Independent variables

Last compiled on May, 2025

---

This is the code with which we make our independent variables.

---

1 Initatiating R environment

Start out with a custom function to load a set of required packages. And load the necessary data.

# packages and read data
rm(list = ls())

# (c) Jochem Tolsma
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)
        }
    })
}
packages = c("haven", "coda", "matrixStats", "parallel", "MASS", "doParallel", "dplyr", "cowplot", "tidyverse",
    "naniar", "dotwhisker", "gt", "reshape2", "VGAM", "expss")
fpackage.check(packages)
rm(packages)

load(file = "data/dutch_netsize_analyses_deps.rda")
# 1-(1262/1325)

#--------------------------------------------------------------------------------

---

2 Eyeballing the variables

Some initial eyeballing of the data. What are value labels, for instance, that we can use later?

# eyeballing corona == 1 missing
val_lab(df$V4)
table(df$V4)

# moeder == 6 missing
val_lab(df$V5a)
v <- data.frame(table(df$V5a_anders))
m <- data.frame(table(df$V5b_anders))
# table(df$migr) hier kunnen we w/nw van maken, voor nu is me dat te veel werk

# vader == 4 missing
val_lab(df$V5b)
table(df$V5b)

# elect auto -== 1 missing
val_lab(df$V6)
table(df$V6)

# scooter == no missing
val_lab(df$V7)
table(df$V7)

# vegan == 3 missing
val_lab(df$V8)
table(df$V8)

# geslacht == no missing
val_lab(df$geslacht)
table(df$geslacht)

# opleiding == no missing, 8 'wil niet zeggen'
val_lab(df$opleiding)
table(df$opleiding)

# leeftijd == no missing
table(df$leeftijd)
class(df$leeftijd)

# zipcode == no missing
table(df$postcode)
class(df$postcode)

# new vars io
val_lab(df$werk)  # werk

val_lab(df$woonsituatie)  # wonen

val_lab(df$huishoudomvang)  # huishouden

val_lab(df$inkomen)  # income

val_lab(df$politiek)  # politics

table(df$politiek)  # politics

---

3 Education

Next, we make the education variable, not the value labels below and how I create the variable.

#--------------------------------------------------------------------------------
# opleiding

# geen onderwijs / basisonderwijs / cursus inburgering / cursus Nederlandse taal 1 LBO / VBO / VMBO
# (kader- of beroepsgerichte leerweg) / MBO 1 (assistentenopleidi 2 MAVO / HAVO of VWO (eerste drie
# jaar) / ULO / MULO / VMBO (TL of GL) / VSO 3 MBO 2, 3, 4 (basisberoeps-, vak-, middenkader- of
# specialistenopleiding) of MBO 4 HAVO of VWO (overgegaan naar de 4e klas) / HBS / MMS / HBO
# propedeuse of WO Prop 5 HBO (behalve HBO-master) / WO-kandidaats -of WO-bachelor 6 WO-doctoraal
# of WO-master of HBO-master / postdoctoraal onderwijs 7 weet niet/wil niet zeggen 8 we recreate to
# 3 cats
table(df$opleiding)
df$opl[df$opleiding == 1] <- 1  # prim/sec
df$opl[df$opleiding == 2] <- 1  # prim/sec
df$opl[df$opleiding == 3] <- 1  # prim/sec
df$opl[df$opleiding == 4] <- 2  # lower tert
df$opl[df$opleiding == 5] <- 3  # higher tert/higher prep
df$opl[df$opleiding == 6] <- 3  # higher tert
df$opl[df$opleiding == 7] <- 3  # higher tert
table(df$opl)
df$opl <- as.factor(df$opl)
df <- within(df, opl <- relevel(opl, ref = 3))

---

4 Self-reported gender

Next up, self reported gender, no missings.

#--------------------------------------------------------------------------------
# nothing wrong with geslacht
table(df$geslacht)
df$woman[df$geslacht == 2] <- 1
df$woman[df$geslacht == 1] <- 0
#--------------------------------------------------------------------------------

---

5 Migration background

Migration background is a bit more challenging as there are some open answers of countries that we do want to include in this variable.

# migration background
table(df$V5a_anders)
table(df$V5b_anders)

table(df$V5a)
table(df$V5b)
# df <- df[!df$V5a == 3, ] # drop the missings
# df <- df[!df$V5b == 3, ]

# so if not a 1 (Dutch?) always get a migrant background
df$migr <- ifelse(df$V5a == 2 & df$V5a != 3 | 
                    df$V5b == 2 & df$V5b != 3, 1, # either mother father abroad == 1
                  ifelse(df$V5a == 3 & df$V5b == 3, 3, 0)) # both missing == 3, otherwise 0
table(df$migr)

nrow(df[is.na(df$migr),])

df$vader[df$V5a == 1] <- 1

df$vader[df$V5a_anders=="Amerika"]<-2
df$vader[df$V5a_anders=="Aruba"]<-3
df$vader[df$V5a_anders=="Belgie"]<-2
df$vader[df$V5a_anders=="België"]<-2
df$vader[df$V5a_anders=="Brazilië"]<-3
df$vader[df$V5a_anders=="Cameroun"]<-3
df$vader[df$V5a_anders=="China"]<-3
df$vader[df$V5a_anders=="Curaçao"]<-3
df$vader[df$V5a_anders=="duitsland"]<-2
df$vader[df$V5a_anders=="Duitsland"]<-2
df$vader[df$V5a_anders=="Engeland"]<-2
df$vader[df$V5a_anders=="EU"]<-2
df$vader[df$V5a_anders=="Filipijnen"]<-3
df$vader[df$V5a_anders=="frankrijk"]<-2
df$vader[df$V5a_anders=="Frankrijk"]<-2
df$vader[df$V5a_anders=="indie"]<-2
df$vader[df$V5a_anders=="indonesie"]<-2
df$vader[df$V5a_anders=="Indonesie"]<-2
df$vader[df$V5a_anders=="Indonesië"]<-2
df$vader[df$V5a_anders=="Iran"]<-3
df$vader[df$V5a_anders=="Italie"]<-2
df$vader[df$V5a_anders=="Italië"]<-2
df$vader[df$V5a_anders=="Lettland"]<-2
df$vader[df$V5a_anders=="Marokko"]<-3
df$vader[df$V5a_anders=="Nederlands Indie"]<-2
df$vader[df$V5a_anders=="Nederlands Indië"]<-2
df$vader[df$V5a_anders=="Nieuw-Zeeland"]<-2
df$vader[df$V5a_anders=="Nigeria"]<-3
df$vader[df$V5a_anders=="Oeganda"]<-3
df$vader[df$V5a_anders=="Oostenrijk"]<-3
df$vader[df$V5a_anders=="Polen"]<-3
df$vader[df$V5a_anders=="Polen en Nederland"]<-2
df$vader[df$V5a_anders=="Spanje"]<-2
df$vader[df$V5a_anders=="Sri Lanka"]<-3
df$vader[df$V5a_anders=="suriname"]<-3
df$vader[df$V5a_anders=="Suriname"]<-3
df$vader[df$V5a_anders=="Syrië"]<-3
df$vader[df$V5a_anders=="Turkije"]<-3
df$vader[df$V5a_anders=="Venezuela"]<-3
df$vader[df$V5a_anders=="Verenigd Koninkrijk"]<-2
df$vader[df$V5a_anders=="Yugoslavia"]<-2
df$vader[df$V5a_anders=="Zuid Afrika"]<-3

table(df$vader)
nrow(df[is.na(df$vader), ])

So now we coded the fathers, and we move on with the mothers.

# df$vader <- ifelse(df$V5a == 3 & df$V5a_anders == '', NA, df$vader)

df$moeder[df$V5b == 1] <- 1
table(df$V5b_anders)


df$moeder[df$V5b_anders == "algerije"] <- 3
df$moeder[df$V5b_anders == "Aruba"] <- 3
df$moeder[df$V5b_anders == "Australie"] <- 2
df$moeder[df$V5b_anders == "Belgie"] <- 2
df$moeder[df$V5b_anders == "China"] <- 3

df$moeder[df$V5b_anders == "Curaçao"] <- 3
df$moeder[df$V5b_anders == "duitsland"] <- 2
df$moeder[df$V5b_anders == "Duitsland"] <- 3
df$moeder[df$V5b_anders == "Engelabd"] <- 2
df$moeder[df$V5b_anders == "Engeland"] <- 2
df$moeder[df$V5b_anders == "EU"] <- 2
df$moeder[df$V5b_anders == "Filipijnen"] <- 3
df$moeder[df$V5b_anders == "frankrijk"] <- 2
df$moeder[df$V5b_anders == "Frankrijk"] <- 2
df$moeder[df$V5b_anders == "GB"] <- 2
df$moeder[df$V5b_anders == "indonesie"] <- 2
df$moeder[df$V5b_anders == "Indonesie"] <- 2
df$moeder[df$V5b_anders == "Indonesië"] <- 2
df$moeder[df$V5b_anders == "Irak"] <- 3
df$moeder[df$V5b_anders == "Italie"] <- 2
df$moeder[df$V5b_anders == "Kongo"] <- 3
df$moeder[df$V5b_anders == "Marokko"] <- 3
df$moeder[df$V5b_anders == "Nederlands indie"] <- 2
df$moeder[df$V5b_anders == "Nederlands Indie"] <- 2
df$moeder[df$V5b_anders == "Nederlands Indië"] <- 2
df$moeder[df$V5b_anders == "Nigeria"] <- 3
df$moeder[df$V5b_anders == "Oostenrijk"] <- 2
df$moeder[df$V5b_anders == "Phnom Penh"] <- 3
df$moeder[df$V5b_anders == "Polen"] <- 2
df$moeder[df$V5b_anders == "Rusland"] <- 2
df$moeder[df$V5b_anders == "singapore"] <- 3
df$moeder[df$V5b_anders == "Spanje"] <- 2
df$moeder[df$V5b_anders == "suriname"] <- 3
df$moeder[df$V5b_anders == "Suriname"] <- 3
df$moeder[df$V5b_anders == "Syrië"] <- 3
df$moeder[df$V5b_anders == "Turkije"] <- 3
df$moeder[df$V5b_anders == "Venezuela"] <- 3
df$moeder[df$V5b_anders == "zie 5A"] <- 1
df$moeder[df$V5b_anders == "Zwitserland"] <- 2

# df$moeder <- ifelse(df$V5b == 3 & df$V5b_anders == '', NA, df$moeder)

Now that we created mother and father birth country, we implement a heuristic for migration background: majority, westerns, non-western migration background.

table(df$moeder)
nrow(df[is.na(df$moeder), ])

df$migr3[df$moeder == 1 & df$vader == 1] <- 1
df$migr3[df$moeder == 2 & df$vader == 2] <- 2
df$migr3[df$moeder == 3 & df$vader == 3] <- 3

df$migr3[df$moeder == 3 & df$vader == 2] <- 3
df$migr3[df$moeder == 3 & df$vader == 1] <- 3

df$migr3[df$moeder == 2 & df$vader == 1] <- 2
df$migr3[df$moeder == 2 & df$vader == 3] <- 2

df$migr3[df$moeder == 1 & df$vader == 2] <- 2
df$migr3[df$moeder == 1 & df$vader == 3] <- 3

df$migr3[is.na(df$moeder) & df$vader == 1] <- 1
df$migr3[is.na(df$moeder) & df$vader == 2] <- 2
df$migr3[is.na(df$moeder) & df$vader == 3] <- 3

df$migr3[is.na(df$vader) & df$moeder == 1] <- 1
df$migr3[is.na(df$vader) & df$moeder == 2] <- 2
df$migr3[is.na(df$vader) & df$moeder == 3] <- 3

table(df$migr3)
table(df$migr)

# 1 Maj 2 west 3 nonwest

---

6 Age

We create age and age squared as variables.

#--------------------------------------------------------------------------------
# age
df$leeftijd10 <- df$leeftijd/10
table(df$leeftijd10)
df$leeftijd10sq <- df$leeftijd10^2

df$agecat <- NA
df$agecat[df$leeftijd < 31] <- "18-30"
df$agecat[df$leeftijd > 30 & df$leeftijd < 46] <- "31-45"
df$agecat[df$leeftijd > 45 & df$leeftijd < 66] <- "46-65"
df$agecat[df$leeftijd > 65] <- ">65"

table(df$agecat)
# table(df$leeftijd)
nrow(df[df$leeftijd < 31, ])
nrow(df[df$leeftijd > 30 & df$leeftijd < 46, ])
nrow(df[df$leeftijd > 45 & df$leeftijd < 66, ])
nrow(df[df$leeftijd > 65, ])

---

7 Work situation

We create the current working situation of a respondent. (These are new variable obtained from I&O research, hence the hassle with the identifier before.)

#--------------------------------------------------------------------------------
# NEW IO VARS
#--------------------------------------------------------------------------------

# working val_lab(df$werk)
df$werk <- as.character(df$werk)
df$work <- 1
df$work[df$werk == "anders"] <- 0
df$work[df$werk == "werkloos / werkzoekend / bijstand"] <- 0
df$work[df$werk == "studerend /schoolgaand"] <- 0
df$work[df$werk == "arbeidsongeschikt"] <- 0
df$work[df$werk == "gepensioneerd of VUT"] <- 0
df$work[df$werk == "huisvrouw / huisman"] <- 0

table(df$work)

table(df$werk)

---

8 Living situation

Whether respondents live alone or not.

#--------------------------------------------------------------------------------
# livingalone livingalone
df$woonsituatie <- as.character(df$woonsituatie)

df$livingalone <- 1
df$livingalone[df$woonsituatie == "ik ben gehuwd/woon samen zonder thuiswonende kinderen"] <- 0
df$livingalone[df$woonsituatie == "ik woon bij mijn ouder(s)/verzorger(s)"] <- 0
df$livingalone[df$woonsituatie == "ik ben gehuwd/woon samen met thuiswonende kinderen"] <- 0
df$livingalone[df$woonsituatie == "ik woon alleen (zonder partner) met kinderen"] <- 0
df$livingalone[df$woonsituatie == "anders"] <- 0


df$livingalone[is.na(df$woonsituatie)] <- 2
df$livingalone[df$woonsituatie == "weet niet/wil niet zeggen"] <- 2  # --> should drop this one!

table(df$livingalone)
table(df$woonsituatie)

---

9 Household size

Respondent household size.

#--------------------------------------------------------------------------------
# householdsize

df$hhsize <- as.character(df$huishoudomvang)
table(df$hhsize)
df$hhsize[df$hhsize == "1 (alleen ikzelf)"] <- 1
df$hhsize <- as.numeric(df$hhsize)
table(df$hhsize)
df$hhsize[df$livingalone == 1] <- 1  # this overrides with prior answer
table(df$hhsize)
table(df$huishoudomvang)

---

10 Income

Respondent income.

#--------------------------------------------------------------------------------
# income


table(df$inkomen)
df$inkomen <- as.character(df$inkomen)
nrow(df[is.na(df$inkomen), ])  # --> 37 missings

df$income[df$inkomen == "minimum (Minder dan € 14.100)"] <- 1
df$income[df$inkomen == "beneden modaal (€ 14.100 tot € 29.500)"] <- 1
df$income[df$inkomen == "bijna modaal (€ 29.500 tot € 36.500)"] <- 1
df$income[df$inkomen == "modaal (€ 36.500 tot € 43.500)"] <- 1
df$income[df$inkomen == "tussen 1 en 2 keer modaal (€ 43.500 tot € 73.000)"] <- 2
df$income[df$inkomen == "twee keer modaal (€ 73.000 tot € 87.100)"] <- 2
df$income[df$inkomen == "meer dan 2 keer modaal (€ 87.100 of meer)"] <- 2
df$income[df$inkomen == "weet ik niet / wil ik niet zeggen"] <- 3
df$income[is.na(df$inkomen)] <- 3

table(df$income)
# 1 up to modal 2 modal and up 3 NA/doesn't know/doesn't want to say

---

11 Zipcode info

We create a number of variables that relate to folks’ zipcodes.

#--------------------------------------------------------------------------------
# zip info
zip <- read.csv("data/postcode4.csv")  # --> https://www.cbs.nl/nl-nl/dossier/nederland-regionaal/geografische-data/gegevens-per-postcode
zip <- zip[-2, ]
zipnames <- zip[1, ]
names(zip) <- zipnames
zipnames <- zip[-1, ]

zip <- zip %>%
    replace_with_na_all(condition = ~.x == -99997)
df$postcode <- as.numeric(df$postcode)
zip$PC4 <- as.numeric(zip$PC4)
df <- left_join(df, zip[, c("PC4", "OAD", "UITKMINAOW", "P_KOOPWON", "WOZWONING", "P_NL_ACHTG", "INWONER")],
    by = c(postcode = "PC4"))

df$neighdens <- as.numeric(df$OAD)/1000
df$pbenrec <- (as.numeric(df$UITKMINAOW)/as.numeric(df$INWONER)) * 100
df$pbuyhouse <- as.numeric(df$P_KOOPWON)
df$worthhouse <- as.numeric(df$WOZWONING)/100
df$pdutch <- as.numeric(df$P_NL_ACHTG)

nrow(df[is.na(df$neighdens), ])

summary(df$pbenrec)
summary(df$neighdens)
summary(df$pbuyhouse)
summary(df$worthhouse)
summary(df$pdutch)

---

12 Unused variables

There are a number of remaining variables to the data that we do not use in this particular paper.

#--------------------------------------------------------------------------------
# politics
table(df$politiek)

df$politiek <- as.character(df$politiek)

# right
df$politics <- 3
df$politics[df$politiek == "VVD"] <- 1
df$politics[df$politiek == "PVV"] <- 1
df$politics[df$politiek == "CDA"] <- 1
df$politics[df$politiek == "Forum voor Democratie"] <- 1
df$politics[df$politiek == "ChristenUnie"] <- 1
df$politics[df$politiek == "SGP"] <- 1
df$politics[df$politiek == "BoerBurgerBeweging"] <- 1
df$politics[df$politiek == "JA21"] <- 1
df$politics[df$politiek == "50PLUS"] <- 1


# left
df$politics[df$politiek == "D66"] <- 2
df$politics[df$politiek == "SP"] <- 2
df$politics[df$politiek == "PvdA"] <- 2
df$politics[df$politiek == "GroenLinks"] <- 2
df$politics[df$politiek == "Partij voor de Dieren"] <- 2
df$politics[df$politiek == "Volt"] <- 2
df$politics[df$politiek == "DENK"] <- 2
df$politics[df$politiek == "BIJ1"] <- 2

table(df$politics)
# 1 right 2 left 3 no voting, blanco, no voting rights, doesn't wanna say, etc.

#--------------------------------------------------------------------------------
# unused vars drop missings and create corona var
table(df$V4)
df$coronaiv <- df$V4
table(df$coronaiv)

# create electric vehicle var
table(df$V6)
df$elecaut <- df$V6
table(df$elecaut)

# create scooter var
table(df$V7)
df$scooteriv <- df$V7
table(df$scooteriv)

# create vagan var
table(df$V8)
df$veganiv <- df$V8

# progressive construction: elec vehicle or vegan == progressive
df$progr <- ifelse(df$veganiv == 1 | df$elecaut == 1, 1, 0)
table(df$progr)

---

13 Sample selections

Finally, we apply some sample selection based on the missigness in the data. Not much, as you can see below.

#--------------------------------------------------------------------------------
# missingness and sample selection nrow(df[is.na(df$leeftijd10),]) # no missing
nrow(df[is.na(df$agecat), ])  # no missing
nrow(df[is.na(df$geslacht), ])  # no missing
nrow(df[is.na(df$opl), ])  # no missing
nrow(df[is.na(df$V5a), ])  # six missing
nrow(df[is.na(df$V5b), ])  # four missing
nrow(df[df$migr == 3, ])  # four missing on effective variable
table(df$migr)
nrow(df[is.na(df$migr3), ])  # four missing

# nrow(df[is.na(df$pbenrec), ]) nrow(df[is.na(df$neighdens), ]) # four missing
# nrow(df[is.na(df$pbuyhouse), ])
nrow(df[is.na(df$worthhouse), ])  # eight missing
# nrow(df[is.na(df$pdutch), ])

nrow(df[is.na(df$werk), ])  # no missing
nrow(df[is.na(df$work), ])  # no missing
nrow(df[is.na(df$hhsize), ])  # 1 missing

nrow(df[is.na(df$income), ])  # no missing

nrow(df[df$income == 3, ])  # though 184 won't say/know
nrow(df[is.na(df$inkomen), ])  # oif which 37 system missing
# nrow(df[is.na(df$politics),]) # no missing nrow(df[df$politics == 3,]) # though 223 won't
# say/know nrow(df[is.na(df$politiek),]) # oif which 135 system missing

# df <- df[!is.na(df$pbenrec), ] df <- df[!is.na(df$neighdens), ] df <- df[!is.na(df$pbuyhouse), ]
df <- df[!is.na(df$worthhouse), ]
df <- df[!is.na(df$migr3), ]
df <- df[!is.na(df$hhsize), ]
# df <- df[!is.na(df$pdutch), ]

# df <- df[!df$coronaiv == 3,] df <- df[!df$elecauto == 3,] df <- df[!df$veganiv == 3,] df <-
# df[!df$scooteriv == 3,] summary(df$hhsize)

#--------------------------------------------------------------------------------

---

14 Data storage

Finally, we save the data that we use for the analyses.

# save data
save(df, file = "data/dutch_netsize_analyses.rda")

#--------------------------------------------------------------------------------