrix
    mm <- matchingL[[ind]]
    #unlist the answers given
    ans <- unlist(mm, use.names=FALSE)
    #use `stringr` to extract numbers in these answers
    id <- unlist(stringr::str_extract_all(ans, "\\(?[0-9,.]+\\)?"))
    id <- id[!is.na(id)] #remove NA
    
    for (k in 1:length(which(test$respnr==i))) {
      #for w2-alters of ego i, with ids 1:k,
      #if alter was renamed, give score 1,
      #if he/she was not renamed give 0,
      #alters who had no chance to reappear due to ego not filling out the survey keep NA.
      test$surviveW3[which(test$respnr == i & test$alterid == k)] <- ifelse(test$alterid[ which(test$respnr == i & test$alterid == k)] %in% id, 1,0)}
  }
}

# remaining NAs on `survivew3` among egos who did fill out w3, are due to ego not listing any alters..
#any(test$w3participation==1 & is.na(test$surviveW3))
#test$respnr[which(test$w3participation==1 & is.na(test$surviveW3))]
#so, set to 0...
test$surviveW3[which(is.na(test$surviveW3) & test$w3participation == 1)] <- 0

# in which egonet did they (re)appear?
test$cdn3 <- NA
test$study3 <- NA
test$bff3 <- NA
test$csn3 <- NA

# i use the matching matrices of w3; where columns refer to unique w2 alters (maintained and created; those of our constructed dataframe) and rows referring to unique w3 alters
# so, if w2-alter/column j is matched to w3-alter/row i, then w2-alter did survive (but we already knew that); but more importantly, j was the i^th alter mentioned
# and this allows me to infer in what network j (re-)appeared
# since cdn = 1-5; study = 6-10; bff = 11-15; csn = 16-20.
# however, since the rows only include non-duplicate w3-alters;
# if alter j at t2 was named more than once at t3, i only know the first
# egonet in which j was named...
# thus, before i proceed, i use the matching matrices that allowed ego to match alters from different egonets to make alter ids for w3 alters; and figure out to which egonets they belonged

#first, subset data5, bc we only want names data of egos with w1/w2 alters who participated in w3
data5a <- data5[which(data5$respnr %in% unique(test$respnr[which(test$w3participation==1)])),]

# i make a dataframe of w3-alters, with potential duplicates...
df_names <- data.frame(p1 = data5a$egonet1.SQ001.,p2 = data5a$egonet1.SQ002.,p3 = data5a$egonet1.SQ003.,p4 = data5a$egonet1.SQ004.,p5 = data5a$egonet1.SQ005.,p6 = data5a$egonet2.SQ001.,p7 = data5a$egonet2.SQ002.,p8 = data5a$egonet2.SQ003.,p9 = data5a$egonet2.SQ004.,p10= data5a$egonet2.SQ005.,p11= data5a$egonet3.SQ001.,p12= data5a$egonet3.SQ002.,p13= data5a$egonet3.SQ003.,p14= data5a$egonet3.SQ004.,p15= data5a$egonet3.SQ005.,p16= data5a$egonet4.SQ001., p17= data5a$egonet4.SQ002., p18= data5a$egonet4.SQ003.,p19= data5a$egonet4.SQ004.,p20= data5a$egonet4.SQ005.)

# list of dataframes per ego with rows reflecting alters
# and columns indicating the name(s) of the particular alters
alterL <- list()
# loop over all egos, that filled out w3
for ( i in 1:nrow(data5a)) {
  alterL[[i]] <- data.frame(
    alterid = 1:20,name1 = NA,name2 = NA, name3 = NA,name4 = NA)}

# fill the names based on the names data-frame
for ( i in 1:length(alterL)) {
  alterL[[i]]$name1 <- unlist(df_names[i, ], use.names=FALSE)
  alterL[[i]]$name1 <- ifelse(alterL[[i]]$name1=="", NA, alterL[[i]]$name1)
  }

# calculate netsize for each net to get the correct matching matrix
{
  net1 <- cbind(data5a$egonet1.SQ001.,data5a$egonet1.SQ002., data5a$egonet1.SQ003.,data5a$egonet1.SQ004., data5a$egonet1.SQ005.)
  net1 <- ifelse(net1=="", NA, net1)
  ns1 <- vector()
  for (i in 1:nrow(net1)) {
    ns1[i] <- length(net1[i,][which(!is.na(net1[i,]))])
  }
  net2 <- cbind(data5a$egonet2.SQ001.,data5a$egonet2.SQ002., data5a$egonet2.SQ003.,data5a$egonet2.SQ004., data5a$egonet2.SQ005.)
  net2 <- ifelse(net2=="", NA, net2)
  ns2 <- vector()
  for (i in 1:nrow(net2)) {
    ns2[i] <- length(net2[i,][which(!is.na(net2[i,]))])
  }
  net3 <- cbind(data5a$egonet3.SQ001.,data5a$egonet3.SQ002., data5a$egonet3.SQ003.,data5a$egonet3.SQ004., data5a$egonet3.SQ005.)
  net3 <- ifelse(net3=="", NA, net3)
  ns3 <- vector()
  for (i in 1:nrow(net3)) {
    ns3[i] <- length(net3[i,][which(!is.na(net3[i,]))])
  }
  net4 <- cbind(data5a$egonet4.SQ001.,data5a$egonet4.SQ002., data5a$egonet4.SQ003.,data5a$egonet4.SQ004., data5a$egonet4.SQ005.)
  net4 <- ifelse(net4=="", NA, net4)
  ns4 <- vector()
  for (i in 1:nrow(net4)) {
    ns4[i] <- length(net4[i,][which(!is.na(net4[i,]))])
  }
}
# construct the matching matrices list for egonet1-2
matchingList <- list()
for (i in 1:length(alterL)) {
  matchingL <- list()
  matchingL[[1]] <- cbind(data5a$matching1N1.SQ001_SQ001.[i],data5a$matching1N1.SQ002_SQ001.[i],data5a$matching1N1.SQ003_SQ001.[i],data5a$matching1N1.SQ004_SQ001.[i],data5a$matching1N1.SQ005_SQ001.[i])
  matchingL[[2]]<- rbind(
    cbind(data5a$matching1N2.SQ001_SQ001.[i], data5a$matching1N2.SQ002_SQ001.[i], data5a$matching1N2.SQ003_SQ001.[i], data5a$matching1N2.SQ004_SQ001.[i], data5a$matching1N2.SQ005_SQ001.[i]),
    cbind(data5a$matching1N2.SQ001_SQ002.[i], data5a$matching1N2.SQ002_SQ002.[i], data5a$matching1N2.SQ003_SQ002.[i], data5a$matching1N2.SQ004_SQ002.[i], data5a$matching1N2.SQ005_SQ002.[i]))
  matchingL[[3]]<- rbind(
    cbind(data5a$matching1N3.SQ001_SQ001.[i], data5a$matching1N3.SQ002_SQ001.[i], data5a$matching1N3.SQ003_SQ001.[i], data5a$matching1N3.SQ004_SQ001.[i], data5a$matching1N3.SQ005_SQ001.[i]),
    cbind(data5a$matching1N3.SQ001_SQ002.[i], data5a$matching1N3.SQ002_SQ002.[i], data5a$matching1N3.SQ003_SQ002.[i], data5a$matching1N3.SQ004_SQ002.[i], data5a$matching1N3.SQ005_SQ002.[i]),
    cbind(data5a$matching1N3.SQ001_SQ003.[i], data5a$matching1N3.SQ002_SQ003.[i], data5a$matching1N3.SQ003_SQ003.[i], data5a$matching1N3.SQ004_SQ003.[i], data5a$matching1N3.SQ005_SQ003.[i]))
  matchingL[[4]]<- rbind(
    cbind(data5a$matching1N4.SQ001_SQ001.[i], data5a$matching1N4.SQ002_SQ001.[i], data5a$matching1N4.SQ003_SQ001.[i], data5a$matching1N4.SQ004_SQ001.[i], data5a$matching1N4.SQ005_SQ001.[i]),
    cbind(data5a$matching1N4.SQ001_SQ002.[i], data5a$matching1N4.SQ002_SQ002.[i], data5a$matching1N4.SQ003_SQ002.[i], data5a$matching1N4.SQ004_SQ002.[i], data5a$matching1N4.SQ005_SQ002.[i]),
    cbind(data5a$matching1N4.SQ001_SQ003.[i], data5a$matching1N4.SQ002_SQ003.[i], data5a$matching1N4.SQ003_SQ003.[i], data5a$matching1N4.SQ004_SQ003.[i], data5a$matching1N4.SQ005_SQ003.[i]),
    cbind(data5a$matching1N4.SQ001_SQ004.[i], data5a$matching1N4.SQ002_SQ004.[i], data5a$matching1N4.SQ003_SQ004.[i], data5a$matching1N4.SQ004_SQ004.[i], data5a$matching1N4.SQ005_SQ004.[i]))
  matchingL[[5]]<- rbind(
    cbind(data5a$matching1N5.SQ001_SQ001.[i], data5a$matching1N5.SQ002_SQ001.[i], data5a$matching1N5.SQ003_SQ001.[i], data5a$matching1N5.SQ004_SQ001.[i], data5a$matching1N5.SQ005_SQ001.[i]),
    cbind(data5a$matching1N5.SQ001_SQ002.[i], data5a$matching1N5.SQ002_SQ002.[i], data5a$matching1N5.SQ003_SQ002.[i], data5a$matching1N5.SQ004_SQ002.[i], data5a$matching1N5.SQ005_SQ002.[i]),
    cbind(data5a$matching1N5.SQ001_SQ003.[i], data5a$matching1N5.SQ002_SQ003.[i], data5a$matching1N5.SQ003_SQ003.[i], data5a$matching1N5.SQ004_SQ003.[i], data5a$matching1N5.SQ005_SQ003.[i]),
    cbind(data5a$matching1N5.SQ001_SQ004.[i], data5a$matching1N5.SQ002_SQ004.[i], data5a$matching1N5.SQ003_SQ004.[i], data5a$matching1N5.SQ004_SQ004.[i], data5a$matching1N5.SQ005_SQ004.[i]),
    cbind(data5a$matching1N5.SQ001_SQ005.[i], data5a$matching1N5.SQ002_SQ005.[i], data5a$matching1N5.SQ003_SQ005.[i], data5a$matching1N5.SQ004_SQ005.[i], data5a$matching1N5.SQ005_SQ005.[i]))
  matchingList[[i]] <- matchingL
} # so... matchingL[[1]][[5]] is matchingmatrix 5 (i.e., 5 alters named in egonet2) for ego 1

# the combination of the matching matrices and the netsizes for ego allows me to match the names myself.
for (i in 1:length(matchingList)) {     # for ego i
  mL <- matchingList[[i]]               # get the matching list
  ns <- ns2[[i]]                        # get the size of egonet2
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # retrieve the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net2[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name2[which(alterL[[i]]$alterid==matched[,2])] <- net[matched[,1]]
    }
  }
}#ignore warning

# again, make a matching list for the second matching matrices
# (i.e., matching egonet3 alters to prev. alters);
matchingList2 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching2L <- list()
  matching2L[[1]] <- cbind(data5a$matching2N1.SQ001_SQ001.[i],data5a$matching2N1.SQ002_SQ001.[i],data5a$matching2N1.SQ003_SQ001.[i],data5a$matching2N1.SQ004_SQ001.[i],data5a$matching2N1.SQ005_SQ001.[i],data5a$matching2N1.SQ006_SQ001.[i],data5a$matching2N1.SQ007_SQ001.[i],data5a$matching2N1.SQ008_SQ001.[i],data5a$matching2N1.SQ009_SQ001.[i],data5a$matching2N1.SQ010_SQ001.[i])
  matching2L[[2]] <- rbind(
    cbind(data5a$matching2N2.SQ001_SQ001.[i],data5a$matching2N2.SQ002_SQ001.[i],data5a$matching2N2.SQ003_SQ001.[i],data5a$matching2N2.SQ004_SQ001.[i],data5a$matching2N2.SQ005_SQ001.[i],data5a$matching2N2.SQ006_SQ001.[i],data5a$matching2N2.SQ007_SQ001.[i],data5a$matching2N2.SQ008_SQ001.[i],data5a$matching2N2.SQ009_SQ001.[i],data5a$matching2N2.SQ010_SQ001.[i]),
    cbind(data5a$matching2N2.SQ001_SQ002.[i],data5a$matching2N2.SQ002_SQ002.[i],data5a$matching2N2.SQ003_SQ002.[i],data5a$matching2N2.SQ004_SQ002.[i],data5a$matching2N2.SQ005_SQ002.[i],data5a$matching2N2.SQ006_SQ002.[i],data5a$matching2N2.SQ007_SQ002.[i],data5a$matching2N2.SQ008_SQ002.[i],data5a$matching2N2.SQ009_SQ002.[i],data5a$matching2N2.SQ010_SQ002.[i]))
  matching2L[[3]] <- rbind(
    cbind(data5a$matching2N3.SQ001_SQ001.[i],data5a$matching2N3.SQ002_SQ001.[i],data5a$matching2N3.SQ003_SQ001.[i],data5a$matching2N3.SQ004_SQ001.[i],data5a$matching2N3.SQ005_SQ001.[i],data5a$matching2N3.SQ006_SQ001.[i],data5a$matching2N3.SQ007_SQ001.[i],data5a$matching2N3.SQ008_SQ001.[i],data5a$matching2N3.SQ009_SQ001.[i],data5a$matching2N3.SQ010_SQ001.[i]),
    cbind(data5a$matching2N3.SQ001_SQ002.[i],data5a$matching2N3.SQ002_SQ002.[i],data5a$matching2N3.SQ003_SQ002.[i],data5a$matching2N3.SQ004_SQ002.[i],data5a$matching2N3.SQ005_SQ002.[i],data5a$matching2N3.SQ006_SQ002.[i],data5a$matching2N3.SQ007_SQ002.[i],data5a$matching2N3.SQ008_SQ002.[i],data5a$matching2N3.SQ009_SQ002.[i],data5a$matching2N3.SQ010_SQ002.[i]),
    cbind(data5a$matching2N3.SQ001_SQ003.[i],data5a$matching2N3.SQ002_SQ003.[i],data5a$matching2N3.SQ003_SQ003.[i],data5a$matching2N3.SQ004_SQ003.[i],data5a$matching2N3.SQ005_SQ003.[i],data5a$matching2N3.SQ006_SQ003.[i],data5a$matching2N3.SQ007_SQ003.[i],data5a$matching2N3.SQ008_SQ003.[i],data5a$matching2N3.SQ009_SQ003.[i],data5a$matching2N3.SQ010_SQ003.[i]))
  matching2L[[4]] <- rbind(
    cbind(data5a$matching2N4.SQ001_SQ001.[i],data5a$matching2N4.SQ002_SQ001.[i],data5a$matching2N4.SQ003_SQ001.[i],data5a$matching2N4.SQ004_SQ001.[i],data5a$matching2N4.SQ005_SQ001.[i],data5a$matching2N4.SQ006_SQ001.[i],data5a$matching2N4.SQ007_SQ001.[i],data5a$matching2N4.SQ008_SQ001.[i],data5a$matching2N4.SQ009_SQ001.[i],data5a$matching2N4.SQ010_SQ001.[i]),
    cbind(data5a$matching2N4.SQ001_SQ002.[i],data5a$matching2N4.SQ002_SQ002.[i],data5a$matching2N4.SQ003_SQ002.[i],data5a$matching2N4.SQ004_SQ002.[i],data5a$matching2N4.SQ005_SQ002.[i],data5a$matching2N4.SQ006_SQ002.[i],data5a$matching2N4.SQ007_SQ002.[i],data5a$matching2N4.SQ008_SQ002.[i],data5a$matching2N4.SQ009_SQ002.[i],data5a$matching2N4.SQ010_SQ002.[i]),
    cbind(data5a$matching2N4.SQ001_SQ003.[i],data5a$matching2N4.SQ002_SQ003.[i],data5a$matching2N4.SQ003_SQ003.[i],data5a$matching2N4.SQ004_SQ003.[i],data5a$matching2N4.SQ005_SQ003.[i],data5a$matching2N4.SQ006_SQ003.[i],data5a$matching2N4.SQ007_SQ003.[i],data5a$matching2N4.SQ008_SQ003.[i],data5a$matching2N4.SQ009_SQ003.[i],data5a$matching2N4.SQ010_SQ003.[i]),
    cbind(data5a$matching2N4.SQ001_SQ004.[i],data5a$matching2N4.SQ002_SQ004.[i],data5a$matching2N4.SQ003_SQ004.[i],data5a$matching2N4.SQ004_SQ004.[i],data5a$matching2N4.SQ005_SQ004.[i],data5a$matching2N4.SQ006_SQ004.[i],data5a$matching2N4.SQ007_SQ004.[i],data5a$matching2N4.SQ008_SQ004.[i],data5a$matching2N4.SQ009_SQ004.[i],data5a$matching2N4.SQ010_SQ004.[i]))
  matching2L[[5]] <- rbind(
    cbind(data5a$matching2N5.SQ001_SQ001.[i],data5a$matching2N5.SQ002_SQ001.[i],data5a$matching2N5.SQ003_SQ001.[i],data5a$matching2N5.SQ004_SQ001.[i],data5a$matching2N5.SQ005_SQ001.[i],data5a$matching2N5.SQ006_SQ001.[i],data5a$matching2N5.SQ007_SQ001.[i],data5a$matching2N5.SQ008_SQ001.[i],data5a$matching2N5.SQ009_SQ001.[i],data5a$matching2N5.SQ010_SQ001.[i]),
    cbind(data5a$matching2N5.SQ001_SQ002.[i],data5a$matching2N5.SQ002_SQ002.[i],data5a$matching2N5.SQ003_SQ002.[i],data5a$matching2N5.SQ004_SQ002.[i],data5a$matching2N5.SQ005_SQ002.[i],data5a$matching2N5.SQ006_SQ002.[i],data5a$matching2N5.SQ007_SQ002.[i],data5a$matching2N5.SQ008_SQ002.[i],data5a$matching2N5.SQ009_SQ002.[i],data5a$matching2N5.SQ010_SQ002.[i]),
    cbind(data5a$matching2N5.SQ001_SQ003.[i],data5a$matching2N5.SQ002_SQ003.[i],data5a$matching2N5.SQ003_SQ003.[i],data5a$matching2N5.SQ004_SQ003.[i],data5a$matching2N5.SQ005_SQ003.[i],data5a$matching2N5.SQ006_SQ003.[i],data5a$matching2N5.SQ007_SQ003.[i],data5a$matching2N5.SQ008_SQ003.[i],data5a$matching2N5.SQ009_SQ003.[i],data5a$matching2N5.SQ010_SQ003.[i]),
    cbind(data5a$matching2N5.SQ001_SQ004.[i],data5a$matching2N5.SQ002_SQ004.[i],data5a$matching2N5.SQ003_SQ004.[i],data5a$matching2N5.SQ004_SQ004.[i],data5a$matching2N5.SQ005_SQ004.[i],data5a$matching2N5.SQ006_SQ004.[i],data5a$matching2N5.SQ007_SQ004.[i],data5a$matching2N5.SQ008_SQ004.[i],data5a$matching2N5.SQ009_SQ004.[i],data5a$matching2N5.SQ010_SQ004.[i]),
    cbind(data5a$matching2N5.SQ001_SQ005.[i],data5a$matching2N5.SQ002_SQ005.[i],data5a$matching2N5.SQ003_SQ005.[i],data5a$matching2N5.SQ004_SQ005.[i],data5a$matching2N5.SQ005_SQ005.[i],data5a$matching2N5.SQ006_SQ005.[i],data5a$matching2N5.SQ007_SQ005.[i],data5a$matching2N5.SQ008_SQ005.[i],data5a$matching2N5.SQ009_SQ005.[i],data5a$matching2N5.SQ010_SQ005.[i]))
  matchingList2[[i]] <- matching2L
}

for (i in 1:length(matchingList2)) {    # for ego i
  mL <- matchingList2[[i]]              # get the matching list 2
  ns <- ns3[[i]]                        # get the size of egonet3
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net3[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name3[matched[,2]] <- net[matched[,1]]
    }
  }
}

# same for matching 3 (i.e., egonet4 with egonets 1-3)
matchingList3 <- list()
for (i in 1:length(alterL)) { # for ego i
  matching3L <- list()
  matching3L[[1]] <- cbind(data5a$matching3N1.SQ001_SQ001.[i],data5a$matching3N1.SQ002_SQ001.[i],data5a$matching3N1.SQ003_SQ001.[i],data5a$matching3N1.SQ004_SQ001.[i],data5a$matching3N1.SQ005_SQ001.[i],data5a$matching3N1.SQ006_SQ001.[i],data5a$matching3N1.SQ007_SQ001.[i],data5a$matching3N1.SQ008_SQ001.[i],data5a$matching3N1.SQ009_SQ001.[i],data5a$matching3N1.SQ010_SQ001.[i], data5a$matching3N1.SQ011_SQ001.[i], data5a$matching3N1.SQ012_SQ001.[i], data5a$matching3N1.SQ013_SQ001.[i], data5a$matching3N1.SQ014_SQ001.[i], data5a$matching3N1.SQ015_SQ001.[i])
  matching3L[[2]] <- rbind(
    cbind(data5a$matching3N2.SQ001_SQ001.[i],data5a$matching3N2.SQ002_SQ001.[i],data5a$matching3N2.SQ003_SQ001.[i],data5a$matching3N2.SQ004_SQ001.[i],data5a$matching3N2.SQ005_SQ001.[i],data5a$matching3N2.SQ006_SQ001.[i],data5a$matching3N2.SQ007_SQ001.[i],data5a$matching3N2.SQ008_SQ001.[i],data5a$matching3N2.SQ009_SQ001.[i],data5a$matching3N2.SQ010_SQ001.[i], data5a$matching3N2.SQ011_SQ001.[i], data5a$matching3N2.SQ012_SQ001.[i], data5a$matching3N2.SQ013_SQ001.[i], data5a$matching3N2.SQ014_SQ001.[i], data5a$matching3N2.SQ015_SQ001.[i]),
    cbind(data5a$matching3N2.SQ001_SQ002.[i],data5a$matching3N2.SQ002_SQ002.[i],data5a$matching3N2.SQ003_SQ002.[i],data5a$matching3N2.SQ004_SQ002.[i],data5a$matching3N2.SQ005_SQ002.[i],data5a$matching3N2.SQ006_SQ002.[i],data5a$matching3N2.SQ007_SQ002.[i],data5a$matching3N2.SQ008_SQ002.[i],data5a$matching3N2.SQ009_SQ002.[i],data5a$matching3N2.SQ010_SQ002.[i], data5a$matching3N2.SQ011_SQ002.[i], data5a$matching3N2.SQ012_SQ002.[i], data5a$matching3N2.SQ013_SQ002.[i], data5a$matching3N2.SQ014_SQ002.[i], data5a$matching3N2.SQ015_SQ002.[i]))
  matching3L[[3]] <- rbind(
    cbind(data5a$matching3N3.SQ001_SQ001.[i],data5a$matching3N3.SQ002_SQ001.[i],data5a$matching3N3.SQ003_SQ001.[i],data5a$matching3N3.SQ004_SQ001.[i],data5a$matching3N3.SQ005_SQ001.[i],data5a$matching3N3.SQ006_SQ001.[i],data5a$matching3N3.SQ007_SQ001.[i],data5a$matching3N3.SQ008_SQ001.[i],data5a$matching3N3.SQ009_SQ001.[i],data5a$matching3N3.SQ010_SQ001.[i], data5a$matching3N3.SQ011_SQ001.[i], data5a$matching3N3.SQ012_SQ001.[i], data5a$matching3N3.SQ013_SQ001.[i], data5a$matching3N3.SQ014_SQ001.[i], data5a$matching3N3.SQ015_SQ001.[i]),
    cbind(data5a$matching3N3.SQ001_SQ002.[i],data5a$matching3N3.SQ002_SQ002.[i],data5a$matching3N3.SQ003_SQ002.[i],data5a$matching3N3.SQ004_SQ002.[i],data5a$matching3N3.SQ005_SQ002.[i],data5a$matching3N3.SQ006_SQ002.[i],data5a$matching3N3.SQ007_SQ002.[i],data5a$matching3N3.SQ008_SQ002.[i],data5a$matching3N3.SQ009_SQ002.[i],data5a$matching3N3.SQ010_SQ002.[i], data5a$matching3N3.SQ011_SQ002.[i], data5a$matching3N3.SQ012_SQ002.[i], data5a$matching3N3.SQ013_SQ002.[i], data5a$matching3N3.SQ014_SQ002.[i], data5a$matching3N3.SQ015_SQ002.[i]),
    cbind(data5a$matching3N3.SQ001_SQ003.[i],data5a$matching3N3.SQ002_SQ003.[i],data5a$matching3N3.SQ003_SQ003.[i],data5a$matching3N3.SQ004_SQ003.[i],data5a$matching3N3.SQ005_SQ003.[i],data5a$matching3N3.SQ006_SQ003.[i],data5a$matching3N3.SQ007_SQ003.[i],data5a$matching3N3.SQ008_SQ003.[i],data5a$matching3N3.SQ009_SQ003.[i],data5a$matching3N3.SQ010_SQ003.[i], data5a$matching3N3.SQ011_SQ003.[i], data5a$matching3N3.SQ012_SQ003.[i], data5a$matching3N3.SQ013_SQ003.[i], data5a$matching3N3.SQ014_SQ003.[i], data5a$matching3N3.SQ015_SQ003.[i]))
  matching3L[[4]] <- rbind(
    cbind(data5a$matching3N4.SQ001_SQ001.[i],data5a$matching3N4.SQ002_SQ001.[i],data5a$matching3N4.SQ003_SQ001.[i],data5a$matching3N4.SQ004_SQ001.[i],data5a$matching3N4.SQ005_SQ001.[i],data5a$matching3N4.SQ006_SQ001.[i],data5a$matching3N4.SQ007_SQ001.[i],data5a$matching3N4.SQ008_SQ001.[i],data5a$matching3N4.SQ009_SQ001.[i],data5a$matching3N4.SQ010_SQ001.[i], data5a$matching3N4.SQ011_SQ001.[i], data5a$matching3N4.SQ012_SQ001.[i], data5a$matching3N4.SQ013_SQ001.[i], data5a$matching3N4.SQ014_SQ001.[i], data5a$matching3N4.SQ015_SQ001.[i]),
    cbind(data5a$matching3N4.SQ001_SQ002.[i],data5a$matching3N4.SQ002_SQ002.[i],data5a$matching3N4.SQ003_SQ002.[i],data5a$matching3N4.SQ004_SQ002.[i],data5a$matching3N4.SQ005_SQ002.[i],data5a$matching3N4.SQ006_SQ002.[i],data5a$matching3N4.SQ007_SQ002.[i],data5a$matching3N4.SQ008_SQ002.[i],data5a$matching3N4.SQ009_SQ002.[i],data5a$matching3N4.SQ010_SQ002.[i], data5a$matching3N4.SQ011_SQ002.[i], data5a$matching3N4.SQ012_SQ002.[i], data5a$matching3N4.SQ013_SQ002.[i], data5a$matching3N4.SQ014_SQ002.[i], data5a$matching3N4.SQ015_SQ002.[i]),
    cbind(data5a$matching3N4.SQ001_SQ003.[i],data5a$matching3N4.SQ002_SQ003.[i],data5a$matching3N4.SQ003_SQ003.[i],data5a$matching3N4.SQ004_SQ003.[i],data5a$matching3N4.SQ005_SQ003.[i],data5a$matching3N4.SQ006_SQ003.[i],data5a$matching3N4.SQ007_SQ003.[i],data5a$matching3N4.SQ008_SQ003.[i],data5a$matching3N4.SQ009_SQ003.[i],data5a$matching3N4.SQ010_SQ003.[i], data5a$matching3N4.SQ011_SQ003.[i], data5a$matching3N4.SQ012_SQ003.[i], data5a$matching3N4.SQ013_SQ003.[i], data5a$matching3N4.SQ014_SQ003.[i], data5a$matching3N4.SQ015_SQ003.[i]),
    cbind(data5a$matching3N4.SQ001_SQ003.[i],data5a$matching3N4.SQ002_SQ003.[i],data5a$matching3N4.SQ003_SQ003.[i],data5a$matching3N4.SQ004_SQ003.[i],data5a$matching3N4.SQ005_SQ003.[i],data5a$matching3N4.SQ006_SQ003.[i],data5a$matching3N4.SQ007_SQ003.[i],data5a$matching3N4.SQ008_SQ003.[i],data5a$matching3N4.SQ009_SQ003.[i],data5a$matching3N4.SQ010_SQ003.[i], data5a$matching3N4.SQ011_SQ003.[i], data5a$matching3N4.SQ012_SQ003.[i], data5a$matching3N4.SQ013_SQ003.[i], data5a$matching3N4.SQ014_SQ003.[i], data5a$matching3N4.SQ015_SQ003.[i]),
    cbind(data5a$matching3N4.SQ001_SQ004.[i],data5a$matching3N4.SQ002_SQ004.[i],data5a$matching3N4.SQ003_SQ004.[i],data5a$matching3N4.SQ004_SQ004.[i],data5a$matching3N4.SQ005_SQ004.[i],data5a$matching3N4.SQ006_SQ004.[i],data5a$matching3N4.SQ007_SQ004.[i],data5a$matching3N4.SQ008_SQ004.[i],data5a$matching3N4.SQ009_SQ004.[i],data5a$matching3N4.SQ010_SQ004.[i], data5a$matching3N4.SQ011_SQ004.[i], data5a$matching3N4.SQ012_SQ004.[i], data5a$matching3N4.SQ013_SQ004.[i], data5a$matching3N4.SQ014_SQ004.[i], data5a$matching3N4.SQ015_SQ004.[i]))
  matching3L[[5]] <- rbind(
    cbind(data5a$matching3N5.SQ001_SQ001.[i],data5a$matching3N5.SQ002_SQ001.[i],data5a$matching3N5.SQ003_SQ001.[i],data5a$matching3N5.SQ004_SQ001.[i],data5a$matching3N5.SQ005_SQ001.[i],data5a$matching3N5.SQ006_SQ001.[i],data5a$matching3N5.SQ007_SQ001.[i],data5a$matching3N5.SQ008_SQ001.[i],data5a$matching3N5.SQ009_SQ001.[i],data5a$matching3N5.SQ010_SQ001.[i], data5a$matching3N5.SQ011_SQ001.[i], data5a$matching3N5.SQ012_SQ001.[i], data5a$matching3N5.SQ013_SQ001.[i], data5a$matching3N5.SQ014_SQ001.[i], data5a$matching3N5.SQ015_SQ001.[i]),
    cbind(data5a$matching3N5.SQ001_SQ002.[i],data5a$matching3N5.SQ002_SQ002.[i],data5a$matching3N5.SQ003_SQ002.[i],data5a$matching3N5.SQ004_SQ002.[i],data5a$matching3N5.SQ005_SQ002.[i],data5a$matching3N5.SQ006_SQ002.[i],data5a$matching3N5.SQ007_SQ002.[i],data5a$matching3N5.SQ008_SQ002.[i],data5a$matching3N5.SQ009_SQ002.[i],data5a$matching3N5.SQ010_SQ002.[i], data5a$matching3N5.SQ011_SQ002.[i], data5a$matching3N5.SQ012_SQ002.[i], data5a$matching3N5.SQ013_SQ002.[i], data5a$matching3N5.SQ014_SQ002.[i], data5a$matching3N5.SQ015_SQ002.[i]),
    cbind(data5a$matching3N5.SQ001_SQ003.[i],data5a$matching3N5.SQ002_SQ003.[i],data5a$matching3N5.SQ003_SQ003.[i],data5a$matching3N5.SQ004_SQ003.[i],data5a$matching3N5.SQ005_SQ003.[i],data5a$matching3N5.SQ006_SQ003.[i],data5a$matching3N5.SQ007_SQ003.[i],data5a$matching3N5.SQ008_SQ003.[i],data5a$matching3N5.SQ009_SQ003.[i],data5a$matching3N5.SQ010_SQ003.[i], data5a$matching3N5.SQ011_SQ003.[i], data5a$matching3N5.SQ012_SQ003.[i], data5a$matching3N5.SQ013_SQ003.[i], data5a$matching3N5.SQ014_SQ003.[i], data5a$matching3N5.SQ015_SQ003.[i]),
    cbind(data5a$matching3N5.SQ001_SQ003.[i],data5a$matching3N5.SQ002_SQ003.[i],data5a$matching3N5.SQ003_SQ003.[i],data5a$matching3N5.SQ004_SQ003.[i],data5a$matching3N5.SQ005_SQ003.[i],data5a$matching3N5.SQ006_SQ003.[i],data5a$matching3N5.SQ007_SQ003.[i],data5a$matching3N5.SQ008_SQ003.[i],data5a$matching3N5.SQ009_SQ003.[i],data5a$matching3N5.SQ010_SQ003.[i], data5a$matching3N5.SQ011_SQ003.[i], data5a$matching3N5.SQ012_SQ003.[i], data5a$matching3N5.SQ013_SQ003.[i], data5a$matching3N5.SQ014_SQ003.[i], data5a$matching3N5.SQ015_SQ003.[i]),
    cbind(data5a$matching3N5.SQ001_SQ005.[i],data5a$matching3N5.SQ002_SQ005.[i],data5a$matching3N5.SQ003_SQ005.[i],data5a$matching3N5.SQ004_SQ005.[i],data5a$matching3N5.SQ005_SQ005.[i],data5a$matching3N5.SQ006_SQ005.[i],data5a$matching3N5.SQ007_SQ005.[i],data5a$matching3N5.SQ008_SQ005.[i],data5a$matching3N5.SQ009_SQ005.[i],data5a$matching3N5.SQ010_SQ005.[i], data5a$matching3N5.SQ011_SQ005.[i], data5a$matching3N5.SQ012_SQ005.[i], data5a$matching3N5.SQ013_SQ005.[i], data5a$matching3N5.SQ014_SQ005.[i], data5a$matching3N5.SQ015_SQ005.[i]))
  matchingList3[[i]] <- matching3L
}

for (i in 1:length(matchingList3)) {    # for ego i
  mL <- matchingList3[[i]]              # get the matching list 2
  ns <- ns4[[i]]                        # get the size of egonet4
  if(ns>0) {                            # if ns=0, no matching was done!
    mm <- as.matrix(mL[[ns]])           # and the corresponding matrix
    matched <- which(mm==1, arr.ind=T)  # retrieve array indices
    net <- net4[i,]
    if(length(matched)>0) {             # if matching was performed!
      alterL[[i]]$name4[matched[,2]] <- net[matched[,1]]
    }
  }
}

# make long df with w3 alters in ego,
# and indicators for 4 egonets;
df2 <- data.frame(ego=rep(1:length(alterL),each=20),alter=rep(1:20),cdn=NA, study=NA,bff=NA,csn=NA)

for (i in unique(df2$ego)) {  
    for (j in 1:20) { # for alters nested in ego
    # find out if names denoting alter j appear in the 4 egonets
    alter <- unlist(alterL[[i]][j,-1], use.names = F) # get names alter j
    alter <- alter[!is.na(alter)] # exclude NAs
    cdn <- alter %in% net1[i,]
    study <- alter %in% net2[i,]
    bff <- alter %in% net3[i,]
    csn <- alter %in% net4[i,]
    # and if so, give alter j score 1 on indicators; 0 otherwise
    df2$cdn[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% cdn, 1, 0)
    df2$study[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% study, 1, 0)
    df2$bff[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% bff, 1, 0)
    df2$csn[which(df2$ego==i & df2$alter==j)] <- ifelse("TRUE" %in% csn, 1, 0)
  }
}

# now that i have, for each alter of ego (including duplicates) at t2,
# the nets to which they belong..
# i continue with the w1-w2 matching matrices

# we already subsetted the matching matrices (in `w1w2`), but now we take only the rows corresponding to w3 egos who have w2 maintained/created alters
w1w2 <- w1w2[which(w1w2$respnr %in% unique(test$respnr[which(test$w3participation==1)])),]

for ( i in 1:length(alterL)) {

  matchingL <- vector("list", 20) #pre-allocate empty list of length 20, to store matching matrices
  {
    matchingL[[1]] <- w1w2[i, 1:20]
    matchingL[[2]] <- w1w2[i, 21:40]
    matchingL[[3]] <- w1w2[i,41:60]
    matchingL[[4]] <- w1w2[i ,61:80]
    matchingL[[5]] <- w1w2[i ,81:100]
    matchingL[[6]] <- w1w2[i ,101:120]
    matchingL[[7]] <- w1w2[i ,121:140]
    matchingL[[8]] <- w1w2[i ,141:160]
    matchingL[[9]] <- w1w2[i ,161:180]
    matchingL[[10]] <- w1w2[i ,181:200]
    matchingL[[11]] <- w1w2[i ,201:220]
    matchingL[[12]] <- w1w2[i ,221:240]
    matchingL[[13]] <- w1w2[i ,241:260]
    matchingL[[14]] <- w1w2[i ,261:280]
    matchingL[[15]] <- w1w2[i ,281:300]
    matchingL[[16]] <- w1w2[i ,301:320]
    matchingL[[17]] <- w1w2[i ,321:340]
    matchingL[[18]] <- w1w2[i ,341:360]
    matchingL[[19]] <- w1w2[i ,361:380]
    matchingL[[20]] <- w1w2[i ,381:400]
  }
  
  #find the 'right' matching matrix in this list, by taking the one that contains answers
  ind <- NULL
  for (j in seq_along(matchingL)) {
    #check along the sequence of elements j in the matching matrix list if they are non-empty and not NA
    check <- FALSE
    for (col in matchingL[[j]]) {
      if (!all(is.na(col)) && any(nchar(col) > 0)) {
        check <- TRUE
        break  
      }
      }
    if (check) {
      ind <- j
      break  
    }
  }
  if(length(ind)>0) {
    # get the  matrix
    mm <- matchingL[[ind]]
    
    # extract alter ids
    ans <- stringr::str_extract_all(mm,"\\(?[0-9,.]+\\)?") 
    # here, the object refers to the w2-alter j;
    # and the element indicator in the list refers to the w3-alter 
     
    for (j in unique( test$alterid[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i])] )) { # for wave-2 alter j
      
      #to which row/w3-alter was alter j matched?
      match <- which(ans == j)

      # and in which networks did this w3-alter belong?
      nets <- df2[which(df2$ego == i & df2$alter == match[1]),]
      
      if(length(match)>0) { # if j was matched to w3-alters...
        
        # assign to alter j the networks in which the w3-alter appeared
        test$cdn3[ which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <- nets$cdn[1]
        test$study3[ which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <- nets$study[1]
        test$bff3[ which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <- nets$bff[1]
        test$csn3[ which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <- nets$csn[1]
      }
    }
  }
}

#also add dynamic relational info on closeness and communication frequency
#in w3, this was only asked for alters from w2 (those maintained from w1 or w2-created)!
test$frequency.t3 <- NA
test$closeness.t3 <- NA

# subset w3 name interpreters on contact freq. of w1 alters (sq021 - sq040!)
#tail(names(data5a),234)
freq <- data5a[,c(1106:1125)]
close <- data5a[,c(1126:1145)]

# recode into numeric values;
freq <- ifelse(freq=="(Bijna) elke dag",7,
                       ifelse(freq=="1-2 keer per week",6,
                              ifelse(freq=="Aantal keer per maand",5, 
                                     ifelse(freq=="Ong. 1 keer per maand",4, 
                                            ifelse(freq=="Aantal keer per jaar",3,
                                                   ifelse(freq=="Ong. 1 keer per jaar",2,
                                                          ifelse(freq=="Nooit",1,
                                                                      NA   )))))))
close <- ifelse(close=="Heel erg hecht", 4, ifelse(close=="Hecht", 3, ifelse(close=="Enigszins hecht",2, ifelse(close=="Niet hecht", 1, NA))))

for ( i in 1:length(alterL)) { #for ego i
   for (j in unique( test$alterid[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i])] ) ) { # for wave-2 alter j 
     
     test$frequency.t3[ which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <-
       freq[i,][!is.na(freq[i,])][j]
     
     test$closeness.t3[ which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <-
       close[i,][!is.na(close[i,])][j]
   }
}

#last, add respnodent's reasons for not renaming a w2-alter.
#this was only asked for w2 alters who were not relisted
test$reason <- NA

#subset reasons given for not renaming a w2-alter
#tail(names(data5a),243)
reasons <- data5a[,c(1146:1165)]
#unique(reasons$vergeten.SQ001.)
# if, according to ego, the relationship with alter changed ("Onze relatie is veranderd."), we probed further
reasons2 <- data5a[,c(1186:1205)]

for ( i in 1:length(alterL)) { #for ego i


  #if any w2-alter of i went unlisted (thus, scored 0 on `surviveW3`)...
  
  if ( 0 %in% test$surviveW3[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i])] ) {
    
    # for *non-renamed* alter j
     for (j in unique( test$alterid[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] &
                                       test$surviveW3 == 0)])) {
       
       #get respondent i's reason for not re-naming j,
       reasonij <- reasons[i,j]
       #if it was due to a change in relationship, probe further
       reasonij <- ifelse(reasonij == "Onze relatie is veranderd.", reasons2[i,j], reasonij)
       #and if the relationship changed in other ways than we offered as choices, just set to "other"
       reasonij <- ifelse(reasonij == "De relatie is op een andere manier veranderd.", "Andere reden.", reasonij)
       
       test$reason[which(test$ego == unique(test$ego[which(test$w3participation == 1)])[i] & test$alterid == j)] <-
         reasonij
     }
  }
}

#recode...
#unique(test$reason)
test$reason <- ifelse(test$reason == "Andere reden.", "miscellaneous",
                    ifelse(test$reason == "Er is geen gelegenheid geweest voor ons om contact te hebben.",
                           "no occasion to get together",
                           ifelse(test$reason == "Eén van ons is verhuisd.", "someone moved",
                                         ifelse(test$reason == "Eén van ons heeft een belangrijke verandering doorgemaakt  (zoals het stoppen met studeren, het aangaan/beëindigen van een relatie, het krijgen van een kind, enzovoort).", "other life event",
                                                ifelse(test$reason == "Eén van ons heeft gezondheidsproblemen.", 
                                                       "other life event",
                                                       ifelse(test$reason == "We zijn uit elkaar gegroeid/de relatie is verwaterd.", "drifted apart",
                                                              ifelse(test$reason =="We hadden meningsverschillen of ruzie.", "disagreement / conflict",  
                                                                     ifelse(test$reason == "Ik ben simpelweg vergeten om deze persoon opnieuw te noemen.", "forgotten", test$reason))))))))
    
data_alters23 <- test

```


```{r, eval=FALSE}
#now 'bind back together'
test <- df_alters[which(df_alters$statusW2=="Dropped"),]
df_alters <- bind_rows(test, data_alters23)

#rearrange, by ego, and status
df_alters <- df_alters %>% arrange(ego, factor(statusW2, levels=c("Maintained", "Created", "Dropped")))
row.names(df_alters) <- 1:nrow(df_alters)

#new alter ids, 1:n_alter
for (i in unique(df_alters$ego)) {
  df_alters$alterid[which(df_alters$ego == i)] <- 1:length(df_alters$alterid[which(df_alters$ego == i)])
}

#exclude 'names'
df_alters.nn <- df_alters[,-c(5:8)]

#save the resulting dataframe (exclude names)
fsave(df_alters.nn, "tie_maintenance.RDa")
```

<br>


# data_nested.Rda

To account for the  nesting structure of our data (i.e., ties are nested in dyads and networks, that are both nested in ego), we use a 3-level design. So transform the data accordingly. We do this separately for wave 1 --> wave 2, and wave 2 --> wave 3. 

First, tidy up the data-frame (deleting stuff that we won't need + re-ordering columns)

```{r, eval=FALSE}
#in our analyses, we only include non-kin
df <- df_alters.nn[which(df_alters.nn$kin == 0),]
nrow(df) #n=4653

#set gender to binary
df$ego_female <- ifelse(df$ego_gender==2,1,df$ego_gender)
df$alter_female <- ifelse(df$alter_gender==2,1,df$alter_gender)

#drop columns we won't need
names(df)
delete <- c(1,5,15,21:25,27,28,30,33,38,52,72:88)
df <- df[,-delete]

#reorder
names(df)

order <- c("ego", "ego_female", "ego_age", "ego_educ", "study.year", "romantic", "extraversion", "fin_restr",
           "housing.transition_bin", "occupation.transition_bin", "educ_father", "educ_mother",
           "cdn.size", "cdn.size2", "study.size", "study.size2", "bff.size", "bff.size2", "csn.size", "csn.size2",
           "cdn.density", "cdn.density2", "study.density", "study.density2", "bff.density", "bff.density2", "csn.density", "csn.density2",
           "alterid", "alter_female", "alter_age", "alter_educ",
           "same_gender", "dif_age", "sim_educ",
           "proximity", "duration",
           "frequency.t1", "frequency.t2", "frequency.t3", "closeness.t1", "closeness.t2", "closeness.t3", "multiplex", "multiplex.t2",
           "cdn1","cdn2","cdn3","study1", "study2","study3","bff1","bff2","bff3","csn1","csn2","csn3",
           "cdn_embed.t1", "cdn_embed.t2", "study_embed.t1", "study_embed.t2", "bff_embed.t1", "bff_embed.t2", "csn_embed.t1", "csn_embed.t2",
           "statusW2", "w3participation",  "reason")

df <- df[,order]

#multiplex t2
df$multiplex.t2 <- rowSums(df[,c("cdn2","study2","bff2","csn2")]) - 1
```

<br>

## wave 1 --> wave 2

```{r, eval = FALSE}
#in analyzing waves 1 and 2, we look at w1 alters and see whether they are maintained or dropped
#so, drop w2-created alters
df1 <- df[-which(df$statusW2 == "Created"),]
nrow(df1) #n=3174

#make seperate datasets for each relational dimension (at t1)
df.cdn <- df1[which(df1$cdn1==1),]
df.study <- df1[which(df1$study1==1),]
df.bff <- df1[which(df1$bff1==1),]
df.csn <- df1[which(df1$csn1==1),]

# rbind the multiple datasets

#but first, a. make new Y (i.e., remain in *same* egonet); make new variables for features depending on the particular egonet (eg size, embeddedness)
df.cdn$Y <- ifelse(is.na(df.cdn$cdn2),0,df.cdn$cdn2)
df.study$Y <- ifelse(is.na(df.study$study2),0,df.study$study2)
df.bff$Y <- ifelse(is.na(df.bff$bff2),0,df.bff$bff2)
df.csn$Y <- ifelse(is.na(df.csn$csn2),0,df.csn$csn2)

df.cdn$tie <- "Confidant"
df.study$tie <- "Study"
df.bff$tie <- "Friend"
df.csn$tie <- "Sport"

df.cdn$density <- df.cdn$cdn.density
df.bff$density <- df.bff$bff.density
df.study$density <- df.study$study.density
df.csn$density <- df.csn$csn.density

df.cdn$size <- df.cdn$cdn.size
df.bff$size <- df.bff$bff.size
df.study$size <- df.study$study.size
df.csn$size <- df.csn$csn.size

df.cdn$embed <- df.cdn$cdn_embed.t1
df.bff$embed <- df.bff$bff_embed.t1
df.study$embed <- df.study$study_embed.t1
df.csn$embed <- df.csn$csn_embed.t1

#also include "extended" embeddedness, or a the extent to which an alter appearing in the egonet of ego is also embedded in additional layers of ego's network.
#we calculate the sum of embeddedness scores in additional networks, where NA = 0.
df.cdn$embed.ext <- rowSums(cbind(ifelse(is.na(df.cdn$bff_embed.t1),0,df.cdn$bff_embed.t1), ifelse(is.na(df.cdn$study_embed.t1),0,df.cdn$study_embed.t1), ifelse(is.na(df.cdn$csn_embed.t1),0,df.cdn$csn_embed.t1)))
df.bff$embed.ext <- rowSums(cbind(ifelse(is.na(df.bff$cdn_embed.t1),0,df.bff$cdn_embed.t1), ifelse(is.na(df.bff$study_embed.t1),0,df.bff$study_embed.t1), ifelse(is.na(df.bff$csn_embed.t1),0,df.bff$csn_embed.t1)))
df.study$embed.ext <- rowSums(cbind(ifelse(is.na(df.study$cdn_embed.t1),0,df.study$cdn_embed.t1), ifelse(is.na(df.study$bff_embed.t1),0,df.study$bff_embed.t1), ifelse(is.na(df.study$csn_embed.t1),0,df.study$csn_embed.t1)))
df.csn$embed.ext <- rowSums(cbind(ifelse(is.na(df.csn$cdn_embed.t1),0,df.csn$cdn_embed.t1), ifelse(is.na(df.csn$bff_embed.t1),0,df.csn$bff_embed.t1), ifelse(is.na(df.csn$study_embed.t1),0,df.csn$study_embed.t1)))

data <- rbind(df.cdn,df.bff,df.study, df.csn)

#add combination of tie and ego, to make a 'network' variable (i.e., networks are nested in ego, and thus unique to ego)
data$network <- paste0(data$tie, data$ego)

#reorder by ego and alter
data <- data[order(data$ego, data$alterid),]
row.names(data) <- 1:nrow(data)

#set NAs on str. embededness to 0 (for those that were named as the only alter)
data$embed[is.na(data$embed)] <- 0

#keep only columns of interest
names(data)
data <- data[,c(1:12,29:38, 41, 44, 46,49,52,55,66,69:75 )]

#make unique alter id
data$alterid <- paste0(data$ego, "X", data$alterid)

#listwise deletion
colSums(is.na(data)) #missings per variable

#table(data$proximity, useNA="always")
#prop.table(table(data$proximity, useNA="always"))
data$proximity <- ifelse(is.na(data$proximity), "far", data$proximity) #some don't knows on alter proximity, set to "far"

data <- data[which(complete.cases(data)),]

data12 <- data
```

```{r test, eval=FALSE, echo=FALSE}
#check if everything went allright
length(unique(data$ego)) #N_ego = 513
length(unique(paste0(data$ego, "x",data$alterid))) #N_alter = 3104
nrow(data) #N_tie = 4939

#multiplexity, at the alter-level, times n_alter should be n_tie.
#subset from `df` (alter-level) only alters from our tie-level data
testdf <- df
#make an egoXalter id
testdf$alterid <- paste0(testdf$ego, "X", testdf$alterid )
testdf <- testdf[which(testdf$alterid %in% data$alterid),]
(mean(testdf$multiplex)+1) * nrow(testdf) == nrow(data) 
#OKAY!
```
<br>

## wave 2 --> wave 3

```{r, eval = FALSE}
#in analyzing waves 2 and 3, we look at w2-alters (maintained / created)
#so, drop w2-dropped
#and only of egos who participated in w3
df2 <- df[-which(df$statusW2 == "Dropped"),]
df2 <- df2[df2$w3participation == 1,]

#make seperate datasets for each relational dimension (at t2)
df.cdn <- df2[which(df2$cdn2==1),]
df.study <- df2[which(df2$study2==1),]
df.bff <- df2[which(df2$bff2==1),]
df.csn <- df2[which(df2$csn2==1),]

# rbind the multiple datasets

#but first, a. make new Y (i.e., remain in *same* egonet); make new variables for features depending on the particular egonet (eg size, embeddedness)

df.cdn$Y <- ifelse(is.na(df.cdn$cdn3),0,df.cdn$cdn3)
df.study$Y <- ifelse(is.na(df.study$study3),0,df.study$study3)
df.bff$Y <- ifelse(is.na(df.bff$bff3),0,df.bff$bff3)
df.csn$Y <- ifelse(is.na(df.csn$csn3),0,df.csn$csn3)

df.cdn$tie <- "Confidant"
df.study$tie <- "Study"
df.bff$tie <- "Friend"
df.csn$tie <- "Sport"

df.cdn$density <- df.cdn$cdn.density2
df.bff$density <- df.bff$bff.density2
df.study$density <- df.study$study.density2
df.csn$density <- df.csn$csn.density2

df.cdn$size <- df.cdn$cdn.size2
df.bff$size <- df.bff$bff.size2
df.study$size <- df.study$study.size2
df.csn$size <- df.csn$csn.size2

df.cdn$embed <- df.cdn$cdn_embed.t2
df.bff$embed <- df.bff$bff_embed.t2
df.study$embed <- df.study$study_embed.t2
df.csn$embed <- df.csn$csn_embed.t2

#also include "extended" embeddedness, or a the extent to which an alter appearing in the egonet of ego is also embedded in additional layers of ego's network.
#we calculate the sum of embeddedness scores in additional networks, where NA = 0.
df.cdn$embed.ext <- rowSums(cbind(ifelse(is.na(df.cdn$bff_embed.t2),0,df.cdn$bff_embed.t2), ifelse(is.na(df.cdn$study_embed.t2),0,df.cdn$study_embed.t2), ifelse(is.na(df.cdn$csn_embed.t2),0,df.cdn$csn_embed.t2)))
df.bff$embed.ext <- rowSums(cbind(ifelse(is.na(df.bff$cdn_embed.t2),0,df.bff$cdn_embed.t2), ifelse(is.na(df.bff$study_embed.t2),0,df.bff$study_embed.t2), ifelse(is.na(df.bff$csn_embed.t2),0,df.bff$csn_embed.t2)))
df.study$embed.ext <- rowSums(cbind(ifelse(is.na(df.study$cdn_embed.t2),0,df.study$cdn_embed.t2), ifelse(is.na(df.study$bff_embed.t2),0,df.study$bff_embed.t2), ifelse(is.na(df.study$csn_embed.t2),0,df.study$csn_embed.t2)))
df.csn$embed.ext <- rowSums(cbind(ifelse(is.na(df.csn$cdn_embed.t2),0,df.csn$cdn_embed.t2), ifelse(is.na(df.csn$bff_embed.t2),0,df.csn$bff_embed.t2), ifelse(is.na(df.csn$study_embed.t2),0,df.csn$study_embed.t2)))

data <- rbind(df.cdn,df.bff,df.study, df.csn)

#add combination of tie and ego, to make a 'network' variable (i.e., networks are nested in ego, and thus unique to ego)
data$network <- paste0(data$tie, data$ego)

#reorder by ego and alter
data <- data[order(data$ego, data$alterid),]
row.names(data) <- 1:nrow(data)

#set NAs on str. embededness to 0 (for those that were named as the only alter)
data$embed[is.na(data$embed)] <- 0

#keep only columns of interest
names(data)

data <- data[,c(1:12, 29:37, 39, 42, 45, 47,50,53,56,66, 68:75 )]
#here include also `statusW2` (created vs maintained) + `reason` for not renaming alter

data$multiplex <- data$multiplex.t2
names(data)
data <- data[,-24]

#make unique alter id
data$alterid <- paste0(data$ego, "X", data$alterid)

#listwise deletion
colSums(is.na(data)) #missings per variable
data$proximity <- ifelse(is.na(data$proximity), "far", data$proximity)

i=unique(data$ego)[1]

#ego covariates among 'created' alters are still empty, so match based on ego id
#note that transitions were not measured between w2w3, so they are NAs
for (i in unique(data$ego)) {
  
  data$educ_father[data$ego == i & data$statusW2 == "Created"] <- df$educ_father[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$educ_mother[data$ego == i & data$statusW2 == "Created"] <- df$educ_mother[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$study.year[data$ego == i & data$statusW2 == "Created"] <- df$study.year[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$extraversion[data$ego == i & data$statusW2 == "Created"] <- df$extraversion[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$fin_restr[data$ego == i & data$statusW2 == "Created"] <- df$fin_restr[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$romantic[data$ego == i & data$statusW2 == "Created"] <- df$romantic[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$housing.transition_bin[data$ego == i & data$statusW2 == "Created"] <- df$housing.transition_bin[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
  data$occupation.transition_bin[data$ego == i & data$statusW2 == "Created"] <- df$occupation.transition_bin[df$ego == i & (df$statusW2 == "Maintained" | df$statusW2 == "Dropped")][1]
}

#missings on 'reason' are no missings, but refer to alters that were renamed
data$reason[is.na(data$reason)] <- "-"
data <- data[which(complete.cases(data)),]

#transitions are not measured, set to 0
data$housing.transition <- 0
data$occupation.transition <- 0

data23 <- data

length(unique(data$ego)) #N_ego = 281
length(unique(data$alterid)) #N_alter = 1859
```

<br>

Make some last alterations, list the data-frames for waves 1-2 and 2-3.
Also row-bind the data, and add a column indicating when the tie appeared and when it (did not) reappear.

```{r,eval=FALSE}
# make colnames similar
colnames(data12)[colnames(data12) == "frequency.t1"] <- "frequency.t"
colnames(data12)[colnames(data12) == "closeness.t1"] <- "closeness.t"

colnames(data23)[colnames(data23) == "frequency.t2"] <- "frequency.t"
colnames(data23)[colnames(data23) == "closeness.t2"] <- "closeness.t"

colnames(data12)[25:28] <- sub("1$", "", colnames(data12)[25:28])
colnames(data23)[24:27] <- sub("2$", "", colnames(data23)[24:27])

data12$housing.transition <- data12$housing.transition_bin
data12$occupation.transition <- data12$occupation.transition_bin

#2 list data-frames
dataList <- list(data12, data23)
fsave(dataList, "data_nested_List.Rda")

#3 row-bind
data12$period <- "w1 -> w2"
#data23$period <- ifelse(data23$statusW2 == "Created", "w2 -> w3", "w1 - w2 -> w3")
data23$period <- "w2 -> w3"

data <- bind_rows(data12,data23)

length(unique(data$ego))
length(unique(data$alterid))
nrow(data)

names(data)
data <- data[,-c(9,10,29)]

fsave(data, "data_nested.Rda")

```

<br>

# {.unlisted .unnumbered}

----


## References




Data preparation

Last compiled on november, 2024

Getting started

clean up

general custom functions

necessary packages

Download data

load in raw data

make public data-sets to deposit

Import data

tie_maintenance.RDa

wave 1 –> wave 2 (maintained vs dropped)

wave 2 created

wave 2 –> wave 3

data_nested.Rda

wave 1 –> wave 2

wave 2 –> wave 3

References