-
Notifications
You must be signed in to change notification settings - Fork 0
/
FinalVersion.Rmd
294 lines (212 loc) · 11.1 KB
/
FinalVersion.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
---
title: "Does giving information about labor market and migrant's education level affect the perceived skills transferability?"
author: "William Fernandez, Fernanda Gomes, Cintya Huaire"
date: "`r format(Sys.time(), '%B %d, %Y | %H:%M:%S | %Z')`"
output: rmdformats::downcute
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
# Packages
```{r, results='hide', message = FALSE, warning = FALSE}
library(tidyverse) # To use dplyr functions and the pipe operator when needed
library(ggplot2) # To visualize data
library(stargazer) # To format model output
library(modelsummary) # To customize the information displayed in your model summary tables
```
# Loading Dataset
```{r, results='hide', message = FALSE, warning = FALSE}
dta_experiment <- read.csv("final_data.csv", sep = ";")
```
# Processing Dataset
```{r, results='hide', message = FALSE, warning = FALSE}
##Cleaning
# excluding unfinished survey
attach(dta_experiment)
newdata <- dta_experiment[ which(FINISHED==1 ),]
detach(dta_experiment)
# removing 3 observations that were not assign to any group, and surveys without consent
newdata <- newdata[ which(newdata$IL02!=" " | newdata$C103 =="I agree"),]
#dropping observations of non students
newdata <- newdata[ which(newdata$A108!="No" ),]
##Transformation and creation of variables
#changing age into numeric
newdata$A101 <- as.numeric(newdata$A101)
#creating dummies for descriptive
#Can't be named Europe because people who are not from Europe have marked none of the above as an option
newdata$no_docs_req<- ifelse(newdata$A110=="None of the above", 1,0)
regressiondata <- newdata
#Fixing vars names
regressiondata$age <- regressiondata$A101
regressiondata$country_birth <- regressiondata$A104
#creating dummies for regressions
regressiondata$dmale <- ifelse(regressiondata$A102 == "Male", 1,0)
regressiondata$dgerman <- ifelse(regressiondata$A104 == "Germany", 1,0)
regressiondata$dunemployed <- ifelse(regressiondata$A105 == "Unemployed", 1,0)
regressiondata$dwhite <- ifelse(regressiondata$A107 == "Caucasian/White",1,0)
regressiondata$dprivate <- ifelse(regressiondata$A109 == "Private",1,0)
regressiondata$dno_docs_req <- ifelse(regressiondata$A110 == "None of the above",1,0)
regressiondata$dcontrol <- ifelse(regressiondata$IL02 == "Control group", 1,0)
regressiondata$treatment1 <- ifelse(regressiondata$IL02 == "Group 1",1,0)
regressiondata$treatment2 <- ifelse(regressiondata$IL02 == "Group 2",1,0)
regressiondata$dG101_ca <- ifelse(regressiondata$G101 == "Completely agree" , 1, 0)
regressiondata$dG102_ca <- ifelse(regressiondata$G102 == "Completely agree" , 1, 0)
regressiondata$dG103_ca <- ifelse(regressiondata$G103 == "Completely agree" , 1, 0)
regressiondata$dG104_ca <- ifelse(regressiondata$G104 == "Completely agree" , 1, 0)
# transforming the outcome variables and the educ variable in factors
regressiondata$educ_level <- factor(regressiondata$A106, levels = c ("Master's Degree", "Bachelor's Degree", "High school"))
regressiondata$G101 <- factor(regressiondata$G101, levels=c("Completely agree","Somewhat agree", "Neither agree nor disagree", "Somewhat disagree", "Completely disagree"))
regressiondata$G102 <- factor(regressiondata$G102, levels = c("Completely agree","Somewhat agree", "Neither agree nor disagree", "Somewhat disagree", "Completely disagree"))
regressiondata$G103 <- factor(regressiondata$G103, levels = c("Completely agree","Somewhat agree", "Neither agree nor disagree", "Somewhat disagree", "Completely disagree"))
regressiondata$G104 <- factor(regressiondata$G104, levels = c("Completely agree","Somewhat agree", "Neither agree nor disagree", "Somewhat disagree", "Completely disagree"))
# selecting relevant variables to work it
regressiondata <- regressiondata %>% dplyr::select("age", "country_birth", "G101","G102", "G103", "G104", "dmale", "educ_level", "dgerman", "dunemployed", "dwhite", "dprivate", "dno_docs_req", "dcontrol", "treatment1", "treatment2", "dG101_ca", "dG102_ca", "dG103_ca", "dG104_ca")
```
# Descriptive
```{r, results='hide', message = FALSE, warning = FALSE}
ggplot(regressiondata, aes(x = G101, fill=G101)) +
geom_bar(alpha = 0.6) +
theme_classic() +
labs(title = "G101: Immigrants in Germany are well educated",
x = "",
y = "") +
theme(legend.position = "none")
ggsave("../distribution_g101.png")
ggplot(regressiondata, aes(x = G102, fill=G102)) +
geom_bar(alpha = 0.6) +
theme_classic() +
labs(title = "G102: Immigrants can apply their skills in the German labor market (PST)",
x = "",
y = "") +
theme(legend.position = "none")
ggsave("../distribution_g102.png")
ggplot(regressiondata, aes(x = G103, fill=G103)) +
geom_bar(alpha = 0.6) +
theme_classic() +
labs(title = "G103: Immigrants will increase competition in the labor market for me personally (L1)",
x = "",
y = "") +
theme(legend.position = "none",text = element_text(size=10))
ggsave("../distribution_g103.png")
ggplot(regressiondata, aes(x = G104, fill=G104)) +
geom_bar(alpha = 0.6) +
theme_classic() +
labs(title = "G104: In general, immigrants will increase competition in the labor market (L2)",
x = "",
y = "") +
theme(legend.position = "none",text = element_text(size=10))
ggsave("../distribution_g104.png")
#Distribution of age
regressiondata %>%
dplyr::mutate(group_mean = mean(age)) %>%
dplyr::mutate(group_median = median(age)) %>%
ggplot(., aes(x = age)) +
geom_density(alpha = 0.5, fill="#FF6666") +
geom_vline(aes(xintercept = group_mean), linetype = "longdash") +
geom_vline(aes(xintercept = group_median), linetype = "solid") +
theme_classic() +
theme(legend.position = "bottom", plot.caption=element_text(hjust=0)) +
labs(title = "Distribution of age in the sample",
x = "Age*",
y = "",
caption = "*Solid line represents the median and dashed line, the mean.")
ggsave("../age.png")
```
# Balance Check
<em>done in Stata as we had two treatments</em>
Code used:
<pre><code>
*Variables
gen master = (educ_level==1)
gen bachelor = (educ_level==2)
gen high_school = (educ_level==3)
gen treatment_status = 1 if dcontrol==1
replace treatment_status = 2 if treatment1==1
replace treatment_status = 3 if treatment2==1
*Labels
label var age "Age"
label var dmale "Gender: male"
label var dgerman "Country of birth: Germany"
label var dunemployed "Unemployed"
label var dwhite "Ethnicity: white"
label var dprivate "Goes to private university"
label var dno_docs_req "Docs. required: none"
label var master "Highest level attained: Master"
label var bachelor "Highest level attained: Bachelor"
label var high_school "Highest level attained: High school"
global covariates age dmale dgerman dunemployed dwhite dprivate dno_docs_req master bachelor high_school
foreach var in $covariates {
mean `var', over(treatment_status)
matrix a1=e(b)
local mean_control = a1[1,1]
local mean_T1 = a1[1,2]
local mean_T2 = a1[1,3]
reg `var' treatment1 treatment2
local p_diff_1 = (2 * ttail(e(df_r), abs(_b[treatment1]/_se[treatment1])))
local p_diff_2 = (2 * ttail(e(df_r), abs(_b[treatment2]/_se[treatment2])))
lincom treatment1
local diff_C_T1 = r(estimate)
lincom treatment2
local diff_C_T2 = r(estimate)
lincom treatment2-treatment1
local diff_T1_T2 = r(estimate)
local sd_diff_T1_T2 = r(se)
local p_diff_T1_T2 = (2 * ttail(e(df_r), abs(`diff_T1_T2'/`sd_diff_T1_T2')))
matrix Balance_`var' = [ `mean_control', `mean_T1', `mean_T2', `diff_C_T1', `diff_C_T2', `diff_T1_T2']
matrix Stars_`var' = [0,0,0,0,0,0]
matrix Stars_`var'[1,4]=(`p_diff_1'<=0.1)+(`p_diff_1'<=0.05)+(`p_diff_1'<=0.01)
matrix Stars_`var'[1,5]=(`p_diff_2'<=0.1)+(`p_diff_2'<=0.05)+(`p_diff_2'<=0.01)
matrix Stars_`var'[1,6]=(`p_diff_T1_T2'<=0.1)+(`p_diff_T1_T2'<=0.05)+(`p_diff_T1_T2'<=0.01)
matrix Balance = [nullmat(Balance)\ Balance_`var']
matrix Stars = [nullmat(Stars)\Stars_`var']
}
frmttable using "$william\Output\Balance.tex", ///
tex statmat(Balance) fragment ///
ctitles("Variable", "Control", "T1", "T2", "T1 vs. C", "T2 vs. C", "T2 vs. T1") ///
rtitles("Age" \ "Gender: male" \ "Country of birth: Germany" \ "Unemployed" \ "Ethnicity: white" \ ///
"Goes to private university" \ "Docs. required: none" \ "Highest level attained: Master" \ ///
"Highest level attained: Bachelor" \ "Highest level attained: High school") ///
annotate(Stars) asymbol(*, **, ***) note("* p<0.1, ** p<0.05, *** p<0.01") replace
cap mat drop Balance Stars
</code></pre>
# Regressions
```{r, results='hide', message = FALSE, warning = FALSE}
model_1 <- lm(G102 ~ treatment1 + treatment2, data=regressiondata)
model_2 <- lm(dG102_ca ~ treatment1 + treatment2, data=regressiondata)
model_3 <- lm(G103 ~ treatment1 + treatment2, data=regressiondata)
model_4 <- lm(dG103_ca ~ treatment1 + treatment2, data=regressiondata)
model_5 <- lm(G104 ~ treatment1 + treatment2, data=regressiondata)
model_6 <- lm(dG104_ca ~ treatment1 + treatment2, data=regressiondata)
stargazer::stargazer(model_1, model_2,
type = "latex",
report = ("vc*"),
title = "Effect of treatment on Perceived Skills Transferability",
model.numbers = FALSE,
covariate.labels = c("Treatment 1", "Treatment 2"),
dep.var.labels = c("PST-Factor", "PST-Dummy"),
dep.var.caption = "",
omit.stat=c("rsq","adj.rsq","f","ser"))
stargazer::stargazer(model_3, model_4, model_5, model_6,
type = "latex",
report = ("vc*"),
title = "Effect of treatment on secondary outcomes",
model.numbers = FALSE,
covariate.labels = c("Treatment 1", "Treatment 2"),
dep.var.labels = c("L1-Factor", "L1-Dummy","L2-Factor","L2-Dummy"),
dep.var.caption = "",
omit.stat=c("rsq","adj.rsq","f","ser"))
model_7 <- lm(dG103_ca ~ treatment1 + treatment2 + age + dmale + dgerman + dunemployed + dwhite + dprivate + dno_docs_req, data = regressiondata)
model_8 <- lm(dG103_ca ~ treatment1 + treatment2 + treatment2*age + age + dmale + dgerman + dunemployed + dwhite + dprivate + dno_docs_req, data = regressiondata)
model_9 <- lm(dG103_ca ~ treatment1 + treatment2 + treatment2*dunemployed + age + dmale + dgerman + dunemployed + dwhite + dprivate + dno_docs_req, data = regressiondata)
stargazer::stargazer(model_7, model_8, model_9,
type = "latex",
report = ("vc*"),
title = "Effect of treatment on L1: Immigrants will increase competition in the labor market for me personally",
model.numbers = FALSE,
dep.var.labels= "",
covariate.labels = c("Treatment 1", "Treatment 2", "Age", "Male", "German", "Unemployed", "White", "Private univ.", "No docs. req."),
column.labels = c("Additive", "Interactive 1", "Interactive 2"),
dep.var.caption = "",
omit = c("Constant"),
omit.stat=c("rsq","adj.rsq","f","ser"))
```