-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathExtension-SCM-In-Place-PE.Rmd
253 lines (206 loc) · 10.2 KB
/
Extension-SCM-In-Place-PE.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
---
output: pdf_document
latex_engine: xelatex
header-includes:
- "\\usepackage{listings}"
- "\\lstset{breaklines=true}"
---
# Extension: In-Place Placebo for Pernambuco
`install.packages("dplyr")`\
`install.packages("ggplot2")`\
`install.packages("Synth")`\
`install.packages("xtable")`\
`install.packages("ggpubr")`
```{r setup, include=FALSE}
#-----------------------------------------------------------------------------------------#
# Initial Setup
#-----------------------------------------------------------------------------------------#
# Load Required Libraries
library(dplyr)
library(ggplot2)
library(Synth)
library(xtable)
library(ggpubr)
```
```{r}
# Load Data
load("data/DATA_COMPLETE.RData")
load("data/abbr_code.RData")
# Subsetting data by grade and subject (for all states, including PE):
PRIMARY_M <- as.data.frame(filter(DATA_COMPLETE, grade == "P", subject == "math"))
PRIMARY_P <- as.data.frame(filter(DATA_COMPLETE, grade == "P", subject == "port"))
LOWERS_M <- as.data.frame(filter(DATA_COMPLETE, grade == "LS", subject == "math"))
LOWERS_P <- as.data.frame(filter(DATA_COMPLETE, grade == "LS", subject == "port"))
UPPERS_M <- as.data.frame(filter(DATA_COMPLETE, grade == "US", subject == "math"))
UPPERS_P <- as.data.frame(filter(DATA_COMPLETE, grade == "US", subject == "port"))
#-----------------------------------------------------------------------------------------#
# Function: prepare_p_ls_PE()
# Description: prepares data for the synth function - primary and lower secondary school
#-----------------------------------------------------------------------------------------#
prepare_p_ls_PE <- function(data){
library(Synth)
predictors <- c("homicides", "TWh", "ln_pop", "unemployment", "edu_invest_pc")
DATA_PM <- dataprep(foo = data,
predictors = predictors,
dependent = "score",
unit.variable = "code_state",
time.variable = "year",
unit.names.variable = "abbr_state",
treatment.identifier = 26, # Pernambuco Code
controls.identifier = c(11:17, 21, 22, 24:25, 27:29, 31:33, 35, 41:43, 50:53),
# Exclude PE (26) and CE (23) from controls
time.predictors.prior = seq(1995, 2007, 2),
time.optimize.ssr = seq(1995, 2007, 2),
time.plot = seq(1995, 2019, 2))
return(DATA_PM)
};
#-----------------------------------------------------------------------------------------#
# Function: prepare_us_PE()
# Description: prepares data for the synth function - upper secondary school
#-----------------------------------------------------------------------------------------#
prepare_us_PE <- function(data){
library(Synth)
predictors <- c("homicides", "TWh", "unemployment", "ln_pop", "edu_invest_pc")
DATA_PM <- dataprep(foo = data,
predictors = predictors,
dependent = "score",
unit.variable = "code_state",
time.variable = "year",
unit.names.variable = "abbr_state",
treatment.identifier = 26, # Pernambuco Code
controls.identifier = c(11:17, 21, 22, 24:25, 27:29, 31:33, 35, 41:43, 50:53),
# Exclude PE (26) and CE (23) from controls
time.predictors.prior = seq(1995, 2009, 2),
time.optimize.ssr = seq(1995, 2009, 2),
time.plot = seq(1995, 2019, 2))
return(DATA_PM)
};
#-----------------------------------------------------------------------------------------#
# Preparing data for Synth for PE
#-----------------------------------------------------------------------------------------#
# Prepare data for SCM (specifying PE as the treatment unit):
DATA_PE_PM <- prepare_p_ls_PE(PRIMARY_M)
DATA_PE_PP <- prepare_p_ls_PE(PRIMARY_P)
DATA_PE_LSM <- prepare_p_ls_PE(LOWERS_M)
DATA_PE_LSP <- prepare_p_ls_PE(LOWERS_P)
# DATA_PE_USM <- prepare_us_PE(UPPERS_M)
# DATA_PE_USP <- prepare_us_PE(UPPERS_P)
# Primary School Mathematics and Portuguese
DATA_PE_PM <- prepare_p_ls_PE(PRIMARY_M)
DATA_PE_PP <- prepare_p_ls_PE(PRIMARY_P)
# Lower Secondary School Mathematics and Portuguese
DATA_PE_LSM <- prepare_p_ls_PE(LOWERS_M)
DATA_PE_LSP <- prepare_p_ls_PE(LOWERS_P)
# Upper Secondary School Mathematics and Portuguese
# DATA_PE_USM <- prepare_us_PE(UPPERS_M)
# DATA_PE_USP <- prepare_us_PE(UPPERS_P)
#-----------------------------------------------------------------------------------------#
# Function: plot_scm_PE()
# Description: prepares data from the synthetic control output to be plotted with ggplot
#-----------------------------------------------------------------------------------------#
plot_scm_PE <- function(original_data, synth.tables){
library(tidyverse)
W <- as.data.frame(synth.tables[["tab.w"]])
str(W)
W <- W %>%
filter(w.weights > 0.01) %>%
mutate(w.weights = round(w.weights, digits = 3)) %>%
rename(abbr_state = unit.names)
str(original_data)
str(W)
SC <- left_join(original_data, select(W, -unit.numbers), by = "abbr_state") %>%
na.omit() %>%
group_by(year) %>%
summarise(sc = weighted.mean(score, w.weights))
PE <- original_data %>%
filter(abbr_state == "PE") %>%
select(year, score)
GAP <- left_join(PE, SC, by = "year") %>%
mutate(gap = score - sc)
GAP$grade <- unique(original_data$grade)
GAP$subject <- unique(original_data$subject)
GG_DATA <- left_join(PE, SC, by = "year") %>%
pivot_longer(!year, names_to = "unit", values_to = "score")
GG_DATA$unit[GG_DATA$unit == "score"] <- "Pernambuco"
GG_DATA$unit[GG_DATA$unit == "sc"] <- "Synthetic Control"
GG_DATA$grade <- unique(original_data$grade)
GG_DATA$subject <- unique(original_data$subject)
return(list(GG_DATA, GAP))
}
#-----------------------------------------------------------------------------------------#
# Data Preparation for Plotting
#-----------------------------------------------------------------------------------------#
# Run SCM for Primary School Mathematics (PE)
SCM_PE_PM <- synth(DATA_PE_PM)
TABLES_PE_PM <- synth.tab(dataprep.res = DATA_PE_PM, synth.res = SCM_PE_PM)
# Run SCM for Primary School Portuguese (PE)
SCM_PE_PP <- synth(DATA_PE_PP)
TABLES_PE_PP <- synth.tab(dataprep.res = DATA_PE_PP, synth.res = SCM_PE_PP)
# Run SCM for Lower Secondary School Mathematics (PE)
SCM_PE_LSM <- synth(DATA_PE_LSM)
TABLES_PE_LSM <- synth.tab(dataprep.res = DATA_PE_LSM, synth.res = SCM_PE_LSM)
# Run SCM for Lower Secondary School Portuguese (PE)
SCM_PE_LSP <- synth(DATA_PE_LSP)
TABLES_PE_LSP <- synth.tab(dataprep.res = DATA_PE_LSP, synth.res = SCM_PE_LSP)
# Run SCM for Upper Secondary School Mathematics (PE)
# SCM_PE_USM <- synth(DATA_PE_USM)
# TABLES_PE_USM <- synth.tab(dataprep.res = DATA_PE_USM, synth.res = SCM_PE_USM)
# Run SCM for Upper Secondary School Portuguese (PE)
# SCM_PE_USP <- synth(DATA_PE_USP)
# TABLES_PE_USP <- synth.tab(dataprep.res = DATA_PE_USP, synth.res = SCM_PE_USP)
#-----------------------------------------------------------------------------------------#
# Graphs in ggplot for Pernambuco
PM_PE <- plot_scm_PE(PRIMARY_M, TABLES_PE_PM)
PM_PE_SC <- PM_PE[[1]]
PM_PE_GAP <- PM_PE[[2]]
PP_PE <- plot_scm_PE(PRIMARY_P, TABLES_PE_PP)
PP_PE_SC <- PP_PE[[1]]
PP_PE_GAP <- PP_PE[[2]]
LSM_PE <- plot_scm_PE(LOWERS_M, TABLES_PE_LSM)
LSM_PE_SC <- LSM_PE[[1]]
LSM_PE_GAP <- LSM_PE[[2]]
LSP_PE <- plot_scm_PE(LOWERS_P, TABLES_PE_LSP)
LSP_PE_SC <- LSP_PE[[1]]
LSP_PE_GAP <- LSP_PE[[2]]
# Combining Graph Data for Pernambuco
DATA_GRAPH_PE <- rbind(PM_PE_SC, PP_PE_SC, LSM_PE_SC, LSP_PE_SC)
# Adjusting Labels for Pernambuco data
DATA_GRAPH_PE$grade[DATA_GRAPH_PE$grade=="P"] <- "Primary Education"
DATA_GRAPH_PE$grade[DATA_GRAPH_PE$grade=="LS"] <- "Lower Secondary Education"
DATA_GRAPH_PE$grade <- factor(DATA_GRAPH_PE$grade, levels = c("Primary Education", "Lower Secondary Education"))
DATA_GRAPH_PE$subject[DATA_GRAPH_PE$subject=="math"] <- "Mathematics"
DATA_GRAPH_PE$subject[DATA_GRAPH_PE$subject=="math"] <- "Mathematics"
DATA_GRAPH_PE$subject[DATA_GRAPH_PE$subject=="port"] <- "Portuguese"
#-----------------------------------------------------------------------------------------#
# Plotting
#-----------------------------------------------------------------------------------------#
# Figure for Primary Education in PE
a_06_PE <- ggplot(data = filter(DATA_GRAPH_PE, grade == "Primary Education"), aes(x = year, y = score, color = unit)) +
geom_vline(xintercept = 2008, color = "#636363", linetype = "dashed", size = 0.9) +
geom_vline(xintercept = 2011, color = "#636363", linetype = "dashed", size = 0.9) +
geom_line(size = 0.9) +
scale_color_manual(values = c("#42B1BD", "#D26B5F"), labels = c("Pernambuco", "Synthetic Pernambuco"), name = "") +
ylab("Score") +
xlab("") +
annotate("text", x = 2007, y = 220, label = "TI", color = "#636363", size = 4) +
annotate("text", x = 2013, y = 152, label = "TI + TA", color = "#636363", size = 4) +
theme_bw() +
theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), legend.position = "bottom") +
facet_grid(vars(grade), vars(subject))
# Figure for Lower Secondary Education in PE
b_06_PE <- ggplot(data = filter(DATA_GRAPH_PE, grade == "Lower Secondary Education"), aes(x = year, y = score, color = unit)) +
geom_vline(xintercept = 2008, color = "#636363", linetype = "dashed", size = 0.9) +
geom_vline(xintercept = 2015, color = "#636363", linetype = "dashed", size = 0.9) +
geom_line(size = 0.9) +
scale_color_manual(values = c("#42B1BD", "#D26B5F"), labels = c("Pernambuco", "Synthetic Pernambuco"), name = "") +
ylab("Score") +
xlab("Year") +
annotate("text", x = 2007, y = 250, label = "TI", color = "#636363", size = 4) +
annotate("text", x = 2017, y = 190, label = "TI + TA", color = "#636363", size = 4) +
theme_bw() +
theme(panel.grid.major = element_blank(), panel.grid.minor = element_blank(), legend.position = "bottom") +
facet_grid(vars(grade), vars(subject))
# Arrange and Save the Plots
ggarrange(a_06_PE, b_06_PE, ncol = 1, nrow = 2, common.legend = TRUE, legend = "bottom")
ggsave(filename = "figure_PE.png", path = "plots", width = 21, height = 15, units = "cm")
```