Data is taken from the PISA 2003 database which can be found at the ACER website. The data is read into R from the developed SPSS datafile (studentData.sav).
University expectations variable is also created from the SISCED variable in the PISA database.
# read in spss data Export as RData
library(foreign)
setwd("/Users/philipparker/Dropbox/Projects_Research/Social Forces")
studentData <- read.spss("./PISA2003/data/studentData.sav", to.data.frame = TRUE,
use.value.labels = FALSE)
reducedData <- studentData[, c("CNT", "SCHOOLID", "AGE", "GRADE", "ST03Q01",
"HISEI", "HISCED", "ESCS", "IMMIG", "OECD", "SISCED", "W_FSTUWT", grep("^PV.*(MATH|READ|SCIE)$",
names(studentData), value = TRUE))]
rm(studentData)
# Remove non-OECD countries
reducedData <- reducedData[reducedData$OECD == 1, ]
# Remove non-OECD from levels
reducedData$CNT <- factor(reducedData$CNT)
# Add university expectation binary variable
reducedData$UNI <- ifelse(reducedData$SISCED == 5, 1, ifelse(reducedData$SISCED ==
9, NA, 0))
# For pooled standadization z-score HISEI and HISCED
reducedData[, 6:8] <- scale(reducedData[, 6:8])
names(reducedData)
## [1] "CNT" "SCHOOLID" "AGE" "GRADE" "ST03Q01" "HISEI"
## [7] "HISCED" "ESCS" "IMMIG" "OECD" "SISCED" "W_FSTUWT"
## [13] "PV1MATH" "PV2MATH" "PV3MATH" "PV4MATH" "PV5MATH" "PV1READ"
## [19] "PV2READ" "PV3READ" "PV4READ" "PV5READ" "PV1SCIE" "PV2SCIE"
## [25] "PV3SCIE" "PV4SCIE" "PV5SCIE" "UNI"
save(reducedData, file = "./PISA2003/data/reducedData19013013.RData")
rm(list = ls(all = TRUE))
To manage memory the smaller analysis dataset is saved, the workplace is cleared, and then the workspace is cleared. The smaller data set is then sourced for the remaining analysis.
Principals components of PISA math, verbal, and science tests was undertaken using the psych package.
# Load RData
load("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/data/reducedData19013013.RData")
# Z-score achievement
reducedData[, grep("^PV[0-9]", names(reducedData))] <- scale(reducedData[, grep("^PV[0-9]",
names(reducedData))])
# Principal Value analaysis
library(psych)
# Achievement scores
Ach <- list()
for (i in 1:5) {
Ach[[i]] <- principal(reducedData[, grep(paste0("^PV", i, ".*$"), names(reducedData))],
1, score = TRUE)$scores
}
Ach <- data.frame(do.call(cbind, Ach))
names(Ach) <- paste0("Ach", 1:5)
# add principal components to dataframe
AchData <- data.frame(reducedData, Ach)
AchAgg <- list()
for (i in 1:5) {
AchAgg[[i]] <- aggregate(AchData[, 28 + i] ~ SCHOOLID, FUN = mean, data = AchData)
}
AchAgg <- data.frame(do.call(cbind, AchAgg))[, c(1, 2, 4, 6, 8, 10)]
names(AchAgg) <- c("SCHOOLID", paste0("AchAgg", 1:5))
AchData <- merge(AchData, AchAgg, by = "SCHOOLID")
# Save the data with the ICCs and principal components
save(AchData, file = "/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/data/AchData19013013.RData")
rm(list = ls(all = TRUE))
Country ICCs created using the psych package from a set of anova models. The data is split by country and the ICCs are created in a loop.
# ICCs
load("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/data/AchData19013013.RData")
library(multilevel)
ICCdata <- split(AchData, AchData$CNT)
ICCs <- lapply(ICCdata, function(x) ICC1(aov(Ach1 ~ as.factor(SCHOOLID), data = x,
weights = W_FSTUWT)))
ICCs <- data.frame(do.call(rbind, ICCs))
ICCs$Countries <- row.names(ICCs)
names(ICCs) <- c("ICC", "CNT")
row.names(ICCs) <- NULL
ICCs <- data.frame(ICCs)
ICCs2 <- data.frame(ICCs[order(ICCs), ])
ICCs2 <- ICCs2[1:30, ]
row.names(ICCs2) <- ICCs2$CNT
ICCplot <- as.table(t(ICCs2[1:30, ]))
barplot(ICCplot, horiz = TRUE, cex.names = 0.5, las = 1, main = "ICC by Country")
abline(v = mean(ICCs$ICC, na.rm = TRUE))
# Will need to report this in a table Aggregate the ICCs back into the data
# set
AchData <- merge(AchData, ICCs, by = "CNT")
Descriptive results consist of:
Plots ordered by ICC size.
# Human development index
HDI <- read.csv("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/data/Table_1__Human_Development_Index_and_its_components.csv",
na.strings = "..")
descriptives <- data.frame(n.obs = tapply(AchData$CNT, AchData$CNT, length),
ICC = round(tapply(AchData$ICC, AchData$CNT, mean), 3), ExpectTables = round(tapply(AchData$UNI,
AchData$CNT, mean, na.rm = TRUE), 3))
descriptives$CNT <- row.names(descriptives)
descriptives$strat <- c(-1.043, 1.817, 1.018, -1.321, -0.138, 1.621, 1.862,
-0.087, -1.02, -0.87, -0.474, -1.043, -0.474, 1.421, -0.302, -0.063, 0.166,
-0.474, 0.072, 0.7, NA, 0.937, -1.043, -0.419, -0.083, -0.327, 1.621, -0.138,
1.201, -1.321)
descriptives$HDI <- HDI[HDI$Abbreviation %in% descriptives$CNT, 5]
# ifelse(descriptives$selection <16,1,0)
descriptives2 <- data.frame(descriptives[order(descriptives$ExpectTables), ])
DESplot <- as.table(t(descriptives2[, c("ExpectTables", "CNT")]))
barplot(DESplot, horiz = TRUE, cex.names = 0.5, las = 1, main = "University Expectations by Country")
abline(v = mean(descriptives$ExpectTables, na.rm = TRUE))
cor(descriptives[, c(2:3, 5:6)], use = "complete.obs")
## ICC ExpectTables strat HDI
## ICC 1.00000 0.08432 0.6906 -0.3323
## ExpectTables 0.08432 1.00000 -0.2138 -0.3142
## strat 0.69055 -0.21378 1.0000 -0.3737
## HDI -0.33233 -0.31420 -0.3737 1.0000
pairs(descriptives[, c(2:3, 5:6)])
First the R data file is saved for use in Mplus using the MplusAutomation package. Five data sets are created each with one plausable value for the Achievement principal component. Note that here we display the unscaled logistic regression coefficents. The scaled coefficents are avaliable from the author with results showing even more similarity with probit results. Scaled logit results were done in the Mediaton package by Imia and colleagues. Here we create a Mplus template file that is read in modified and run using a set of R functions. These functions are designed to loop through countries, SES measures, and logit or probit link functions. This means 30 by 3 by 2 Mplus models are created and run.
library(MplusAutomation)
for (i in 1:5){
prepareMplusData(AchData[,c(1:28,39,28+i, 33+i)],
filename=paste0("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/data/PISA2003_PV", i, ".dat")
)
}
All models are run using a single Mplus template file. The template file is:
TITLE: Country XXX !Outcome = Expectation of university entry
!Treatment mediator Interaction;
DATA: FILE = "./PISA2003/data/PISA2003_PV1.dat";
VARIABLE: NAMES = CNT SCHOOLID AGE GRADE ST03Q01 HISEI HISCED ESCS IMMIG OECD
SISCED W_FSTUWT PV1MATH PV2MATH PV3MATH PV4MATH PV5MATH PV1READ PV2READ PV3READ
PV4READ PV5READ PV1SCIE PV2SCIE PV3SCIE PV4SCIE PV5SCIE UNI ICC PVAch;
MISSING=.;
USEVARIABLES UNI HISEI PVAch int;
categorical = UNI;
useobs CNT EQ 1;
cluster=schoolid;
weight=W_FSTUWT;
ANALYSIS: type=complex; estimator = MLR; LINK = PROBIT;
DEFINE:
int = HISEI*PVAch;
STANDARDIZE HISEI PVAch int;
MODEL:
[UNI$1] (mbeta0);
UNI on HISEI (beta2);
UNI on PVAch (beta1);
UNI on int (beta3);
[PVAch](gamma0);
PVAch on HISEI (gamma1);
PVAch(sig2);
MODEL CONSTRAINT:
new(ind dir arg11 arg10 arg01 arg00 v1 v0
probit11 probit10 probit01 probit00
tie de pie ortie orde orpie Ctotal total prop);
dir=beta3*gamma0+beta2;
ind=beta1*gamma1+beta3*gamma1;
arg11=-mbeta0+beta2+(beta1+beta3)*(gamma0+gamma1);
arg10=-mbeta0+beta2+(beta1+beta3)*gamma0;
arg01=-mbeta0+beta1*(gamma0+gamma1);
arg00=-mbeta0+beta1*gamma0;
v1=(beta1+beta3)^2*sig2+1;
v0=beta1^2*sig2+1;
probit11=arg11/sqrt(v1);
probit10=arg10/sqrt(v1);
probit01=arg01/sqrt(v0);
probit00=arg00/sqrt(v0);
! Phi function needed below:
tie=phi(probit11)-phi(probit10);
de=phi(probit10)-phi(probit00);
pie=phi(probit01)-phi(probit00);
ortie=(phi(probit11)/(1-phi(probit11)))/
(phi(probit10)/(1-phi(probit10)));
orde=(phi(probit10)/(1-phi(probit10)))/
(phi(probit00)/(1-phi(probit00)));
orpie=(phi(probit01)/(1-phi(probit01)))/
(phi(probit00)/(1-phi(probit00)));
Ctotal = de + tie;
total= ind + dir;
prop = (tie)/ (Ctotal+.00001);
The template file is iterative over in order to create country specific syntax files. The creation and running of files is done by:
fileCreation <- function(ses.var = "HISEI", link = "PROBIT") {
temp <- readLines("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/MPLUS/TEMPLATEFILE19012013.inp")
temp[10:60] <- unlist(sapply(temp[10:60], function(x) gsub("HISEI", ses.var,
x)))
countries <- levels(AchData$CNT)
if (link == "LOGIT") {
link = "LOGIT"
temp[17] <- gsub("PROBIT", "LOGIT", temp[17])
} else {
link = "PROBIT"
}
print(link)
print(ses.var)
for (i in 1:30) {
x <- temp
x[1] <- gsub("XXX", countries[i], x[1])
x[13] <- paste0("useobs CNT EQ ", i, ";")
cat(x, sep = "\n", file = paste0("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/MPLUSforMARKDOWN/",
ses.var, "/", link, "/", countries[i], ".inp"))
}
}
fileCreation(ses.var = "HISEI", link = "LOGIT") # Highest occupation prestige of parents
## [1] "LOGIT"
## [1] "HISEI"
fileCreation(ses.var = "HISEI", link = "PROBIT") # Highest occupation prestige of parents
## [1] "PROBIT"
## [1] "HISEI"
fileCreation(ses.var = "HISCED", link = "LOGIT") # Highest education level of parents
## [1] "LOGIT"
## [1] "HISCED"
fileCreation(ses.var = "HISCED", link = "PROBIT") # Highest education level of parents
## [1] "PROBIT"
## [1] "HISCED"
fileCreation(ses.var = "ESCS", link = "LOGIT") # SES composite
## [1] "LOGIT"
## [1] "ESCS"
fileCreation(ses.var = "ESCS", link = "PROBIT") # SES composite
## [1] "PROBIT"
## [1] "ESCS"
All the models in each of the folders using MplusAutomation package in R:
runModels("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/MPLUSforMARKDOWN/HISEI", recursive=TRUE)
runModels("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/MPLUSforMARKDOWN/HISCED", recursive=TRUE)
runModels("/Users/philipparker/Dropbox/Projects_Research/Social Forces/PISA2003/MPLUSforMARKDOWN/ESCS", recursive=TRUE)
Extraction of Results was done using a series of R functions. The Mplus output files differed slightly from those from other countries and thus there entries were added manually.
# Run on mplus files one at a time
setwd("/Users/philipparker/Dropbox/Projects_Research/Social Forces")
paths <- c("./PISA2003/MPLUSforMARKDOWN/HISEI/PROBIT", "./PISA2003/MPLUSforMARKDOWN/HISEI/LOGIT",
"./PISA2003/MPLUSforMARKDOWN/HISCED/PROBIT", "./PISA2003/MPLUSforMARKDOWN/HISCED/LOGIT",
"./PISA2003/MPLUSforMARKDOWN/ESCS/PROBIT", "./PISA2003/MPLUSforMARKDOWN/ESCS/LOGIT")
parameterExtract <- function(path) {
files <- list.files(path, full.names = TRUE, pattern = ".out$")
# Extract relavent output
temp <- grep("New/Additional Parameters", readLines(files[1]))
lines <- c(temp + c(13, 14, 19, 21), temp - 20)
output <- sapply(files, function(x) readLines(x)[lines])
# Function for extracting relavent parameters
f1 <- function(x) unlist(strsplit(x, "[[:space:]]+"))[c(2:4)]
results <- apply(output, c(1, 2), f1)
TIE <- matrix(results[, 1, 1:30], 30, 3, byrow = TRUE)
DE <- matrix(results[, 2, 1:30], 30, 3, byrow = TRUE)
total <- matrix(results[, 3, 1:30], 30, 3, byrow = TRUE)
prop <- matrix(results[, 4, 1:30], 30, 3, byrow = TRUE)
sch <- matrix(results[, 5, 1:30], 30, 3, byrow = TRUE)
results <- cbind(total, TIE, DE, prop, sch)
results <- data.frame(results[, c(2:3, 5:6, 8:9, 11:12, 14:15)], descriptives$ICC,
descriptives$HDI, descriptives$strat)
names(results) <- c("total", "totalse", "ind", "indse", "dir", "dirse",
"prop", "propse", "sch", "schse", "ICC", "HDI", "strat")
results <- data.frame(apply(results, 2, as.numeric))
results$CNT <- levels(AchData$CNT)
results
}
resultsList <- lapply(paths, parameterExtract)
names(resultsList) <- c("HISEI_PROBIT", "HISEI_LOGIT", "HISCED_PROBIT", "HISCED_LOGIT",
"ESCS_PROBIT", "ESCS_LOGIT")
# Parsing was incomplete for poland so results added by hand. Better parsing
# would solve this problem.
resultsList[[3]][25, ] <- c(0.18, 0.01, 0.072, 0.005, 0.108, 0.009, 0.402, 0.028,
-0.01, 0.028, 0.163, 15, 1, "POL")
resultsList[[3]][, 1:13] <- lapply(resultsList[[3]][, 1:13], as.numeric)
resultsList[[4]][25, ] <- c(0.243, 0.014, 0.097, 0.007, 0.146, 0.012, 0.4, 0.026,
-0.02, 0.048, 0.163, 15, 1, "POL")
resultsList[[4]][, 1:13] <- lapply(resultsList[[4]][, 1:13], as.numeric)
resultsList
## $HISEI_PROBIT
## total totalse ind indse dir dirse prop propse sch schse ICC
## 1 0.109 0.006 0.058 0.003 0.051 0.006 0.530 0.038 0.043 0.023 0.233
## 2 0.151 0.013 0.082 0.007 0.069 0.009 0.542 0.032 0.084 0.041 0.589
## 3 0.176 0.008 0.103 0.006 0.074 0.007 0.582 0.030 0.031 0.031 0.564
## 4 0.105 0.006 0.040 0.003 0.065 0.006 0.381 0.032 -0.016 0.024 0.189
## 5 0.119 0.015 0.065 0.008 0.054 0.010 0.544 0.044 0.025 0.050 0.381
## 6 0.196 0.009 0.085 0.006 0.112 0.009 0.431 0.029 0.025 0.033 0.509
## 7 0.148 0.010 0.087 0.006 0.061 0.008 0.590 0.038 0.074 0.036 0.633
## 8 0.107 0.009 0.049 0.004 0.058 0.009 0.457 0.045 -0.053 0.028 0.156
## 9 0.169 0.008 0.077 0.005 0.092 0.007 0.457 0.029 -0.027 0.029 0.245
## 10 0.096 0.008 0.033 0.004 0.063 0.009 0.344 0.052 0.035 0.020 0.050
## 11 0.139 0.011 0.091 0.007 0.048 0.010 0.654 0.051 0.030 0.035 0.535
## 12 0.130 0.009 0.082 0.006 0.048 0.007 0.632 0.037 -0.001 0.043 0.265
## 13 0.157 0.013 0.064 0.009 0.093 0.009 0.409 0.037 0.143 0.067 0.390
## 14 0.211 0.011 0.094 0.008 0.116 0.010 0.447 0.032 0.037 0.050 0.570
## 15 0.118 0.009 0.058 0.005 0.061 0.010 0.487 0.053 -0.016 0.026 0.191
## 16 0.103 0.013 0.026 0.004 0.078 0.011 0.250 0.034 0.019 0.020 0.042
## 17 0.172 0.008 0.044 0.007 0.127 0.008 0.258 0.038 0.021 0.046 0.574
## 18 0.132 0.012 0.045 0.007 0.088 0.011 0.337 0.047 0.070 0.049 0.538
## 19 0.070 0.007 0.028 0.005 0.041 0.007 0.407 0.064 -0.033 0.059 0.422
## 20 0.171 0.013 0.084 0.007 0.087 0.013 0.490 0.048 0.194 0.083 0.325
## 21 0.148 0.014 0.059 0.006 0.089 0.011 0.399 0.031 0.094 0.027 0.511
## 22 0.142 0.012 0.097 0.008 0.044 0.012 0.688 0.062 -0.018 0.053 0.625
## 23 0.132 0.010 0.043 0.004 0.088 0.010 0.329 0.033 0.002 0.031 0.082
## 24 0.113 0.009 0.059 0.005 0.054 0.009 0.522 0.052 -0.023 0.031 0.189
## 25 0.176 0.010 0.073 0.006 0.104 0.009 0.412 0.032 -0.004 0.027 0.163
## 26 0.163 0.011 0.084 0.008 0.079 0.009 0.517 0.038 0.024 0.028 0.382
## 27 0.190 0.010 0.098 0.007 0.092 0.008 0.517 0.031 0.123 0.039 0.466
## 28 0.121 0.008 0.041 0.004 0.080 0.009 0.335 0.036 0.024 0.024 0.110
## 29 0.085 0.011 0.050 0.008 0.035 0.009 0.587 0.073 0.046 0.054 0.624
## 30 0.093 0.006 0.042 0.003 0.052 0.006 0.447 0.040 0.000 0.026 0.265
## HDI strat CNT
## 1 0.938 -1.043 AUS
## 2 0.895 1.817 AUT
## 3 0.897 1.018 BEL
## 4 0.911 -1.321 CAN
## 5 0.913 -0.138 CHE
## 6 0.873 1.621 CZE
## 7 0.920 1.862 DEU
## 8 0.901 -0.087 DNK
## 9 0.885 -1.020 ESP
## 10 0.892 -0.870 FIN
## 11 0.893 -0.474 FRA
## 12 0.875 -1.043 GBR
## 13 0.860 -0.474 GRC
## 14 0.831 1.421 HUN
## 15 0.916 -0.302 IRL
## 16 0.906 -0.063 ISL
## 17 0.881 0.166 ITA
## 18 0.912 -0.474 JPN
## 19 0.909 0.072 KOR
## 20 0.875 0.700 LUX
## 21 0.775 NA MEX
## 22 0.921 0.937 NLD
## 23 0.955 -1.043 NOR
## 24 0.919 -0.419 NZL
## 25 0.821 -0.083 POL
## 26 0.816 -0.327 PRT
## 27 0.840 1.621 SVK
## 28 0.916 -0.138 SWE
## 29 0.722 1.201 TUR
## 30 0.937 -1.321 USA
##
## $HISEI_LOGIT
## total totalse ind indse dir dirse prop propse sch schse ICC
## 1 0.128 0.008 0.069 0.004 0.059 0.009 0.537 0.046 0.070 0.039 0.233
## 2 0.185 0.018 0.102 0.009 0.084 0.011 0.548 0.029 0.149 0.070 0.589
## 3 0.236 0.011 0.136 0.007 0.101 0.009 0.575 0.028 0.053 0.052 0.564
## 4 0.129 0.008 0.049 0.003 0.080 0.009 0.380 0.038 -0.026 0.039 0.189
## 5 0.141 0.021 0.074 0.011 0.067 0.012 0.527 0.032 0.051 0.092 0.381
## 6 0.258 0.013 0.113 0.008 0.145 0.012 0.438 0.029 0.044 0.057 0.509
## 7 0.173 0.013 0.102 0.008 0.071 0.008 0.588 0.028 0.136 0.062 0.633
## 8 0.135 0.013 0.064 0.006 0.071 0.011 0.472 0.042 -0.087 0.047 0.156
## 9 0.213 0.009 0.096 0.006 0.117 0.009 0.451 0.029 -0.042 0.049 0.245
## 10 0.142 0.012 0.048 0.005 0.094 0.015 0.340 0.055 0.056 0.033 0.050
## 11 0.184 0.015 0.119 0.008 0.065 0.013 0.648 0.050 0.052 0.059 0.535
## 12 0.171 0.013 0.107 0.008 0.064 0.009 0.628 0.033 0.003 0.073 0.265
## 13 0.169 0.018 0.066 0.011 0.103 0.011 0.391 0.039 0.246 0.115 0.390
## 14 0.256 0.015 0.111 0.010 0.146 0.013 0.432 0.034 0.069 0.086 0.570
## 15 0.167 0.012 0.080 0.007 0.087 0.014 0.478 0.055 -0.027 0.044 0.191
## 16 0.140 0.019 0.037 0.006 0.103 0.016 0.262 0.035 0.036 0.034 0.042
## 17 0.241 0.012 0.058 0.010 0.183 0.013 0.241 0.039 0.033 0.077 0.574
## 18 0.178 0.016 0.058 0.008 0.119 0.016 0.327 0.047 0.117 0.081 0.538
## 19 0.062 0.009 0.026 0.006 0.036 0.008 0.423 0.083 -0.049 0.106 0.422
## 20 0.239 0.019 0.115 0.009 0.124 0.019 0.481 0.047 0.330 0.142 0.325
## 21 0.211 0.016 0.082 0.007 0.129 0.014 0.390 0.032 0.156 0.044 0.511
## 22 0.184 0.016 0.124 0.009 0.059 0.015 0.678 0.061 -0.018 0.088 0.625
## 23 0.170 0.015 0.061 0.006 0.109 0.014 0.357 0.034 0.003 0.054 0.082
## 24 0.163 0.014 0.085 0.007 0.079 0.013 0.519 0.050 -0.039 0.052 0.189
## 25 0.232 0.015 0.100 0.007 0.133 0.013 0.429 0.031 -0.006 0.047 0.163
## 26 0.215 0.013 0.109 0.009 0.106 0.012 0.505 0.039 0.041 0.047 0.382
## 27 0.244 0.013 0.126 0.008 0.119 0.012 0.515 0.031 0.208 0.067 0.466
## 28 0.177 0.013 0.060 0.005 0.117 0.013 0.339 0.034 0.039 0.040 0.110
## 29 0.084 0.012 0.047 0.008 0.037 0.009 0.559 0.079 0.071 0.094 0.624
## 30 0.114 0.008 0.050 0.004 0.064 0.008 0.436 0.045 -0.001 0.044 0.265
## HDI strat CNT
## 1 0.938 -1.043 AUS
## 2 0.895 1.817 AUT
## 3 0.897 1.018 BEL
## 4 0.911 -1.321 CAN
## 5 0.913 -0.138 CHE
## 6 0.873 1.621 CZE
## 7 0.920 1.862 DEU
## 8 0.901 -0.087 DNK
## 9 0.885 -1.020 ESP
## 10 0.892 -0.870 FIN
## 11 0.893 -0.474 FRA
## 12 0.875 -1.043 GBR
## 13 0.860 -0.474 GRC
## 14 0.831 1.421 HUN
## 15 0.916 -0.302 IRL
## 16 0.906 -0.063 ISL
## 17 0.881 0.166 ITA
## 18 0.912 -0.474 JPN
## 19 0.909 0.072 KOR
## 20 0.875 0.700 LUX
## 21 0.775 NA MEX
## 22 0.921 0.937 NLD
## 23 0.955 -1.043 NOR
## 24 0.919 -0.419 NZL
## 25 0.821 -0.083 POL
## 26 0.816 -0.327 PRT
## 27 0.840 1.621 SVK
## 28 0.916 -0.138 SWE
## 29 0.722 1.201 TUR
## 30 0.937 -1.321 USA
##
## $HISCED_PROBIT
## total totalse ind indse dir dirse prop propse sch schse ICC
## 1 0.122 0.007 0.040 0.003 0.082 0.006 0.330 0.025 0.023 0.028 0.233
## 2 0.153 0.012 0.053 0.007 0.100 0.009 0.346 0.032 0.080 0.038 0.589
## 3 0.176 0.010 0.080 0.006 0.096 0.009 0.454 0.032 0.041 0.031 0.564
## 4 0.117 0.006 0.028 0.002 0.089 0.007 0.240 0.021 -0.021 0.022 0.189
## 5 0.156 0.016 0.052 0.008 0.105 0.012 0.331 0.035 0.023 0.050 0.381
## 6 0.202 0.010 0.089 0.007 0.113 0.009 0.440 0.032 0.017 0.033 0.509
## 7 0.155 0.012 0.085 0.007 0.071 0.010 0.544 0.041 0.086 0.036 0.633
## 8 0.131 0.015 0.051 0.004 0.080 0.013 0.387 0.037 -0.052 0.029 0.156
## 9 0.159 0.009 0.065 0.006 0.094 0.007 0.411 0.027 -0.031 0.032 0.245
## 10 0.097 0.007 0.027 0.003 0.070 0.008 0.280 0.038 0.042 0.020 0.050
## 11 0.138 0.011 0.077 0.007 0.061 0.009 0.560 0.046 0.023 0.036 0.535
## 12 0.151 0.011 0.065 0.005 0.086 0.009 0.433 0.034 0.008 0.039 0.265
## 13 0.139 0.012 0.055 0.006 0.084 0.008 0.393 0.030 0.154 0.066 0.390
## 14 0.223 0.011 0.097 0.008 0.126 0.010 0.434 0.030 0.030 0.047 0.570
## 15 0.119 0.009 0.047 0.004 0.073 0.009 0.392 0.043 -0.015 0.025 0.191
## 16 0.172 0.013 0.033 0.004 0.139 0.012 0.193 0.023 0.015 0.021 0.042
## 17 0.154 0.009 0.046 0.007 0.107 0.008 0.302 0.039 0.021 0.047 0.574
## 18 0.152 0.014 0.053 0.009 0.099 0.011 0.347 0.051 0.073 0.047 0.538
## 19 0.092 0.007 0.031 0.005 0.061 0.006 0.336 0.045 -0.029 0.058 0.422
## 20 0.155 0.017 0.067 0.008 0.088 0.013 0.431 0.042 0.201 0.089 0.325
## 21 0.155 0.014 0.053 0.006 0.102 0.010 0.343 0.025 0.090 0.026 0.511
## 22 0.105 0.012 0.066 0.007 0.039 0.010 0.632 0.069 -0.003 0.054 0.625
## 23 0.136 0.013 0.028 0.003 0.107 0.012 0.209 0.024 -0.003 0.030 0.082
## 24 0.143 0.011 0.053 0.005 0.090 0.010 0.368 0.035 -0.028 0.032 0.189
## 25 0.180 0.010 0.072 0.005 0.108 0.009 0.402 0.028 -0.010 0.028 0.163
## 26 0.134 0.010 0.064 0.007 0.070 0.007 0.478 0.039 0.031 0.029 0.382
## 27 0.195 0.010 0.099 0.008 0.096 0.009 0.507 0.034 0.118 0.039 0.466
## 28 0.131 0.011 0.032 0.004 0.099 0.010 0.243 0.025 0.032 0.025 0.110
## 29 0.105 0.011 0.061 0.008 0.044 0.007 0.580 0.051 0.033 0.054 0.624
## 30 0.131 0.007 0.033 0.003 0.098 0.007 0.253 0.025 0.003 0.026 0.265
## HDI strat CNT
## 1 0.938 -1.043 AUS
## 2 0.895 1.817 AUT
## 3 0.897 1.018 BEL
## 4 0.911 -1.321 CAN
## 5 0.913 -0.138 CHE
## 6 0.873 1.621 CZE
## 7 0.920 1.862 DEU
## 8 0.901 -0.087 DNK
## 9 0.885 -1.020 ESP
## 10 0.892 -0.870 FIN
## 11 0.893 -0.474 FRA
## 12 0.875 -1.043 GBR
## 13 0.860 -0.474 GRC
## 14 0.831 1.421 HUN
## 15 0.916 -0.302 IRL
## 16 0.906 -0.063 ISL
## 17 0.881 0.166 ITA
## 18 0.912 -0.474 JPN
## 19 0.909 0.072 KOR
## 20 0.875 0.700 LUX
## 21 0.775 NA MEX
## 22 0.921 0.937 NLD
## 23 0.955 -1.043 NOR
## 24 0.919 -0.419 NZL
## 25 15.000 1.000 POL
## 26 0.816 -0.327 PRT
## 27 0.840 1.621 SVK
## 28 0.916 -0.138 SWE
## 29 0.722 1.201 TUR
## 30 0.937 -1.321 USA
##
## $HISCED_LOGIT
## total totalse ind indse dir dirse prop propse sch schse ICC
## 1 0.148 0.009 0.046 0.004 0.101 0.009 0.314 0.028 0.037 0.047 0.233
## 2 0.183 0.017 0.067 0.009 0.117 0.012 0.364 0.031 0.143 0.066 0.589
## 3 0.240 0.014 0.105 0.007 0.135 0.013 0.436 0.030 0.065 0.053 0.564
## 4 0.145 0.009 0.033 0.002 0.112 0.010 0.225 0.025 -0.036 0.037 0.189
## 5 0.169 0.025 0.061 0.010 0.107 0.017 0.364 0.028 0.053 0.091 0.381
## 6 0.267 0.013 0.118 0.009 0.150 0.012 0.440 0.030 0.026 0.057 0.509
## 7 0.184 0.016 0.099 0.009 0.085 0.011 0.538 0.034 0.156 0.062 0.633
## 8 0.182 0.023 0.068 0.006 0.115 0.019 0.371 0.035 -0.085 0.048 0.156
## 9 0.203 0.011 0.082 0.006 0.120 0.009 0.406 0.028 -0.046 0.053 0.245
## 10 0.140 0.011 0.039 0.004 0.101 0.013 0.277 0.040 0.067 0.033 0.050
## 11 0.183 0.015 0.099 0.009 0.084 0.013 0.541 0.044 0.040 0.060 0.535
## 12 0.205 0.015 0.087 0.007 0.118 0.014 0.423 0.033 0.013 0.067 0.265
## 13 0.155 0.016 0.060 0.008 0.095 0.010 0.386 0.032 0.265 0.115 0.390
## 14 0.271 0.016 0.111 0.010 0.160 0.013 0.411 0.031 0.056 0.082 0.570
## 15 0.165 0.012 0.063 0.006 0.102 0.014 0.382 0.046 -0.025 0.042 0.191
## 16 0.248 0.020 0.047 0.006 0.200 0.018 0.191 0.021 0.030 0.035 0.042
## 17 0.212 0.012 0.062 0.009 0.150 0.012 0.292 0.040 0.035 0.079 0.574
## 18 0.213 0.018 0.067 0.011 0.147 0.017 0.313 0.050 0.123 0.078 0.538
## 19 0.075 0.010 0.025 0.006 0.050 0.007 0.338 0.058 -0.042 0.104 0.422
## 20 0.217 0.023 0.092 0.010 0.126 0.020 0.421 0.044 0.343 0.150 0.325
## 21 0.220 0.016 0.074 0.007 0.146 0.013 0.337 0.025 0.149 0.043 0.511
## 22 0.134 0.016 0.083 0.009 0.051 0.014 0.621 0.072 0.002 0.090 0.625
## 23 0.187 0.020 0.040 0.005 0.147 0.019 0.213 0.024 -0.004 0.052 0.082
## 24 0.210 0.017 0.073 0.006 0.136 0.016 0.350 0.035 -0.047 0.054 0.189
## 25 0.243 0.014 0.097 0.007 0.146 0.012 0.400 0.026 -0.020 0.048 0.163
## 26 0.171 0.011 0.081 0.008 0.091 0.009 0.472 0.039 0.053 0.049 0.382
## 27 0.252 0.013 0.126 0.009 0.126 0.012 0.500 0.034 0.200 0.066 0.466
## 28 0.204 0.017 0.046 0.005 0.158 0.016 0.226 0.023 0.050 0.041 0.110
## 29 0.095 0.012 0.058 0.008 0.037 0.008 0.607 0.059 0.050 0.093 0.624
## 30 0.149 0.009 0.036 0.004 0.114 0.009 0.238 0.028 0.000 0.043 0.265
## HDI strat CNT
## 1 0.938 -1.043 AUS
## 2 0.895 1.817 AUT
## 3 0.897 1.018 BEL
## 4 0.911 -1.321 CAN
## 5 0.913 -0.138 CHE
## 6 0.873 1.621 CZE
## 7 0.920 1.862 DEU
## 8 0.901 -0.087 DNK
## 9 0.885 -1.020 ESP
## 10 0.892 -0.870 FIN
## 11 0.893 -0.474 FRA
## 12 0.875 -1.043 GBR
## 13 0.860 -0.474 GRC
## 14 0.831 1.421 HUN
## 15 0.916 -0.302 IRL
## 16 0.906 -0.063 ISL
## 17 0.881 0.166 ITA
## 18 0.912 -0.474 JPN
## 19 0.909 0.072 KOR
## 20 0.875 0.700 LUX
## 21 0.775 NA MEX
## 22 0.921 0.937 NLD
## 23 0.955 -1.043 NOR
## 24 0.919 -0.419 NZL
## 25 15.000 1.000 POL
## 26 0.816 -0.327 PRT
## 27 0.840 1.621 SVK
## 28 0.916 -0.138 SWE
## 29 0.722 1.201 TUR
## 30 0.937 -1.321 USA
##
## $ESCS_PROBIT
## total totalse ind indse dir dirse prop propse sch schse ICC
## 1 0.154 0.008 0.061 0.004 0.093 0.008 0.394 0.027 0.012 0.028 0.233
## 2 0.221 0.014 0.097 0.007 0.124 0.011 0.438 0.027 0.073 0.038 0.589
## 3 0.250 0.009 0.129 0.007 0.122 0.009 0.514 0.027 0.031 0.030 0.564
## 4 0.158 0.005 0.040 0.002 0.118 0.006 0.254 0.019 -0.030 0.022 0.189
## 5 0.192 0.019 0.092 0.010 0.100 0.013 0.479 0.033 0.013 0.046 0.381
## 6 0.252 0.010 0.101 0.007 0.151 0.009 0.401 0.026 0.021 0.033 0.509
## 7 0.195 0.012 0.109 0.007 0.086 0.010 0.560 0.036 0.067 0.035 0.633
## 8 0.159 0.013 0.068 0.005 0.092 0.013 0.424 0.040 -0.055 0.029 0.156
## 9 0.222 0.008 0.096 0.006 0.126 0.008 0.433 0.023 -0.036 0.031 0.245
## 10 0.134 0.007 0.042 0.004 0.093 0.009 0.310 0.039 0.033 0.020 0.050
## 11 0.196 0.012 0.116 0.008 0.080 0.011 0.593 0.042 0.024 0.035 0.535
## 12 0.193 0.009 0.100 0.006 0.093 0.008 0.519 0.030 0.002 0.042 0.265
## 13 0.190 0.015 0.065 0.008 0.125 0.010 0.342 0.029 0.143 0.065 0.390
## 14 0.274 0.012 0.102 0.009 0.172 0.012 0.373 0.031 0.024 0.046 0.570
## 15 0.178 0.009 0.061 0.007 0.117 0.011 0.342 0.040 -0.024 0.025 0.191
## 16 0.170 0.013 0.042 0.004 0.128 0.012 0.249 0.025 0.018 0.020 0.042
## 17 0.213 0.010 0.053 0.008 0.160 0.009 0.250 0.033 0.016 0.046 0.574
## 18 0.215 0.012 0.070 0.008 0.144 0.012 0.326 0.038 0.053 0.045 0.538
## 19 0.107 0.008 0.036 0.006 0.071 0.006 0.334 0.045 -0.035 0.058 0.422
## 20 0.204 0.015 0.092 0.007 0.112 0.014 0.453 0.040 0.190 0.078 0.325
## 21 0.198 0.017 0.068 0.009 0.130 0.011 0.344 0.026 0.080 0.025 0.511
## 22 0.177 0.013 0.121 0.008 0.056 0.012 0.685 0.052 -0.014 0.052 0.625
## 23 0.178 0.011 0.050 0.005 0.128 0.010 0.279 0.026 0.000 0.029 0.082
## 24 0.165 0.010 0.077 0.005 0.087 0.010 0.471 0.036 -0.028 0.029 0.189
## 25 0.228 0.010 0.076 0.006 0.152 0.010 0.332 0.027 -0.008 0.029 0.163
## 26 0.196 0.011 0.089 0.009 0.107 0.008 0.453 0.034 0.025 0.028 0.382
## 27 0.268 0.010 0.122 0.008 0.146 0.010 0.456 0.028 0.118 0.038 0.466
## 28 0.172 0.010 0.054 0.005 0.118 0.010 0.315 0.030 0.023 0.023 0.110
## 29 0.119 0.012 0.061 0.009 0.058 0.008 0.514 0.048 0.035 0.054 0.624
## 30 0.159 0.006 0.042 0.004 0.118 0.007 0.262 0.024 -0.004 0.025 0.265
## HDI strat CNT
## 1 0.938 -1.043 AUS
## 2 0.895 1.817 AUT
## 3 0.897 1.018 BEL
## 4 0.911 -1.321 CAN
## 5 0.913 -0.138 CHE
## 6 0.873 1.621 CZE
## 7 0.920 1.862 DEU
## 8 0.901 -0.087 DNK
## 9 0.885 -1.020 ESP
## 10 0.892 -0.870 FIN
## 11 0.893 -0.474 FRA
## 12 0.875 -1.043 GBR
## 13 0.860 -0.474 GRC
## 14 0.831 1.421 HUN
## 15 0.916 -0.302 IRL
## 16 0.906 -0.063 ISL
## 17 0.881 0.166 ITA
## 18 0.912 -0.474 JPN
## 19 0.909 0.072 KOR
## 20 0.875 0.700 LUX
## 21 0.775 NA MEX
## 22 0.921 0.937 NLD
## 23 0.955 -1.043 NOR
## 24 0.919 -0.419 NZL
## 25 0.821 -0.083 POL
## 26 0.816 -0.327 PRT
## 27 0.840 1.621 SVK
## 28 0.916 -0.138 SWE
## 29 0.722 1.201 TUR
## 30 0.937 -1.321 USA
##
## $ESCS_LOGIT
## total totalse ind indse dir dirse prop propse sch schse ICC
## 1 0.179 0.010 0.067 0.004 0.112 0.011 0.373 0.031 0.018 0.048 0.233
## 2 0.281 0.021 0.128 0.009 0.153 0.016 0.457 0.025 0.133 0.066 0.589
## 3 0.349 0.013 0.174 0.008 0.174 0.014 0.500 0.026 0.050 0.051 0.564
## 4 0.184 0.008 0.043 0.003 0.141 0.009 0.233 0.022 -0.052 0.036 0.189
## 5 0.228 0.030 0.112 0.014 0.116 0.018 0.490 0.026 0.028 0.085 0.381
## 6 0.340 0.013 0.138 0.008 0.202 0.014 0.405 0.025 0.035 0.058 0.509
## 7 0.234 0.017 0.132 0.009 0.103 0.011 0.561 0.027 0.122 0.062 0.633
## 8 0.214 0.020 0.092 0.007 0.122 0.017 0.430 0.036 -0.089 0.047 0.156
## 9 0.282 0.010 0.118 0.006 0.163 0.011 0.420 0.024 -0.055 0.052 0.245
## 10 0.196 0.011 0.059 0.006 0.136 0.015 0.303 0.041 0.054 0.033 0.050
## 11 0.266 0.017 0.154 0.010 0.112 0.016 0.578 0.041 0.041 0.059 0.535
## 12 0.267 0.014 0.138 0.008 0.129 0.012 0.518 0.028 0.005 0.073 0.265
## 13 0.195 0.022 0.061 0.011 0.134 0.014 0.313 0.032 0.244 0.113 0.390
## 14 0.332 0.019 0.109 0.012 0.223 0.018 0.327 0.033 0.045 0.080 0.570
## 15 0.251 0.012 0.080 0.009 0.171 0.017 0.318 0.042 -0.038 0.042 0.191
## 16 0.239 0.019 0.062 0.005 0.177 0.018 0.260 0.025 0.034 0.034 0.042
## 17 0.294 0.015 0.067 0.011 0.227 0.014 0.226 0.034 0.025 0.077 0.574
## 18 0.292 0.016 0.087 0.010 0.205 0.019 0.297 0.039 0.091 0.076 0.538
## 19 0.078 0.010 0.026 0.006 0.052 0.007 0.335 0.062 -0.049 0.103 0.422
## 20 0.292 0.022 0.128 0.009 0.163 0.022 0.440 0.042 0.325 0.135 0.325
## 21 0.285 0.017 0.094 0.009 0.191 0.013 0.330 0.026 0.133 0.041 0.511
## 22 0.231 0.018 0.154 0.009 0.077 0.018 0.667 0.055 -0.012 0.088 0.625
## 23 0.238 0.017 0.074 0.006 0.164 0.016 0.310 0.028 0.000 0.051 0.082
## 24 0.247 0.015 0.112 0.007 0.135 0.015 0.453 0.035 -0.047 0.048 0.189
## 25 0.304 0.015 0.109 0.008 0.196 0.015 0.357 0.027 -0.015 0.050 0.163
## 26 0.260 0.012 0.112 0.010 0.148 0.012 0.430 0.035 0.042 0.047 0.382
## 27 0.355 0.013 0.158 0.009 0.198 0.015 0.443 0.028 0.203 0.067 0.466
## 28 0.265 0.015 0.081 0.007 0.184 0.016 0.307 0.029 0.037 0.038 0.110
## 29 0.101 0.013 0.053 0.009 0.049 0.008 0.521 0.061 0.052 0.093 0.624
## 30 0.181 0.008 0.040 0.004 0.140 0.009 0.223 0.025 -0.011 0.042 0.265
## HDI strat CNT
## 1 0.938 -1.043 AUS
## 2 0.895 1.817 AUT
## 3 0.897 1.018 BEL
## 4 0.911 -1.321 CAN
## 5 0.913 -0.138 CHE
## 6 0.873 1.621 CZE
## 7 0.920 1.862 DEU
## 8 0.901 -0.087 DNK
## 9 0.885 -1.020 ESP
## 10 0.892 -0.870 FIN
## 11 0.893 -0.474 FRA
## 12 0.875 -1.043 GBR
## 13 0.860 -0.474 GRC
## 14 0.831 1.421 HUN
## 15 0.916 -0.302 IRL
## 16 0.906 -0.063 ISL
## 17 0.881 0.166 ITA
## 18 0.912 -0.474 JPN
## 19 0.909 0.072 KOR
## 20 0.875 0.700 LUX
## 21 0.775 NA MEX
## 22 0.921 0.937 NLD
## 23 0.955 -1.043 NOR
## 24 0.919 -0.419 NZL
## 25 0.821 -0.083 POL
## 26 0.816 -0.327 PRT
## 27 0.840 1.621 SVK
## 28 0.916 -0.138 SWE
## 29 0.722 1.201 TUR
## 30 0.937 -1.321 USA
Here was conduct country level analysis using correations and partial correlations. Partial correlations where done using the ppcor package. Scatter plots of results and barplots or results orded by size are created using the probit models for occupational prestige.
library(ppcor)
# Preperation of Results Corr between total effect and ICC by country
# Correlation
sapply(resultsList, function(x) cor(x[, "total"], x[, "ICC"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.40956 0.28596 0.23206 0.01751 0.38198
## ESCS_LOGIT
## 0.19200
sapply(resultsList, function(x) cor(x[!c(x$CNT == "MEX"), "total"], x[!c(x$CNT ==
"MEX"), "strat"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.4920 0.4059 0.4911 0.3436 0.4670
## ESCS_LOGIT
## 0.3563
# Partial correlation controlling for age of selection
sapply(resultsList, function(x) pcor(data.frame(x[!c(x$CNT == "MEX"), "total"],
x[!c(x$CNT == "MEX"), "ICC"], x[!c(x$CNT == "MEX"), c("strat", "HDI")]))$estimate[2,
1])
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.09403 -0.01700 -0.07966 -0.26190 0.09346
## ESCS_LOGIT
## -0.08752
# Partial correlation controlling for ICC
sapply(resultsList, function(x) pcor(data.frame(x[!c(x$CNT == "MEX"), "total"],
x[!c(x$CNT == "MEX"), "ICC"], x[!c(x$CNT == "MEX"), c("strat", "HDI")]))$estimate[3,
1])
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.2942 0.2974 0.4039 0.4021 0.3007
## ESCS_LOGIT
## 0.3394
# Corr between proportion indirect and CNT Correlation Probit
sapply(resultsList, function(x) cor(x[, "prop"], x[, "ICC"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.4551 0.4101 0.6598 0.6634 0.5191
## ESCS_LOGIT
## 0.4718
sapply(resultsList, function(x) cor(x[!c(x$CNT == "MEX"), "prop"], x[!c(x$CNT ==
"MEX"), "strat"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.3025 0.2970 0.5426 0.5684 0.4301
## ESCS_LOGIT
## 0.4462
# Logit log-odds
sapply(resultsList, function(x) cor(x[, "ind"], x[, "ICC"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.5531 0.4368 0.6063 0.4891 0.5756
## ESCS_LOGIT
## 0.4166
sapply(resultsList, function(x) cor(x[!c(x$CNT == "MEX"), "ind"], x[!c(x$CNT ==
"MEX"), "strat"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.5460 0.4871 0.6876 0.6381 0.5711
## ESCS_LOGIT
## 0.5021
# Probit Partial correlation controlling for age of selection
sapply(resultsList, function(x) pcor(data.frame(x[!c(x$CNT == "MEX"), "prop"],
x[!c(x$CNT == "MEX"), "ICC"], x[!c(x$CNT == "MEX"), c("strat", "HDI")]))$estimate[2,
1])
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.3905 0.3375 0.5252 0.5146 0.3762
## ESCS_LOGIT
## 0.2901
# Partial correlation controlling for ICC
sapply(resultsList, function(x) pcor(data.frame(x[!c(x$CNT == "MEX"), "prop"],
x[!c(x$CNT == "MEX"), "ICC"], x[!c(x$CNT == "MEX"), c("strat", "HDI")]))$estimate[3,
1])
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## -4.688e-02 -7.922e-05 1.189e-01 1.620e-01 9.254e-02
## ESCS_LOGIT
## 1.644e-01
# Log-odds Partial correlation controlling for age of selection
sapply(resultsList, function(x) pcor(data.frame(x[!c(x$CNT == "MEX"), "ind"],
x[!c(x$CNT == "MEX"), "ICC"], x[!c(x$CNT == "MEX"), c("strat", "HDI")]))$estimate[2,
1])
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.3083 0.1693 0.3576 0.1935 0.3419
## ESCS_LOGIT
## 0.1420
# Partial correlation controlling for ICC
sapply(resultsList, function(x) pcor(data.frame(x[!c(x$CNT == "MEX"), "ind"],
x[!c(x$CNT == "MEX"), "ICC"], x[!c(x$CNT == "MEX"), c("strat", "HDI")]))$estimate[3,
1])
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.2470 0.2796 0.4071 0.4103 0.2881
## ESCS_LOGIT
## 0.3429
# Plots for weights with country specific standadization
par(mfrow = c(2, 2))
results <- resultsList[[1]]
plot(results[, "ICC"], results[, "total"], ylab = "Total Effect", xlab = "ICC",
ylim = c(0, 0.3), pch = "", main = "Total Effect by ICCs")
text(results[, "ICC"], results[, "total"], results[, "CNT"], cex = 0.6)
abline(lm(results[, "total"] ~ results[, "ICC"]), col = "grey")
arrows(x0 = results[, "ICC"], y0 = results[, "total"] - results[, "totalse"],
x1 = results[, "ICC"], y1 = results[, "total"] + results[, "totalse"], length = 0,
col = "grey")
# Indirect Effect
plot(results[, "ICC"], results[, "prop"], ylab = "Proportion of Total Effect ",
xlab = "ICC", ylim = c(0, 0.8), pch = "", main = "Primary Effect by ICCs")
text(results[, "ICC"], results[, "prop"], results[, "CNT"], cex = 0.6)
abline(lm(results[, "prop"] ~ results[, "ICC"]), col = "grey")
arrows(x0 = results[, "ICC"], y0 = results[, "prop"] - results[, "propse"],
x1 = results[, "ICC"], y1 = results[, "prop"] + results[, "propse"], length = 0,
col = "grey")
ReTotal <- data.frame(results[order(results$total), ])
ResPlot <- as.table(t(ReTotal[, c("total", "CNT")]))
colnames(ResPlot) <- ResPlot[2, ]
barplot(ResPlot, horiz = TRUE, cex.names = 0.8, las = 1, main = "Total Effects by Country")
abline(v = mean(results$total, na.rm = TRUE))
ReTotal <- data.frame(results[order(results$prop), ])
ResPlot <- as.table(t(ReTotal[, c("prop", "CNT")]))
colnames(ResPlot) <- ResPlot[2, ]
barplot(ResPlot, horiz = TRUE, cex.names = 0.8, las = 1, main = "Primary Effect by Country")
abline(v = mean(results$prop, na.rm = TRUE))
par(mfrow = c(1, 1))
dev.off()
## null device
## 1
Illustrating the consistency of results we replot figures uing the probit results using the ESCS measure in the pisa database.
results <- resultsList[[5]]
par(mfrow = c(2, 2))
plot(results[, "ICC"], results[, "total"], ylab = "Total Effect", xlab = "ICC",
ylim = c(0, 0.3), pch = "", main = "Relationship between SES \n and Educational Expectations")
text(results[, "ICC"], results[, "total"], results[, "CNT"], cex = 0.6)
abline(lm(results[, 1] ~ results[, "ICC"]), col = "grey")
arrows(x0 = results[, "ICC"], y0 = results[, "total"] - results[, "totalse"],
x1 = results[, "ICC"], y1 = results[, "total"] + results[, "totalse"], length = 0,
col = "grey")
# Indirect Effect
plot(results[, "ICC"], results[, "prop"], ylab = "Proportion of Total Effect ",
xlab = "ICC", ylim = c(0, 0.8), pch = "", main = "Relationship between SES and Educational Expectations \n Explained by Achievement Differentials")
text(results[, "ICC"], results[, "prop"], results[, "CNT"], cex = 0.6)
abline(lm(results[, "prop"] ~ results[, "ICC"]), col = "grey")
arrows(x0 = results[, "ICC"], y0 = results[, "prop"] - results[, "propse"],
x1 = results[, "ICC"], y1 = results[, "prop"] + results[, "propse"], length = 0,
col = "grey")
ReTotal <- data.frame(results[order(results$total), ])
ResPlot <- as.table(t(ReTotal[, c("total", "CNT")]))
colnames(ResPlot) <- ResPlot[2, ]
barplot(ResPlot, horiz = TRUE, cex.names = 0.5, las = 1)
abline(v = mean(results$total, na.rm = TRUE))
ReTotal <- data.frame(results[order(results$prop), ])
ResPlot <- as.table(t(ReTotal[, c("prop", "CNT")]))
colnames(ResPlot) <- ResPlot[2, ]
barplot(ResPlot, horiz = TRUE, cex.names = 0.5, las = 1)
abline(v = mean(results$prop, na.rm = TRUE))
par(mfrow = c(1, 1))
dev.off()
## null device
## 1
In the models it was clear that there were some countries who were outliers. While we report on the sample of countries as a whole here, we also provide sensitivity analysis without outliers:
# Deletion of Italy as outlier. Corr between proportion indirect and CNT
sapply(resultsList, function(x) cor(x[!c(x$CNT == "ITA"), "prop"], x[!c(x$CNT ==
"ITA"), "ICC"]))
## HISEI_PROBIT HISEI_LOGIT HISCED_PROBIT HISCED_LOGIT ESCS_PROBIT
## 0.5776 0.5442 0.7134 0.7184 0.6091
## ESCS_LOGIT
## 0.5665