-
Notifications
You must be signed in to change notification settings - Fork 0
/
google_search_cy_analysis_AI_search.R
330 lines (262 loc) · 17.9 KB
/
google_search_cy_analysis_AI_search.R
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
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
## google_search_cy_analysis_AI_search.R
##########################################################################
##
## Authors: Geoff Dancy and Christopher J. Fariss
##
## Title: "The Search for Human Rights: A Global Analysis Using Google Data"
##
## Contact Information:
## Geoff Dancy <geoff.dancy@utoronto.ca>
## Christopher J. Fariss <cjf0006@gmail.com>
##
## Copyright (c) 2022, under the Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License.
## For more information see: http://creativecommons.org/licenses/by-nc-sa/3.0/us/
## All rights reserved.
##
##########################################################################
## load necessary libraries
## change groundhog to TRUE to install original versions of libraries from April-2022
source("groundhog_library_func.R")
groundhog_library_func(groundhog=FALSE, regular_install=FALSE)
## additional functions to add content to the stargazer table output
source("Rcode_Human_Rights_google_search_country/star_cleaning_functions.R")
## change this logical-value for the alternative count variable
attention_rate <- FALSE ## set to FALSE for count
## keep these values the same (see the other files for alternative model specifications)
## That is don't change these values look at the other files for the alternative model specifications
NGO_amnesty <- TRUE ## set to FALSE for the alternative NGO variable
lowsearch_option <- FALSE ## note that the lowsearch term in the file label means that the lowsearch argument is set to its default FALSE value in the gtrends() search function which is FALSE (sorry this is confusing)
if(lowsearch_option==TRUE) test_dat <- readRDS("Data_output/combined_gsearch_dat_list_merged_AI.RDS")
if(lowsearch_option==FALSE) test_dat <- readRDS("Data_output/combined_gsearch_dat_list_merged_lowsearch_AI.RDS")
## inspect object
length(test_dat)
head(test_dat[[1]])
summary(test_dat[[1]])
#test_dat$hits_mean[is.na(test_dat$hits_mean)] <- 0
#test_dat$hits_median[is.na(test_dat$hits_median)] <- 0
#test_dat$hits_max[is.na(test_dat$hits_max)] <- 0
#test_dat$hits_sd[is.na(test_dat$hits_sd)] <- 0
## linear models (languages are seperated)
fit_mean_robust <- fit_median_robust <- fit_max_robust <- list()
unit_n <- c()
for(i in 1:length(test_dat)){
temp <- subset(test_dat[[i]], !is.na(CCODE))
## scale variable
temp$theta_mean <- -1*temp$theta_mean
temp$amnesty_attention_count <- scale(temp$amnesty_attention_count)
temp$amnesty_attention_rate <- scale(temp$amnesty_attention_rate)
temp$GDP_growth_annual_percent <- scale(temp$GDP_growth_annual_percent)
temp$Foreign_direct_investment_net_inflows_percent_GDP <- scale(temp$Foreign_direct_investment_net_inflows_percent_GDP)
temp$treaty_count <- scale(temp$treaty_count)
temp$v2smgovfilprc <- scale(temp$v2smgovfilprc)
temp$hringo_inter <- scale(temp$hringo_inter)
temp$hringo_inter_rate <- scale(temp$hringo_inter_rate)
temp$hringo_inter_rate <- scale(temp$hringo_inter_rate)
if(attention_rate==FALSE & NGO_amnesty==TRUE) temp$NGO <- temp$amnesty_attention_count
if(attention_rate==TRUE & NGO_amnesty==TRUE) temp$NGO <- temp$amnesty_attention_rate
if(attention_rate==FALSE & NGO_amnesty==FALSE) temp$NGO <- temp$hringo_inter
if(attention_rate==TRUE & NGO_amnesty==FALSE) temp$NGO <- temp$hringo_inter_rate
fit_mean <- lm(hits_mean ~ Foreign_direct_investment_net_inflows_percent_GDP
+ NGO
+ treaty_count
+ GDP_growth_annual_percent
+ theta_mean
+ v2smgovfilprc
,
data=temp)
fit_median <- lm(hits_median ~ Foreign_direct_investment_net_inflows_percent_GDP
+ NGO
+ treaty_count
+ GDP_growth_annual_percent
+ theta_mean
+ v2smgovfilprc
,
data=temp)
fit_max <- lm(hits_max ~Foreign_direct_investment_net_inflows_percent_GDP
+ NGO
+ treaty_count
+ GDP_growth_annual_percent
+ theta_mean
+ v2smgovfilprc
,
data=temp)
#summary(fit)
## robust clustered standard errors
fit_mean_robust[[i]] <- coeftest(fit_mean, vcov = vcovHC, type = "HC1", cluster=~CCODE)
fit_median_robust[[i]] <- coeftest(fit_median, vcov = vcovHC, type = "HC1", cluster=~CCODE)
fit_max_robust[[i]] <- coeftest(fit_max, vcov = vcovHC, type = "HC1", cluster=~CCODE)
unit_n[i] <- length(fit_mean$fitted.values)
}
## print models to screen
unit_n
fit_mean_robust
fit_median_robust
fit_max_robust
## pool the datasets by language group
for(i in 1:length(test_dat)){
if(i>0 & i<=4) test_dat[[i]]$language <- "Amnesty_International" ## English and French
if(i>4 & i<=8) test_dat[[i]]$language <- "Amnistia_Internacional" ## Spanish
if(i>8 & i<=12) test_dat[[i]]$language <- "Anistia_Internacional" ## Portuguese
if(i>12 & i<=16) test_dat[[i]]$language <- "munazamat_aleafw_alduwalia" ## Arabic
}
test_dat_language_pooled <- list()
test_dat_language_pooled[[1]] <- rbind(test_dat[[1]], test_dat[[5]], test_dat[[9]], test_dat[[13]])
test_dat_language_pooled[[2]] <- rbind(test_dat[[2]], test_dat[[6]], test_dat[[10]], test_dat[[14]])
test_dat_language_pooled[[3]] <- rbind(test_dat[[3]], test_dat[[7]], test_dat[[11]], test_dat[[15]])
test_dat_language_pooled[[4]] <- rbind(test_dat[[4]], test_dat[[8]], test_dat[[12]], test_dat[[16]])
## linear models (languages are pooled together)
fit_mean_robust <- fit_median_robust <- fit_max_robust <- r_squared <- list()
unit_n <- c()
for(i in 1:length(test_dat_language_pooled)){
temp <- subset(test_dat_language_pooled[[i]], !is.na(CCODE))
## scale variable
temp$theta_mean <- -1*temp$theta_mean
temp$amnesty_attention_count <- scale(temp$amnesty_attention_count)
temp$amnesty_attention_rate <- scale(temp$amnesty_attention_rate)
temp$GDP_growth_annual_percent <- scale(temp$GDP_growth_annual_percent)
temp$Foreign_direct_investment_net_inflows_percent_GDP <- scale(temp$Foreign_direct_investment_net_inflows_percent_GDP)
temp$treaty_count <- scale(temp$treaty_count)
temp$v2smgovfilprc <- scale(temp$v2smgovfilprc)
temp$hringo_inter <- scale(temp$hringo_inter)
temp$hringo_inter_rate <- scale(temp$hringo_inter_rate)
if(attention_rate==FALSE & NGO_amnesty==TRUE) temp$NGO <- temp$amnesty_attention_count
if(attention_rate==TRUE & NGO_amnesty==TRUE) temp$NGO <- temp$amnesty_attention_rate
if(attention_rate==FALSE & NGO_amnesty==FALSE) temp$NGO <- temp$hringo_inter
if(attention_rate==TRUE & NGO_amnesty==FALSE) temp$NGO <- temp$hringo_inter_rate
fit_mean <- lm(hits_mean ~ -1
+ Foreign_direct_investment_net_inflows_percent_GDP
+ NGO
+ treaty_count
+ GDP_growth_annual_percent
+ theta_mean
+ v2smgovfilprc
+ as.factor(language),
data=temp)
fit_median <- lm(hits_median ~ -1 +
+ Foreign_direct_investment_net_inflows_percent_GDP
+ NGO
+ treaty_count
+ GDP_growth_annual_percent
+ theta_mean
+ v2smgovfilprc
+ as.factor(language),
data=temp)
fit_max <- lm(hits_max ~ -1
+ Foreign_direct_investment_net_inflows_percent_GDP
+ NGO
+ treaty_count
+ GDP_growth_annual_percent
+ theta_mean
+ v2smgovfilprc
+ as.factor(language),
data=temp)
r_squared[[i]] <- c(summary(fit_mean)$r.squared, summary(fit_median)$r.squared, summary(fit_max)$r.squared)
#summary(fit)
## robust clustered standard errors
fit_mean_robust[[i]] <- coeftest(fit_mean, vcov = vcovHC, type = "HC1", cluster=~language)
fit_median_robust[[i]] <- coeftest(fit_median, vcov = vcovHC, type = "HC1", cluster=~language)
fit_max_robust[[i]] <- coeftest(fit_max, vcov = vcovHC, type = "HC1", cluster=~language)
unit_n[i] <- length(fit_mean$fitted.values)
}
unit_n
## print models to screen
#summary(fit_mean)
#summary(fit_median)
#summary(fit_max)
fit_mean_robust[[4]]
fit_median_robust[[4]]
fit_max_robust[[4]]
## save table output for latex
for(j in 1:4){
if(attention_rate==FALSE & NGO_amnesty==TRUE) LABELS <- c("FDI Inflows", "NGO Presence (Count)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship", "English/French (Amnesty International)", "Spanish (Amnistia Internacional)", "Portuguese (Anistia Internacional)", "Arabic (Munazamat Aleafw Alduwalia)")
if(attention_rate==TRUE & NGO_amnesty==TRUE) LABELS <- c("FDI Inflows", "NGO Presence (Rate)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship", "English/French (Amnesty International)", "Spanish (Amnistia Internacional)", "Portuguese (Anistia Internacional)", "Arabic (Munazamat Aleafw Alduwalia)")
if(attention_rate==FALSE & NGO_amnesty==FALSE) LABELS <- c("FDI Inflows", "NGO Presence (Count)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship", "English/French (Amnesty International)", "Spanish (Amnistia Internacional)", "Portuguese (Anistia Internacional)", "Arabic (Munazamat Aleafw Alduwalia)")
if(attention_rate==TRUE & NGO_amnesty==FALSE) LABELS <- c("FDI Inflows", "NGO Presence (Rate)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship", "English/French (Amnesty International)", "Spanish (Amnistia Internacional)", "Portuguese (Anistia Internacional)", "Arabic (Munazamat Aleafw Alduwalia)")
NOTES <- "\\parbox[t]{8cm}{``Amnesty International' searching: Search Mean, Search Median, and Search Max dependent variables are measure of the yearly mean, median, or max country-week search rate value for each country-year unit. Independent variables are measured annually for each country-year unit. $^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01}"
tab_output <- stargazer(fit_mean_robust[[j]], fit_median_robust[[j]], fit_max_robust[[j]], covariate.labels = LABELS, column.labels = c("Search Mean", "Search Median", "Search Max"), dep.var.caption = "Dependent Variable: Google Search Rate", no.space=TRUE, title="Country-Year Regression Analysis", notes=NOTES, notes.align="l", notes.append=FALSE)
tab_output <- star_insert_row(tab_output, paste("R-squared &", round(r_squared[[i]][1],3), "&", round(r_squared[[i]][2],3), "&", round(r_squared[[i]][3],3),"\\\\ "), insert.after=36)
tab_output <- star_insert_row(tab_output, paste("sample size &", unit_n[1], "&", unit_n[2], "&", unit_n[3],"\\\\ "), insert.after=37)
#tab_output
if(j==1 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_count_2012_2016.tex")
if(j==2 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_count_2013_2017.tex")
if(j==3 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_count_2014_2018.tex")
if(j==4 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_count_2015_2019.tex")
if(j==1 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_rate_2012_2016.tex")
if(j==2 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_rate_2013_2017.tex")
if(j==3 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_rate_2014_2018.tex")
if(j==4 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) write(tab_output, file="Tex_tables/main_results_lowsearch_AI_amnesty_report_rate_2015_2019.tex")
}
## save coefficient plots
for(j in 1:4){
coef_mean <- fit_mean_robust[[j]][6:1,1]
coef_median <- fit_median_robust[[j]][6:1,1]
coef_max <- fit_max_robust[[j]][6:1,1]
## select value for first interval
VALUE <- 1
VALUE <- qt(.95, unit_n[j])
coef_mean_1ub <- fit_mean_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_median_1ub <- fit_median_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_max_1ub <- fit_max_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_mean_1lb <- fit_mean_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
coef_median_1lb <- fit_median_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
coef_max_1lb <- fit_max_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
## select value for second interval
VALUE <- 2
VALUE <- qt(.975, unit_n[j])
coef_mean_2ub <- fit_mean_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_median_2ub <- fit_median_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_max_2ub <- fit_max_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_mean_2lb <- fit_mean_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
coef_median_2lb <- fit_median_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
coef_max_2lb <- fit_max_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
## select value for third *optional) interval
VALUE <- 3
VALUE <- qt(.995, unit_n[j])
coef_mean_3ub <- fit_mean_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_median_3ub <- fit_median_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_max_3ub <- fit_max_robust[[j]][6:1,1] + VALUE*fit_mean_robust[[j]][6:1,2]
coef_mean_3lb <- fit_mean_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
coef_median_3lb <- fit_median_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
coef_max_3lb <- fit_max_robust[[j]][6:1,1] - VALUE*fit_mean_robust[[j]][6:1,2]
cbind(coef_mean, coef_median, coef_max)
## coefficient plots
## create graph file
if(j==1 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_count_2012_2016_Rplot.pdf", height=6, width=8)
if(j==2 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_count_2013_2017_Rplot.pdf", height=6, width=8)
if(j==3 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_count_2014_2018_Rplot.pdf", height=6, width=8)
if(j==4 & lowsearch_option==FALSE & attention_rate==FALSE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_count_2015_2019_Rplot.pdf", height=6, width=8)
if(j==1 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_rate_2012_2016_Rplot.pdf", height=6, width=8)
if(j==2 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_rate_2013_2017_Rplot.pdf", height=6, width=8)
if(j==3 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_rate_2014_2018_Rplot.pdf", height=6, width=8)
if(j==4 & lowsearch_option==FALSE & attention_rate==TRUE & NGO_amnesty==TRUE) pdf("Rplots_coefplots/coefs_lowsearch_AI_amnesty_amnesty_report_rate_2015_2019_Rplot.pdf", height=6, width=8)
par(mar=c(4,10,2,1))
plot(coef_mean, 1:6, xlim=c(-6,6), ylim=c(.5,6.5), ylab="", xlab="Standarized Coefficients", xaxt="n", yaxt="n", type="n")
abline(v=0, col=grey(.5), lty=2)
abline(h=1.5, col=grey(.85), lty=1, lwd=.5)
abline(h=2.5, col=grey(.85), lty=1, lwd=.5)
abline(h=3.5, col=grey(.85), lty=1, lwd=.5)
abline(h=4.5, col=grey(.85), lty=1, lwd=.5)
abline(h=5.5, col=grey(.85), lty=1, lwd=.5)
for(i in 1:7){
lines(c(coef_mean_1ub[i], coef_mean_1lb[i]), c(i+.25,i+.25), lwd=2)
lines(c(coef_mean_2ub[i], coef_mean_2lb[i]), c(i+.25,i+.25), lwd=1.25)
#lines(c(coef_mean_3ub[i], coef_mean_3lb[i]), c(i+.25,i+.25), lwd=.5)
lines(c(coef_median_1ub[i], coef_median_1lb[i]), c(i,i), lwd=2)
lines(c(coef_median_2ub[i], coef_median_2lb[i]), c(i,i), lwd=1.25)
#lines(c(coef_median_3ub[i], coef_median_3lb[i]), c(i,i), lwd=.5)
lines(c(coef_max_1ub[i], coef_max_1lb[i]), c(i-.25,i-.25), lwd=2)
lines(c(coef_max_2ub[i], coef_max_2lb[i]), c(i-.25,i-.25), lwd=1.5)
#lines(c(coef_max_3ub[i], coef_max_3lb[i]), c(i-.25,i-.25), lwd=.5)
}
points(coef_mean, 1.25:6.25, col=1, bg=grey(.25), pch=21)
points(coef_median, 1:6, col=1, bg=grey(.65), pch=23)
points(coef_max, 0.75:5.75, col=1, bg=grey(.95), pch=22)
mtext(side=3, line=0.5, text="Lower <----------- Google Search Rate -----------> Higher")
if(attention_rate==FALSE & NGO_amnesty==TRUE) axis(side=2, at=6:1, labels=c("FDI Inflows", "Amnesty Report (Count)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship"), las=2)
if(attention_rate==TRUE & NGO_amnesty==TRUE) axis(side=2, at=6:1, labels=c("FDI Inflows", "Amnesty Report (Rate)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship"), las=2)
if(attention_rate==FALSE & NGO_amnesty==FALSE) axis(side=2, at=6:1, labels=c("FDI Inflows", "NGO Presence (Count)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship"), las=2)
if(attention_rate==TRUE & NGO_amnesty==FALSE) axis(side=2, at=6:1, labels=c("FDI Inflows", "NGO Presence (Rate)", "HR Treaty Ratifications", "GDP Growth", "HR Violations", "Internet Censorship"), las=2)
axis(side=1, at=-6:6)
legend("topleft", legend=c("mean rate", "median rate", "max rate"), pch=c(21, 23, 22), col=1, pt.bg=c(grey(.25), grey(.65), grey(.95)), bty="n", cex=.75)
dev.off()
}