add 32

32/paper.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcf5bb75e799ab4e1e3711ba7844a57acaa2d088d432e5da697abba065bb4d3f
+size 411956

32/replication_package/1-ReadMe.txt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d5234931ac466ad3e769724309900ec8438f8fe1481002da79633f1fbca043a
+size 7074

32/replication_package/ContextAnalysis_Appendix.R

@@ -0,0 +1,719 @@
+# Replication File for Appendix Survey Analyses
+# Table C1 in Appendix C1: The Effect of Excess Males on the Probability of Observing at least One Hate Crime
+# Table C2 in Appendix C2: The Effect of Excess Males on the Probability of Observing at least One Hate Crime (Different Definition of “Excess Males”)
+# Table C3 in Appendix C3: The Effect of Excess Males on the Probability of Observing at least One Hate Crime (linear probability model)
+# Table C4 in Appendix C4: The Effect of Excess Males on the Probability of Observing at least One Physical Attack
+# Table C5 in Appendix C5: Negative Binomial Regression
+# Table C6 in Appendix C6: Interaction  between  Excess  Males  and  East/West  Germany
+# Table C7 in Appendix C7: Interaction between Excess Males and Refugee Inflow
+# Table C9 in Appendix C9: Placebo Analysis
+# Appendix C10: Descriptive Statistics
+
+# R version 4.0.2 (2020-06-22)
+
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+# install.packages("stargazer") # stargazer_5.2.2
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+require(stargazer) # stargazer_5.2.2
+source("Help.R")
+
+dat <- read.dta13("context.dta")
+
+dat_2015 <- dat[dat$year == 2015, ]
+dat_2016 <- dat[dat$year == 2016, ]
+dat_2017 <- dat[dat$year == 2017, ]
+dat_2015$Hate_all_muni_1517 <- dat_2015$Hate_all_muni + dat_2016$Hate_all_muni + dat_2017$Hate_all_muni
+dat_2015$Hate_all_muni_1517_bin <- ifelse(dat_2015$Hate_all_muni_1517 > 0, 1, 0)
+
+# Remove Extreme Value of Excess Males 
+range_x <- quantile(dat_2015$pop_15_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s <- dat_2015[dat_2015$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+                         dat_2015$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+dat_s <- dat[dat$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+               dat$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+
+# ##########################################
+# Main Table (Table C1 in Appendix C1) 
+# ##########################################
+bin_1_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                           pop_15_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           as.factor(ags_county), # county fixed effects
+                         id  = "ags_county", data = dat_2015_s)
+
+bin_1_p <- bin.summary(Hate_all_muni_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         as.factor(ags_county) + as.factor(year), # county + year fixed effects
+                       id  = "ags_county", data = dat_s)
+
+bin_2_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                           pop_15_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           as.factor(ags_county), # county fixed effects
+                         id  = "ags_county", data = dat_2015_s)
+
+bin_2_p <- bin.summary(Hate_all_muni_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                         as.factor(ags_county) + as.factor(year), # county + year fixed effects
+                       id  = "ags_county", data = dat_s)
+
+bin_3_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                           pop_15_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                           pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                           unemp_gendergap_2015 + 
+                           as.factor(ags_state), # state fixed effects
+                         id  = "ags_county", data = dat_2015_s)
+
+bin_3_p <- bin.summary(Hate_all_muni_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                         log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                         pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                         unemp_gendergap_2015 + 
+                         as.factor(ags_state) + as.factor(year), # state + year fixed effects
+                       id  = "ags_county", data = dat_s)
+
+fit_list <- list(bin_1_sum$fit, bin_1_p$fit,  
+                 bin_2_sum$fit, bin_2_p$fit,  
+                 bin_3_sum$fit, bin_3_p$fit)
+se_list  <- list(sqrt(diag(bin_1_sum$vcov)), sqrt(diag(bin_1_p$vcov)),
+                 sqrt(diag(bin_2_sum$vcov)), sqrt(diag(bin_2_p$vcov)),
+                 sqrt(diag(bin_3_sum$vcov)), sqrt(diag(bin_3_p$vcov)))
+
+star_out(stargazer(fit_list, se = se_list,
+                   covariate.labels = c("Excess Males (Age 15 - 44)", 
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage"),
+                   keep=c("pop_15_44_muni_gendergap_2015", 
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_ref_inflow_1514", "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015")),
+         name = "table_C1.tex")
+
+# ###############################################
+# Replicate Table with 25-44 (Table C2 in C2)
+# ###############################################
+range_x2 <- quantile(dat_2015$pop_25_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s2 <- dat_2015[dat_2015$pop_25_44_muni_gendergap_2015 >= range_x2[1] & 
+                          dat_2015$pop_25_44_muni_gendergap_2015 <= range_x2[2], ]
+dat_s2 <- dat[dat$pop_25_44_muni_gendergap_2015 >= range_x2[1] & 
+                dat$pop_25_44_muni_gendergap_2015 <= range_x2[2], ]
+
+
+bin_r_1_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                             pop_25_44_muni_gendergap_2015 + 
+                             log_population_muni_2015 + log_popdens_muni_2015 + 
+                             as.factor(ags_county),
+                           id  = "ags_county", data = dat_2015_s2)
+
+bin_r_1_p <- bin.summary(Hate_all_muni_bin ~ 
+                           pop_25_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           as.factor(ags_county) + as.factor(year),
+                         id  = "ags_county", data = dat_s2)
+
+bin_r_2_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                             pop_25_44_muni_gendergap_2015 + 
+                             log_population_muni_2015 + log_popdens_muni_2015 + 
+                             log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                             as.factor(ags_county),
+                           id  = "ags_county", data = dat_2015_s2)
+
+bin_r_2_p <- bin.summary(Hate_all_muni_bin ~ 
+                           pop_25_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           as.factor(ags_county) + as.factor(year),
+                         id  = "ags_county", data = dat_s2)
+
+bin_r_3_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                             pop_25_44_muni_gendergap_2015 + 
+                             log_population_muni_2015 + log_popdens_muni_2015 + 
+                             log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                             log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                             pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                             unemp_gendergap_2015 + 
+                             as.factor(ags_state),
+                           id  = "ags_county", data = dat_2015_s2)
+
+bin_r_3_p <- bin.summary(Hate_all_muni_bin ~ 
+                           pop_25_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                           pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                           unemp_gendergap_2015 + 
+                           as.factor(ags_state) + as.factor(year),
+                         id  = "ags_county", data = dat_s2)
+
+
+## Table C2 in Appendix C2
+fit_list2 <- list(bin_r_1_sum$fit, bin_r_1_p$fit,  
+                  bin_r_2_sum$fit, bin_r_2_p$fit,  
+                  bin_r_3_sum$fit, bin_r_3_p$fit)
+se_list2  <- list(sqrt(diag(bin_r_1_sum$vcov)), sqrt(diag(bin_r_1_p$vcov)),
+                  sqrt(diag(bin_r_2_sum$vcov)), sqrt(diag(bin_r_2_p$vcov)),
+                  sqrt(diag(bin_r_3_sum$vcov)), sqrt(diag(bin_r_3_p$vcov)))
+
+star_out(stargazer(fit_list2, se = se_list2,
+                   covariate.labels = c("Excess Males (Age 25 - 44)", 
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage"),
+                   keep=c("pop_25_44_muni_gendergap_2015", 
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_ref_inflow_1514", "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015")),
+         name = "table_C2.tex")
+
+# #####################################
+# Linear Probability Model (Table C3 in Appendix C3)
+# #####################################
+lm_1_sum <- lm.summary(Hate_all_muni_1517_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         as.factor(ags_county),
+                       id  = "ags_county", data = dat_2015_s)
+
+lm_1_p <- lm.summary(Hate_all_muni_bin ~ 
+                       pop_15_44_muni_gendergap_2015 + 
+                       log_population_muni_2015 + log_popdens_muni_2015 + 
+                       as.factor(ags_county) + as.factor(year),
+                     id  = "ags_county", data = dat_s)
+
+lm_2_sum <- lm.summary(Hate_all_muni_1517_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                         as.factor(ags_county),
+                       id  = "ags_county", data = dat_2015_s)
+
+lm_2_p <- lm.summary(Hate_all_muni_bin ~ 
+                       pop_15_44_muni_gendergap_2015 + 
+                       log_population_muni_2015 + log_popdens_muni_2015 + 
+                       log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                       as.factor(ags_county) + as.factor(year),
+                     id  = "ags_county", data = dat_s)
+
+lm_3_sum <- lm.summary(Hate_all_muni_1517_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                         log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                         pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                         unemp_gendergap_2015 + 
+                         as.factor(ags_state),
+                       id  = "ags_county", data = dat_2015_s)
+
+lm_3_p <- lm.summary(Hate_all_muni_bin ~ 
+                       pop_15_44_muni_gendergap_2015 + 
+                       log_population_muni_2015 + log_popdens_muni_2015 + 
+                       log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                       log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                       pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                       unemp_gendergap_2015 + 
+                       as.factor(ags_state) + as.factor(year),
+                     id  = "ags_county", data = dat_s)
+
+
+fit_list_lm <- list(lm_1_sum$fit, lm_1_p$fit,  
+                    lm_2_sum$fit, lm_2_p$fit,  
+                    lm_3_sum$fit, lm_3_p$fit)
+se_list_lm  <- list(sqrt(diag(lm_1_sum$vcov)), sqrt(diag(lm_1_p$vcov)),
+                    sqrt(diag(lm_2_sum$vcov)), sqrt(diag(lm_2_p$vcov)),
+                    sqrt(diag(lm_3_sum$vcov)), sqrt(diag(lm_3_p$vcov)))
+
+star_out(stargazer(fit_list_lm, se = se_list_lm,
+                   covariate.labels = c("Excess Males (Age 15 - 44)", 
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage"),
+                   keep=c("pop_15_44_muni_gendergap_2015", 
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_ref_inflow_1514", "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015")),
+         name = "table_C3.tex")
+
+
+## ############################################
+## Physical Attacks (Table C4 in Appendix C4)
+## ############################################
+dat_2015$Physical_muni_1517 <- dat_2015$Physical_muni + dat_2016$Physical_muni + dat_2017$Physical_muni
+dat_2015$Physical_muni_1517_bin <- ifelse(dat_2015$Physical_muni_1517 > 0, 1, 0)
+range_x <- quantile(dat_2015$pop_15_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s <- dat_2015[dat_2015$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+                         dat_2015$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+dat_s <- dat[dat$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+               dat$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+
+bin_phys_1_sum <- bin.summary(Physical_muni_1517_bin ~ 
+                                pop_15_44_muni_gendergap_2015 + 
+                                log_population_muni_2015 + log_popdens_muni_2015 + 
+                                as.factor(ags_county),
+                              id  = "ags_county", data = dat_2015_s)
+
+bin_phys_1_p <- bin.summary(Physical_muni_bin ~ 
+                              pop_15_44_muni_gendergap_2015 + 
+                              log_population_muni_2015 + log_popdens_muni_2015 + 
+                              as.factor(ags_county) + as.factor(year),
+                            id  = "ags_county", data = dat_s)
+
+bin_phys_2_sum <- bin.summary(Physical_muni_1517_bin ~ 
+                                pop_15_44_muni_gendergap_2015 + 
+                                log_population_muni_2015 + log_popdens_muni_2015 + 
+                                log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                                as.factor(ags_county),
+                              id  = "ags_county", data = dat_2015_s)
+
+bin_phys_2_p <- bin.summary(Physical_muni_bin ~ 
+                              pop_15_44_muni_gendergap_2015 + 
+                              log_population_muni_2015 + log_popdens_muni_2015 + 
+                              log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                              as.factor(ags_county) + as.factor(year),
+                            id  = "ags_county", data = dat_s)
+
+bin_phys_3_sum <- bin.summary(Physical_muni_1517_bin ~ 
+                                pop_15_44_muni_gendergap_2015 + 
+                                log_population_muni_2015 + log_popdens_muni_2015 + 
+                                log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                                log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                                pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                                unemp_gendergap_2015 + 
+                                as.factor(ags_state),
+                              id  = "ags_county", data = dat_2015_s)
+
+bin_phys_3_p <- bin.summary(Physical_muni_bin ~ 
+                              pop_15_44_muni_gendergap_2015 + 
+                              log_population_muni_2015 + log_popdens_muni_2015 + 
+                              log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                              log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                              pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                              unemp_gendergap_2015 + 
+                              as.factor(ags_state) + as.factor(year),
+                            id  = "ags_county", data = dat_s)
+
+## Table C4 in Appendix C4
+fit_list_bin_phys <- list(bin_phys_1_sum$fit, bin_phys_1_p$fit,  
+                          bin_phys_2_sum$fit, bin_phys_2_p$fit,  
+                          bin_phys_3_sum$fit, bin_phys_3_p$fit)
+se_list_bin_phys  <- list(sqrt(diag(bin_phys_1_sum$vcov)), sqrt(diag(bin_phys_1_p$vcov)),
+                          sqrt(diag(bin_phys_2_sum$vcov)), sqrt(diag(bin_phys_2_p$vcov)),
+                          sqrt(diag(bin_phys_3_sum$vcov)), sqrt(diag(bin_phys_3_p$vcov)))
+
+star_out(stargazer(fit_list_bin_phys, se = se_list_bin_phys,
+                   covariate.labels = c("Excess Males (Age 15 - 44)", 
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage"),
+                   keep=c("pop_15_44_muni_gendergap_2015", 
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_ref_inflow_1514", "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015")),
+         name = "table_C4.tex")
+
+## ########################################
+## Appendix C5: Count Model 
+## ########################################
+rm(list=ls())
+
+dat <- read.dta13("context.dta")
+source("Help.R")
+
+dat_2015 <- dat[dat$year == 2015, ]
+dat_2016 <- dat[dat$year == 2016, ]
+dat_2017 <- dat[dat$year == 2017, ]
+dat_2015$Hate_all_muni_1517 <- dat_2015$Hate_all_muni + dat_2016$Hate_all_muni + dat_2017$Hate_all_muni
+
+# Remove Extreme Value of Excess Males 
+range_x <- quantile(dat_2015$pop_15_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s <- dat_2015[dat_2015$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+                         dat_2015$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+dat_s <- dat[dat$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+               dat$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+
+for_s <- as.formula(Hate_all_muni_1517 ~ 
+                      pop_15_44_muni_gendergap_2015 + 
+                      log_population_muni_2015 + log_popdens_muni_2015 + 
+                      log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                      log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                      pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                      unemp_gendergap_2015 + 
+                      as.factor(ags_state)) # state fixed effects
+
+for_p <- as.formula(Hate_all_muni ~ 
+                      pop_15_44_muni_gendergap_2015 + 
+                      log_population_muni_2015 + log_popdens_muni_2015 + 
+                      log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                      log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                      pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                      unemp_gendergap_2015 +  as.factor(ags_state) + as.factor(year)) # state + year fixed effects
+
+## Use boostrap to compute standard errors (Use the "out_count.rdata" to get the exact same estimates)
+## Note: the following codes take a while to run
+full_run <- FALSE
+if(full_run == TRUE){
+  # Sum 
+  nb_1_sum_b <- glm.boot(for_s, family = "negative-binomial", 
+                         data = dat_2015_s, 
+                         cluster_id = dat_2015_s$ags_county)
+  
+  # Panel
+  nb_1_p_b <- glm.boot(for_p, family = "negative-binomial", 
+                       data = dat_s, 
+                       cluster_id = dat_s$ags_county)
+  
+  fit_list_nb <- list(nb_1_sum_b$fit, nb_1_p_b$fit)
+  se_list_nb  <- list(nb_1_sum_b$se, nb_1_p_b$se)
+  
+  out_count_table <- list(fit_list_nb, se_list_nb)
+  save(out_count_table, file =  "out_count_table.rdata")
+}
+
+load(file =  "out_count_table.rdata")
+fit_list_nb <- out_count_table[[1]]
+se_list_nb <- out_count_table[[2]]
+
+star_out(stargazer(fit_list_nb, se = se_list_nb,
+                   covariate.labels = c("Excess Males (Age 15 - 44)", 
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage"),
+                   keep=c("pop_15_44_muni_gendergap_2015", 
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_ref_inflow_1514", "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015")),
+         name = "table_C5.tex")
+
+
+
+## ########################################################################
+## Replicate Tables with East/West Interaction (Table C6 in Appendix C6)
+## ########################################################################
+rm(list=ls())
+
+dat <- read.dta13("context.dta")
+source("Help.R")
+
+dat_2015 <- dat[dat$year == 2015, ]
+dat_2016 <- dat[dat$year == 2016, ]
+dat_2017 <- dat[dat$year == 2017, ]
+dat_2015$Hate_all_muni_1517 <- dat_2015$Hate_all_muni + dat_2016$Hate_all_muni + dat_2017$Hate_all_muni
+dat_2015$Hate_all_muni_1517_bin <- as.numeric(dat_2015$Hate_all_muni_1517 > 0)
+
+# Remove Extreme Value of Excess Males 
+range_x <- quantile(dat_2015$pop_15_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s <- dat_2015[dat_2015$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+                         dat_2015$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+dat_s <- dat[dat$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+               dat$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+
+dat_2015_s$west <- 1 - dat_2015_s$east
+dat_s$west <- 1 - dat_s$east
+
+bin_2_sum_ew <- bin.summary(Hate_all_muni_1517_bin ~ 
+                              pop_15_44_muni_gendergap_2015 + west + 
+                              log_population_muni_2015 + log_popdens_muni_2015 + 
+                              log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                              log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                              pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                              unemp_gendergap_2015,
+                            id  = "ags_county", data = dat_2015_s)
+
+bin_2_p_ew <- bin.summary(Hate_all_muni_bin ~ 
+                            pop_15_44_muni_gendergap_2015 + west + 
+                            log_population_muni_2015 + log_popdens_muni_2015 + 
+                            log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                            log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                            pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                            unemp_gendergap_2015 + as.factor(year),
+                          id  = "ags_county", data = dat_s)
+
+bin_3_sum_ew <- bin.summary(Hate_all_muni_1517_bin ~ 
+                              pop_15_44_muni_gendergap_2015*west + 
+                              log_population_muni_2015 + log_popdens_muni_2015 + 
+                              log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                              log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                              pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                              unemp_gendergap_2015,
+                            id  = "ags_county", data = dat_2015_s)
+
+bin_3_p_ew <- bin.summary(Hate_all_muni_bin ~ 
+                            pop_15_44_muni_gendergap_2015*west + 
+                            log_population_muni_2015 + log_popdens_muni_2015 + 
+                            log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                            log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                            pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                            unemp_gendergap_2015 + as.factor(year),
+                          id  = "ags_county", data = dat_s)
+
+fit_list_ew <- list(bin_2_sum_ew$fit, bin_2_p_ew$fit, 
+                    bin_3_sum_ew$fit, bin_3_p_ew$fit)
+se_list_ew  <- list(sqrt(diag(bin_2_sum_ew$vcov)), sqrt(diag(bin_2_p_ew$vcov)),
+                    sqrt(diag(bin_3_sum_ew$vcov)), sqrt(diag(bin_3_p_ew$vcov)))
+
+star_out(stargazer(fit_list_ew, se = se_list_ew,
+                   covariate.labels = c("Excess Males (Age 15 - 44)", "West",
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  
+                                        "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage",
+                                        "Excess Males x West"),
+                   keep=c("pop_15_44_muni_gendergap_2015", "west",
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_ref_inflow_1514", 
+                          "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015",
+                          "pop_15_44_muni_gendergap_2015:west")),
+         name = "table_C6.tex")
+
+
+## ##############################################################
+## Interaction with Refugee Inflow (Table C7 in Appendix C7)
+## ##############################################################
+bin_sum_int <- bin.summary(Hate_all_muni_1517_bin ~ 
+                             pop_15_44_muni_gendergap_2015*log_ref_inflow_1514 + 
+                             log_population_muni_2015 + log_popdens_muni_2015 + 
+                             log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                             log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                             pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                             unemp_gendergap_2015 + 
+                             as.factor(ags_state),
+                           id  = "ags_county", data = dat_2015_s)
+
+bin_p_int <- bin.summary(Hate_all_muni_bin ~ 
+                           pop_15_44_muni_gendergap_2015*log_ref_inflow_1514 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                           pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                           unemp_gendergap_2015 + 
+                           as.factor(ags_state) + as.factor(year),
+                         id  = "ags_county", data = dat_s)
+
+## Table C7 in Appendix C7
+fit_list_int <- list(bin_sum_int$fit, bin_p_int$fit)
+se_list_int  <- list(sqrt(diag(bin_sum_int$vcov)), sqrt(diag(bin_p_int$vcov)))
+
+star_out(stargazer(fit_list_int, se = se_list_int,
+                   covariate.labels = c("Excess Males (Age 15 - 44)", 
+                                        "Log (Refugee Inflow) (2014 vs 2015)",  
+                                        "Log (Population)","Log (Population Density)", 
+                                        "Log (Unemployment Rate)",
+                                        "% of population change (2011 vs 2015)", 
+                                        "Vote share for AfD (2013)", 
+                                        "Log (Refugee Size) (2014)",  
+                                        "Log (General Violence per capita)",
+                                        "% of High Education", 
+                                        "Change in Manufacturing Share (2011 vs 2015)",
+                                        "Share of Manufacturing", "Male Disadvantage",
+                                        "Excess Males × Log (Refugee Inflow)"),
+                   keep=c("pop_15_44_muni_gendergap_2015", "log_ref_inflow_1514", 
+                          "log_population_muni_2015", 
+                          "log_popdens_muni_2015", "log_unemp_all_muni_2015",
+                          "d_pop1511_muni", "vote_afd_2013_muni", 
+                          "log_pop_ref_2014",
+                          "log_violence_percap_2015", 
+                          "pc_hidegree_all2011", "d_manuf1115", "pc_manufacturing_2015", "unemp_gendergap_2015",
+                          "pop_15_44_muni_gendergap_2015:log_ref_inflow_1514")),
+         name = "table_C7.tex")
+
+# ################################################################
+# Appendix C9. Placebo Analysis
+# ###############################################################
+rm(list=ls())
+
+dat_pl <- read.dta13("context_placebo.dta") # data for placebo analysis
+source("Help.R")
+
+dat_2015_s <- dat_pl[dat_pl$year == 2015, ]
+dat_2016_s <- dat_pl[dat_pl$year == 2016, ]
+dat_2017_s <- dat_pl[dat_pl$year == 2017, ]
+
+# ##########################################
+# 2015
+# ##########################################
+# main model + Placebo Treatment 
+bin_15_sum_pl <- bin.summary(Hate_all_muni_bin ~ 
+                               pop_15_44_muni_gendergap_future + 
+                               pop_15_44_muni_gendergap_anu + 
+                               log(population_muni_anu) + log(popdens_muni_anu) + 
+                               log_unemp_all_muni_anu + d_pop_muni_anu + vote_afd_2013_muni + 
+                               log_ref_inflow_anu  + log(pop_ref_anu) + log(violence_percap_anu) + ## county level
+                               pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                               unemp_gendergap_anu + 
+                               as.factor(ags_state), # state fixed effects
+                             id  = "ags_county", data = dat_2015_s)
+
+# ##########################################
+# 2016
+# ##########################################
+# Main model + Placebo 
+bin_16_sum_pl <- bin.summary(Hate_all_muni_bin ~ pop_15_44_muni_gendergap_future + 
+                               pop_15_44_muni_gendergap_anu + 
+                               log(population_muni_anu) + log(popdens_muni_anu) + 
+                               log_unemp_all_muni_anu + d_pop_muni_anu + vote_afd_2013_muni + 
+                               log_ref_inflow_anu  + log(pop_ref_anu) + log(violence_percap_anu) + ## county level
+                               pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                               unemp_gendergap_anu + 
+                               as.factor(ags_state), # state fixed effects
+                             id  = "ags_county", data = dat_2016_s)
+
+# ##########################
+# 2017
+# #########################
+# Main model + Placebo 
+bin_17_sum_pl <- bin.summary(Hate_all_muni_bin ~ pop_15_44_muni_gendergap_future + 
+                               pop_15_44_muni_gendergap_anu + 
+                               log(population_muni_anu) + log(popdens_muni_anu) + 
+                               log_unemp_all_muni_anu + d_pop_muni_anu + vote_afd_2013_muni + 
+                               log_ref_inflow_anu  + log(pop_ref_anu) + log(violence_percap_anu) + ## county level
+                               pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                               unemp_gendergap_anu + 
+                               as.factor(ags_state), # state fixed effects
+                             id  = "ags_county", data = dat_2017_s)
+
+## ####################
+## Pooled Analysis
+## ####################
+# Final model + Placebo 
+bin_pool_sum_pl <- bin.summary(Hate_all_muni_bin ~ pop_15_44_muni_gendergap_future + 
+                                 pop_15_44_muni_gendergap_anu + 
+                                 log(population_muni_anu) + log(popdens_muni_anu) + 
+                                 log_unemp_all_muni_anu + d_pop_muni_anu + vote_afd_2013_muni + 
+                                 log_ref_inflow_anu  + log(pop_ref_anu) + log(violence_percap_anu) + ## county level
+                                 pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                                 unemp_gendergap_anu + 
+                                 as.factor(ags_state) + as.factor(year), # state + year fixed effects
+                               id  = "ags_county", data = dat_pl)
+
+# Repeat the analysis for Large Counties
+dat_pool_s_l <- dat_pl[dat_pl$population_muni_anu > 
+                         quantile(dat_pl$population_muni_anu, prob  = 0.5), ] 
+
+# Main model + Placebo 
+bin_pool_sum_pl_l <- bin.summary(Hate_all_muni_bin ~ pop_15_44_muni_gendergap_future + 
+                                   pop_15_44_muni_gendergap_anu + 
+                                   log(population_muni_anu) + log(popdens_muni_anu) + 
+                                   log_unemp_all_muni_anu + d_pop_muni_anu + vote_afd_2013_muni + 
+                                   log_ref_inflow_anu  + log(pop_ref_anu) + log(violence_percap_anu) + ## county level
+                                   pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                                   unemp_gendergap_anu + 
+                                   as.factor(ags_state) + as.factor(year), # state + year fixed effects
+                                 id  = "ags_county", data = dat_pool_s_l)
+
+# Table
+pl_fit_list_m <- list(bin_15_sum_pl$fit,
+                      bin_16_sum_pl$fit,
+                      bin_17_sum_pl$fit,
+                      bin_pool_sum_pl$fit,
+                      bin_pool_sum_pl_l$fit)
+pl_se_list_m  <- list(sqrt(diag(bin_15_sum_pl$vcov)),
+                      sqrt(diag(bin_16_sum_pl$vcov)),
+                      sqrt(diag(bin_17_sum_pl$vcov)),
+                      sqrt(diag(bin_pool_sum_pl$vcov)),
+                      sqrt(diag(bin_pool_sum_pl_l$vcov)))
+
+star_out(stargazer(pl_fit_list_m, se = pl_se_list_m,
+                   covariate.labels = c("Future-Treatment"),
+                   keep=c("pop_15_44_muni_gendergap_future")),
+         name = "table_C9.tex")
+
+
+# ##############################
+# Appendix C10. Descriptive Statistics
+# ##############################
+rm(list=ls())
+
+dat <- read.dta13("context.dta")
+
+min15 <- round(min(dat$pc_ref_male[dat$year ==  2015], na.rm = TRUE),2)
+min16 <- round(min(dat$pc_ref_male[dat$year ==  2016], na.rm = TRUE),2)
+
+pdf("figure_C10.pdf", height = 5, width = 10)
+par(mfrow = c(1, 2))
+plot(density(dat$pc_ref_male[dat$year ==  2015], na.rm  = TRUE),
+     main = "Proportion of Male Refugees (2015)", 
+     xlim = c(50, 100), xlab = "Percent of Male Refugees Among Refugees (county)")
+text(x = 90, y = 0.08, paste0("min = ", min15), font = 2)
+polygon(density(dat$pc_ref_male[dat$year ==  2015], na.rm  = TRUE)$x,
+        density(dat$pc_ref_male[dat$year ==  2015], na.rm  = TRUE)$y,col='grey80')
+
+plot(density(dat$pc_ref_male[dat$year ==  2016], na.rm  = TRUE),
+     main = "Proportion of Male Refugees (2016)", xlim = c(50, 100),
+     xlab = "Percent of Male Refugees Among Refugees (county)")
+text(x = 90, y = 0.12, paste0("min = ", min16), font = 2)
+polygon(density(dat$pc_ref_male[dat$year ==  2016], na.rm  = TRUE)$x,
+        density(dat$pc_ref_male[dat$year ==  2016], na.rm  = TRUE)$y,col='grey80')
+dev.off()
+

32/replication_package/ContextAnalysis_Main.R

@@ -0,0 +1,214 @@
+# Replication  File for 
+# Figure 1 (Effects of Excess Males on Prob of Hate Crime)
+# Appendix: Figure C8 (Effects of Male Diadvantage on Prob of Hate Crime)
+
+# R version 4.0.2 (2020-06-22)
+
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37  
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+source("Help.R")
+
+dat <- read.dta13("context.dta")
+
+dat_2015 <- dat[dat$year == 2015, ]
+dat_2016 <- dat[dat$year == 2016, ]
+dat_2017 <- dat[dat$year == 2017, ]
+dat_2015$Hate_all_muni_1517 <- dat_2015$Hate_all_muni + dat_2016$Hate_all_muni + dat_2017$Hate_all_muni
+dat_2015$Hate_all_muni_1517_bin <- ifelse(dat_2015$Hate_all_muni_1517 > 0, 1, 0)
+
+## #########################
+## Main Figure (Figure 1)
+## #########################
+# Remove Extreme Value of Excess Males 
+range_x <- quantile(dat_2015$pop_15_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s <- dat_2015[dat_2015$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+                         dat_2015$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+dat_s <- dat[dat$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+               dat$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+
+# sum
+bin_1_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                           pop_15_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                           pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                           unemp_gendergap_2015 + as.factor(ags_state),
+                         id  = "ags_county", data = dat_2015_s)
+
+# annual
+bin_1_p <- bin.summary(Hate_all_muni_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                         log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                         pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                         unemp_gendergap_2015 + as.factor(ags_state) + as.factor(year),
+                       id  = "ags_county", data = dat_s)
+
+# Excess Males
+# Effect Estimation 
+bin_1_sum_effect <- marginal_effect(bin_1_sum,
+                                    newdata = dat_2015_s, family = "logit",
+                                    main_var = "pop_15_44_muni_gendergap_2015",
+                                    difference = TRUE, 
+                                    treat_range = c(1, 1.2))
+
+bin_1_p_effect <- marginal_effect(bin_1_p,
+                                  newdata = dat_s, family = "logit",
+                                  main_var = "pop_15_44_muni_gendergap_2015",
+                                  difference = TRUE, 
+                                  treat_range = c(1, 1.2))
+
+# Dose function
+bin_1_sum_dose <- marginal_effect(bin_1_sum,
+                                  newdata = dat_2015_s, family = "logit",
+                                  main_var = "pop_15_44_muni_gendergap_2015")
+
+bin_1_p_dose <- marginal_effect(bin_1_p,
+                                newdata = dat_s, family = "logit",
+                                main_var = "pop_15_44_muni_gendergap_2015")
+
+
+# Male Diadvantage
+# Effect Estimation 
+bin_1_sum_gap <- marginal_effect(bin_1_sum,
+                                 newdata = dat_2015_s, family = "logit",
+                                 main_var = "unemp_gendergap_2015",
+                                 difference = TRUE, 
+                                 treat_range = c(1, 1.15))
+
+bin_1_p_gap <- marginal_effect(bin_1_p,
+                               newdata = dat_s, family = "logit",
+                               main_var = "unemp_gendergap_2015",
+                               difference = TRUE, 
+                               treat_range = c(1, 1.15))
+
+# Dose Function
+bin_1_sum_gap_dose <- marginal_effect(bin_1_sum,
+                                      newdata = dat_2015_s, family = "logit",
+                                      main_var = "unemp_gendergap_2015")
+
+bin_1_p_gap_dose <- marginal_effect(bin_1_p,
+                                    newdata = dat_s, family = "logit",
+                                    main_var = "unemp_gendergap_2015")
+
+# #####################
+# Plot Effects (Figure 1)
+# #####################
+point <- c(bin_1_sum_effect$out_main[2], bin_1_p_effect$out_main[2])
+high  <- c(bin_1_sum_effect$out_main[3], bin_1_p_effect$out_main[3])
+low   <- c(bin_1_sum_effect$out_main[1], bin_1_p_effect$out_main[1])
+
+## Short Panel 
+marginal_list <- list()
+marginal_list[[1]] <- bin_1_sum_dose
+marginal_list[[2]] <- bin_1_p_dose
+
+title_c <- c("Predicted Probability: Sum", "Predicted Probability: Annual")
+
+# Plot Dose function
+
+pdf("figure_1.pdf", height = 4, width = 11)
+par(mfrow = c(1,3), mar = c(4.5, 2, 4, 1), oma = c(0, 2, 0, 0))
+for(i in 1:2){
+  plot_coef_all  <- do.call("rbind", marginal_list[[i]]$out_main)
+  plot_x <-  marginal_list[[i]]$treat_range
+  if(i == 1){
+    ylim_u <- c(0.14, 0.23)
+    ylab_u <- ""
+  }
+  if(i == 2){
+    ylim_u <- c(0.06, 0.11)
+    ylab_u <- ""
+  }
+  
+  plot(plot_x, plot_coef_all[ ,2], pch = 19, 
+       main = paste("", title_c[i], sep = ""),
+       ylim = ylim_u, 
+       xlab = "Excess Males", 
+       ylab = ylab_u,
+       col = "black", cex = 2, type = "l", lwd = 3, cex.lab = 1.5, cex.main = 1.5, cex.axis = 1.5)
+  lines(plot_x, plot_coef_all[ ,1], col = "black", lty = 2)
+  lines(plot_x, plot_coef_all[ ,3], col = "black", lty = 2)
+  abline(h = marginal_list[[i]]$sample, lty = 2, col = "red", lwd = 2)
+  polygon(c(plot_x, rev(plot_x)), c(plot_coef_all[ ,1], rev(plot_coef_all[ ,3])),
+          col = adjustcolor("black", 0.2), border = NA)
+  par(new=TRUE)
+  hist(dat_2015_s$pop_15_44_muni_gendergap_2015, freq = FALSE, 
+       breaks = seq(from = 0, to = 6, by = 0.01), xlim = c(min(plot_x), max(plot_x)), 
+       xaxt = "n", yaxt = "n", xlab = "", ylab = "", ylim = c(0, 40), main ="")
+}
+par(mar = c(4.5, 5, 4, 1))
+plot(seq(1:2), point, ylim = c(-0.005, 0.045), 
+     ylab = "Effects on Prob (hate crime)", 
+     main = "Effects of Excess Males",
+     xlim = c(0.5, 2.5), xlab = "Outcome Types", xaxt = "n", pch = c(19, 15), 
+     cex.lab = 1.5, cex.axis = 1.5, cex.main = 1.5)
+segments(seq(1:6), low, seq(1:6), high, lwd = 2, c(rep("black",2), rep("black",2), rep("black",2)))
+Axis(side = 1, at = c(1, 2), labels = c("Sum", "Annual"), cex.axis = 1.5)
+abline(h = 0, lty = 2)
+mtext(side = 2, at = 0.5, "Prob (hate crime)", outer = TRUE, line = 0.5)
+dev.off()
+
+# ########################
+# Figure C.8 in Appendix
+# ########################
+
+# Plot Effects 
+point_g <- c(bin_1_sum_gap$out_main[2], bin_1_p_gap$out_main[2])
+high_g  <- c(bin_1_sum_gap$out_main[3], bin_1_p_gap$out_main[3])
+low_g   <- c(bin_1_sum_gap$out_main[1], bin_1_p_gap$out_main[1])
+
+## Short Panel 
+marginal_list_g <- list()
+marginal_list_g[[1]] <- bin_1_sum_gap_dose
+marginal_list_g[[2]] <- bin_1_p_gap_dose
+
+title_c <- c("Predicted Probability: Sum", "Predicted Probability: Annual")
+
+# Plot Dose function
+
+pdf("figure_C8.pdf", height = 4, width = 11)
+par(mfrow = c(1,3), mar = c(4.5, 2, 4, 1), oma = c(0, 2, 0, 0))
+for(i in 1:2){
+  plot_coef_all  <- do.call("rbind", marginal_list_g[[i]]$out_main)
+  plot_x <-  marginal_list_g[[i]]$treat_range
+  if(i <=1) ylim_u <- c(0.14, 0.22)
+  if(i > 1) ylim_u <- c(0.06, 0.11)
+  
+  plot(plot_x, plot_coef_all[ ,2], pch = 19,
+       main = paste("", title_c[i], sep = ""),
+       ylim = ylim_u,
+       xlab = "Male Disadvantage",
+       ylab = "",
+       col = "black", cex = 2, type = "l", lwd = 3, cex.lab = 1.5, cex.axis = 1.5, cex.main = 1.5)
+  lines(plot_x, plot_coef_all[ ,1], col = "black", lty = 2)
+  lines(plot_x, plot_coef_all[ ,3], col = "black", lty = 2)
+  abline(h = marginal_list_g[[i]]$sample, lty = 2, col = "red", lwd = 2)
+  polygon(c(plot_x, rev(plot_x)), c(plot_coef_all[ ,1], rev(plot_coef_all[ ,3])),
+          col = adjustcolor("black", 0.2), border = NA)
+  par(new=TRUE)
+  hist(dat_2015_s$unemp_gendergap_2015, freq = FALSE,
+       breaks = seq(from = 0, to = 6, by = 0.01), xlim = c(min(plot_x), max(plot_x)),
+       xaxt = "n", yaxt = "n", xlab = "", ylab = "", ylim = c(0, 40), main ="")
+}
+par(mar = c(4.5, 5, 4, 1))
+plot(seq(1:2), point_g, ylim = c(-0.01, 0.035), 
+     ylab = "Effects on Prob (hate crime)", 
+     main = "Effects of Male Disadvantage",
+     xlim = c(0.5, 2.5), xlab = "Outcome Types", xaxt = "n", pch = c(19, 15), 
+     cex.lab = 1.5, cex.axis = 1.5, cex.main = 1.5)
+segments(seq(1:6), low_g, seq(1:6), high_g, lwd = 2, c(rep("black",2), rep("black",2), rep("black",2)))
+Axis(side = 1, at = c(1, 2), labels = c("Sum", "Annual"), cex.axis = 1.25)
+abline(h = 0, lty = 2)
+mtext(side = 2, at = 0.5, "Prob (hate crime)", outer = TRUE, line = 0.5)
+dev.off()

32/replication_package/Help.R

@@ -0,0 +1,177 @@
+# Help functions 
+
+marginal_effect <- function(out,
+                            newdata=NULL, 
+                            main_var, family = "logit",
+                            treat_range, difference = FALSE,
+                            seed=1234){
+  
+  fit  <- out$fit
+  # Coef and VCOV 
+  coef_mar <- coef(out$fit)[is.na(coef(out$fit)) == FALSE]
+  vcov_mar <- out$vcov
+  
+  # Sample Mean of Outcomes
+  y_orig <- model.frame(formula(fit), data = newdata)[ ,1]
+  sample_mean_outcome <- mean(y_orig, na.rm = TRUE)
+  
+  # Prepare model.frame and treat_range
+  newdata_use_b <- model.frame(formula(fit), data = newdata)
+  if(missing(treat_range)){
+    treat_range   <- quantile(newdata_use_b[, main_var], 
+                              c(0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95),
+                              na.rm = TRUE)
+  }
+  
+  # Create new_treat and new_control automatically
+  newdata_use_l  <- list()
+  for(i in 1:length(treat_range)){
+    newdata_use_l_b <- newdata_use_b
+    newdata_use_l_b[ , main_var] <- treat_range[i]
+    newdata_use_l_exp <- model.matrix(formula(fit), data = newdata_use_l_b)
+    newdata_use_l_exp <- newdata_use_l_exp[, names(coef_mar)]
+    newdata_use_l[[i]] <- newdata_use_l_exp
+  }
+  
+  set.seed(seed)
+  sim_coef <- mvrnorm(n = 1000, mu=coef_mar, Sigma=vcov_mar)
+  
+  # Linear Part 
+  linear_out <- lapply(newdata_use_l, 
+                       FUN = function(x) as.matrix(sim_coef) %*% as.matrix(t(x)))
+  
+  
+  if(family %in% c("nb", "poisson")){
+    out <- lapply(linear_out, FUN = function(x) apply(exp(x), 1, mean))
+  }else if(family %in% c("ols")){
+    out <- lapply(linear_out, FUN = function(x) apply(x, 1, mean))
+  }else if(family %in% c("logit")){
+    out <- lapply(linear_out, FUN = function(x) apply((exp(x)/(1 + exp(x))), 1, mean))
+  }else{
+    warning("family should be one of 'nb', 'poisson', 'logit', and 'ols'")
+  }
+  
+  if(difference == FALSE){
+    out_main <- lapply(out, FUN = function(x) c(quantile(x, c(0.025)), mean(x), quantile(x, c(0.975))))
+    names(out_main) <- treat_range
+    out_main_percent <- lapply(out, 
+                               FUN = function(x) 
+                                 c(quantile(x, c(0.025)), mean(x), quantile(x, c(0.975)))/sample_mean_outcome)
+    names(out_main_percent) <- treat_range
+  }else if(difference == TRUE){
+    out_b <- out[[2]]  - out[[1]]
+    out_main <- c(quantile(out_b, c(0.025)), mean(out_b), quantile(out_b, c(0.975)), 
+                  quantile(out_b, c(0.05)), quantile(out_b, c(0.95)))
+    out_main_percent <- out_main/sample_mean_outcome
+  }
+  
+  output <- list("out" = out, 
+                 "out_main" = out_main,
+                 "out_main_percent" = out_main_percent,
+                 "sample_mean" = sample_mean_outcome,
+                 "treat_range" = treat_range)
+}
+
+bin.summary <- function(formula, print_var, id, data, digits = 3, type = "logit"){
+  var  <- all.vars(formula)
+  data_use <- model.frame( ~ ., data = data[, c(var, id)])
+  
+  if(type == "logit")  fit <- glm(formula, data = data_use, family = "binomial")
+  if(type == "probit") fit <- glm(formula, data = data_use, family = binomial(link="probit"))
+  tab_p <- coeftest(fit, vcov = vcovCL(fit, cluster = data_use[, id]))
+  
+  if(missing(print_var)) print_var <- seq(1:min(20, nrow(tab_p)))
+  
+  mat <- tab_p[print_var, ]
+  
+  sig <- rep("", length(mat[,4]))
+  sig[mat[ , 4]  < 0.001] <- "***"
+  sig[mat[ , 4]  >= 0.001 & mat[ , 4]  < 0.01] <- "**"
+  sig[mat[ , 4]  >= 0.01 & mat[ , 4]  < 0.05] <- "*"
+  sig[mat[ , 4]  >= 0.05 & mat[ , 4]  < 0.1] <- "."
+  mat <- as.data.frame(mat)
+  mat <- round(mat, digits = digits)
+  mat$Sig <- sig 
+  
+  sample_size <- length(fit$residuals)
+  
+  cat("Coefficients:\n")
+  print(mat[, c(1, 2, 4, 5)], row.names=TRUE)
+  cat(paste("(Sample Size:", sample_size, ")\n", sep = ""))
+  
+  output <- list("fit" = fit, "vcov" = vcovCL(fit, cluster = data_use[, id]),
+                 "sample" = sample_size)
+  
+  return(output)
+}
+
+lm.summary <- function(formula, print_var, id, data, digits = 3){
+  var  <- all.vars(formula)
+  data_use <- model.frame( ~ ., data = data[, c(var, id)])
+  
+  fit <- lm(formula, data = data_use)
+  tab_p <- coeftest(fit, vcov = vcovCL(fit, cluster = data_use[, id]))
+  
+  if(missing(print_var)) print_var <- seq(1:min(20, nrow(tab_p)))
+  
+  mat <- tab_p[print_var, ]
+  
+  sig <- rep("", length(mat[,4]))
+  sig[mat[ , 4]  < 0.001] <- "***"
+  sig[mat[ , 4]  >= 0.001 & mat[ , 4]  < 0.01] <- "**"
+  sig[mat[ , 4]  >= 0.01 & mat[ , 4]  < 0.05] <- "*"
+  sig[mat[ , 4]  >= 0.05 & mat[ , 4]  < 0.1] <- "."
+  mat <- as.data.frame(mat)
+  mat <- round(mat, digits = digits)
+  mat$Sig <- sig 
+  
+  sample_size <- length(fit$residuals)
+  
+  cat("Coefficients:\n")
+  print(mat[, c(1, 2, 4, 5)], row.names=TRUE)
+  cat(paste("(Sample Size:", sample_size, ")\n", sep = ""))
+  
+  output <- list("fit" = fit, "vcov" = vcovCL(fit, cluster = data_use[, id]),
+                 "sample" = sample_size)
+  
+  return(output)
+}
+
+
+glm.boot <- function(formula, data, family, cluster_id, boot = 1000, seed = 1234){
+  
+  set.seed(seed)
+  data$cluster_id <- cluster_id
+  
+  data_u <- data[is.na(data$cluster_id) ==  FALSE, ]
+  coef_boot <- c()
+  for(b in 1:boot){
+    boot_id <- sample(unique(data_u$cluster_id), size = length(unique(data_u$cluster_id)), replace=TRUE)
+    # create bootstap sample with sapply
+    boot_which <- sapply(boot_id, function(x) which(data_u$cluster_id == x))
+    data_boot <- data_u[unlist(boot_which),]
+    if(family == "poisson"){
+      glm_boot <- glm(formula, family = "poisson", data = data_boot)
+      coef_boot <-  cbind(coef_boot, summary(glm_boot)$coef[,1])
+    }else if(family == "negative-binomial"){
+      glm_nb_boot <- glm.nb(formula, data = data_boot)
+      coef_boot <-  cbind(coef_boot, summary(glm_nb_boot)$coef[,1])
+    }
+    if((b%%100) == 0) cat(paste(b, "..."))
+  }
+  if(family == "poisson"){
+    glm_boot_final <- glm(formula, family = "poisson", data = data)
+  }else if(family == "negative-binomial"){
+    glm_boot_final <- glm.nb(formula, data = data)
+  }
+  se <- apply(coef_boot, 1, sd) # bootstrap SE
+  coef <- glm_boot_final$coefficients
+  
+  output <- list("fit" = glm_boot_final, "coef" = coef,  "se" = se)
+  
+  return(output)
+}
+
+star_out <- function(out, name){
+  writeLines(capture.output(out), name)
+}

32/replication_package/SurveyAnalysis_Appendix.R

@@ -0,0 +1,810 @@
+# Replication File for Appendix: Survey Analysis
+# Appendix D2 Figure D2: Replicate Figure 3 with only among anti-refugees
+# Appendix D3 Figure D3: Variables Predicting Mate Competition vs. Other Views About Refugees
+# Appendix D4 Figures D.4.1 and D.4.2: Replicate Figure 4 with wave 1
+# Appendix D5 Table D.5: Table representation of Figure 5
+# Appendix D6 Table D.6.1, Figure.6.2, Table.D.6.3, Table.D.6.4
+# Appendix D8 Table D.8.1: Robustness Check with YouGov Survey Data
+
+# R version 4.0.2 (2020-06-22)
+
+# ##################
+# Data Preparation
+# ##################
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+# install.packages("stargazer") # stargazer_5.2.2
+# install.packages("foreign") # foreign_0.8-80
+# install.packages("list") # list_9.2
+
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6
+require(sandwich)  # sandwich_2.5-1
+require(lmtest) # lmtest_0.9-37
+require(stargazer) # stargazer_5.2.2
+require(foreign) # foreign_0.8-80
+require(list) # list_9.2
+source("Help.R")
+
+dat <- read.dta13(file =  "survey.dta")
+
+# Subset to people who pass the check
+dat_use <- dat[dat$wave == 4, ]
+
+## ###############################
+## 1: Appendix D2 Figure D2
+## ###############################
+# Replicate only among anti-refugee
+quantile(dat_use$refugee_ind, probs = 0.75)
+
+dat_use_r <- dat_use[dat_use$refugee_ind > 0.875, ]
+dat_use_r$MateComp.cont_bin <- ifelse(dat_use_r$MateComp.cont >= 3, 1, 0)
+dat_use_r$excess_c <- ifelse(dat_use_r$pop_15_44_muni_gendergap_2015 < 1.04, "1",
+                             ifelse(dat_use_r$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+dat_male_r <- dat_use_r[dat_use_r$gender == "Male" & dat_use_r$age <= 44 & dat_use_r$age >= 18, ]
+dat_male_y_r <- dat_use_r[dat_use_r$gender == "Male" & dat_use_r$age <= 40 & dat_use_r$age >= 30, ]
+
+mean_all_r <- tapply(dat_use_r$MateComp.cont_bin, dat_use_r$excess_c, mean)
+se_all_r   <- tapply(dat_use_r$MateComp.cont_bin, dat_use_r$excess_c, sd)/sqrt(table(dat_use_r$excess_c))
+
+mean_all_m_r <- tapply(dat_male_r$MateComp.cont_bin, dat_male_r$excess_c, mean)
+se_all_m_r <- tapply(dat_male_r$MateComp.cont_bin, dat_male_r$excess_c, sd)/sqrt(table(dat_male_r$excess_c))
+
+mean_all_y_r <- tapply(dat_male_y_r$MateComp.cont_bin, dat_male_y_r$excess_c, mean)
+se_all_y_r <- tapply(dat_male_y_r$MateComp.cont_bin, dat_male_y_r$excess_c, sd)/sqrt(table(dat_male_y_r$excess_c))
+
+pdf("figure_D2.pdf", height= 6, width = 17.5)
+par(mfrow = c(1, 3), mar = c(2,2,3,2), oma = c(4,4,0,0))
+plot(seq(1:3), mean_all_r, pch = 19, ylim = c(0, 1),
+     xlim = c(0.5, 3.5),
+     main = "All", xaxt = "n", xlab = "", ylab = "", cex.axis = 2.25, cex.main = 2.5,
+     cex = 2.25, cex.lab = 2.5)
+segments(seq(1:3), mean_all_r - 1.96*se_all_r,
+         seq(1:3), mean_all_r + 1.96*se_all_r, pch = 19, lwd = 3)
+Axis(side = 1, at = c(1,2,3), labels = c("1st tercile", "2nd tercile", "3rd tercile"), cex.axis = 2.25)
+
+plot(seq(1:3), mean_all_m_r, pch = 19, ylim = c(0, 1),
+     xlim = c(0.5, 3.5),
+     main = "Male (18-44)", xaxt = "n", xlab = "", ylab = "", cex.axis = 2.25, cex.main = 2.5,
+     cex = 2.25, cex.lab = 2.5)
+segments(seq(1:3), mean_all_m_r - 1.96*se_all_m_r,
+         seq(1:3), mean_all_m_r + 1.96*se_all_m_r, pch = 19, lwd = 3)
+Axis(side = 1, at = c(1,2,3), labels = c("1st tercile", "2nd tercile", "3rd tercile"), cex.axis = 2.25)
+
+plot(seq(1:3), mean_all_y_r, pch = 19, ylim = c(0, 1),
+     xlim = c(0.5, 3.5),
+     main = "Male (30 - 40)", xaxt = "n", xlab = "", ylab = "", cex.axis = 2.25, cex.main = 2.5,
+     cex = 2.25, cex.lab = 2.5)
+segments(seq(1:3), mean_all_y_r - 1.96*se_all_y_r,
+         seq(1:3), mean_all_y_r + 1.96*se_all_y_r, pch = 19, lwd = 3)
+Axis(side = 1, at = c(1,2,3), labels = c("1st tercile", "2nd tercile", "3rd tercile"), cex.axis = 2.25)
+mtext("Proportion Perceiving Mate Competition", side = 2, outer = TRUE, at = 0.5,
+      cex = 1.5, line = 1.75)
+mtext("Excess Males", side = 1, outer = TRUE, at = 0.175,
+      cex = 1.5, line = 1.75)
+mtext("Excess Males", side = 1, outer = TRUE, at = 0.5,
+      cex = 1.5, line = 1.75)
+mtext("Excess Males", side = 1, outer = TRUE, at = 0.825,
+      cex = 1.5, line = 1.75)
+dev.off()
+
+
+## ###############################
+## 2: Appendix D3 Figure D3
+## ###############################
+# Coefficients of Male x Single on Refugee Variables
+
+rm(list=ls())
+dat <- read.dta13(file =  "survey.dta")
+dat_use <- dat[dat$wave == 4, ]
+source("Help.R")
+dat_use$male <- as.numeric(dat_use$gender == "Male")
+
+# outcomes we want to analyze
+outcome_ref <- c("MateComp.cont", "JobComp.cont", "ref_integrating",
+                 "ref_citizenship","ref_reduce","ref_moredone", "ref_cultgiveup",
+                 "ref_economy", "ref_crime", "ref_terror", "ref_loc_services",
+                 "ref_loc_economy", "ref_loc_crime", "ref_loc_culture",
+                 "ref_loc_islam", "ref_loc_schools", "ref_loc_housing", "ref_loc_wayoflife")
+
+outcome_ref_name <- c("Mate competition", "Job competition", "Integration",
+                      "Citizenship for refugees","Number of refugees","More for refugees",
+                      "Culture",
+                      "Economy", "Crime", "Terrorism", "Local social services",
+                      "Local economy", "Local crime", "Local culture",
+                      "Islam", "Local school", "Housing", "Living")
+
+# Fit Ordered Logit
+lm_l <- list()
+lm_out <- list()
+male_mat <- sing_mat <- int_mat <- matrix(NA, nrow = 18, ncol = 2)
+for(i in 1:18){
+  control <- paste(outcome_ref[-i], collapse = "+")
+  for_i <- paste("as.factor(", outcome_ref[i],")", "~ male*singdivsep + ", control, sep = "")
+  lm_l[[i]]  <- polr(for_i, data = dat_use, Hess = TRUE)
+  lm_out[[i]] <- summary(lm_l[[i]])$coef
+  male_mat[i, 1:2] <- summary(lm_l[[i]])$coef["male", 1:2]
+  sing_mat[i, 1:2] <- summary(lm_l[[i]])$coef["singdivsep", 1:2]
+  int_mat[i, 1:2] <- summary(lm_l[[i]])$coef["male:singdivsep", 1:2]
+}
+rownames(int_mat) <- outcome_ref
+
+# Fit linear regression
+lm2_l <- list()
+lm2_out <- list()
+male_mat2 <- sing_mat2 <- int_mat2 <- matrix(NA, nrow = 18, ncol = 2)
+for(i in 1:18){
+  control <- paste(outcome_ref[-i], collapse = "+")
+  for_i <- paste(outcome_ref[i], "~ male*singdivsep + ", control, sep = "")
+  lm2_l[[i]]  <- lm(for_i, data = dat_use)
+  lm2_out[[i]] <- summary(lm2_l[[i]])$coef
+  male_mat2[i, 1:2] <- summary(lm2_l[[i]])$coef["male", 1:2]
+  sing_mat2[i, 1:2] <- summary(lm2_l[[i]])$coef["singdivsep", 1:2]
+  int_mat2[i, 1:2] <- summary(lm2_l[[i]])$coef["male:singdivsep", 1:2]
+}
+rownames(int_mat2) <- outcome_ref
+
+
+# Show Coefficients for Male x Single Interaction (after controlling for other refugee variables)
+# Both Ordered Logit and Linear regression
+col_p <-  rev(c("red", rep("black", 17)))
+
+pdf("figure_D3_1.pdf", height = 6, width = 8)
+par(mfrow = c(1, 2), mar = c(4, 2, 4, 1), oma = c(1, 10, 2, 2))
+plot(rev(int_mat[,1]), seq(1:18), pch = 19, xlim = c(-0.6, 1.0), ylim = c(1, 18),
+     xlab = "Coefficients", ylab = "", yaxt = "n",
+     main = "Ordered logit", col = col_p)
+segments(rev(int_mat[,1]) - 1.96*rev(int_mat[,2]), seq(1:18),
+         rev(int_mat[,1]) + 1.96*rev(int_mat[,2]), seq(1:18), col = col_p)
+abline(v = 0, lty = 2)
+
+plot(rev(int_mat2[,1]), seq(1:18), pch = 19, xlim = c(-0.3, 0.3), ylim = c(1, 18),
+     xlab = "Coefficients", ylab = "", yaxt = "n",
+     main = "Linear regression", col = col_p)
+segments(rev(int_mat2[,1]) - 1.96*rev(int_mat2[,2]), seq(1:18),
+         rev(int_mat2[,1]) + 1.96*rev(int_mat2[,2]), seq(1:18), col = col_p)
+abline(v = 0, lty = 2)
+
+Axis(side = 2, at = seq(1:18), labels = rev(outcome_ref_name), las = 1, tick = 0,
+     outer  = TRUE, hadj = 0, line = 7.5)
+mtext(side = 3, at = 0.5, text = "Coefficients of Male x Single", cex = 1.5, font = 2, outer = TRUE)
+dev.off()
+
+# ######################################
+# Coefficients of Women's Role on Mate Competition
+# ######################################
+# Ordered Logit
+lm_l <- list()
+lm_out <- list()
+role_mat <- matrix(NA, nrow = 18, ncol = 2)
+for(i in 1:18){
+  control <- paste(outcome_ref[-i], collapse = "+")
+  for_i <- paste("as.factor(", outcome_ref[i], ")", "~ women_role + ", control, sep = "")
+  lm_l[[i]]  <- polr(for_i, data = dat_use, Hess = TRUE)
+  lm_out[[i]] <- summary(lm_l[[i]])$coef
+  role_mat[i, 1:2] <- summary(lm_l[[i]])$coef["women_role", 1:2]
+}
+rownames(role_mat) <- outcome_ref
+
+# OLS
+lm_l2 <- list()
+lm_out2 <- list()
+role_mat2 <- matrix(NA, nrow = 18, ncol = 2)
+for(i in 1:18){
+  control <- paste(outcome_ref[-i], collapse = "+")
+  for_i <- paste(outcome_ref[i], "~ women_role + ", control, sep = "")
+  lm_l2[[i]]  <- lm(for_i, data = dat_use)
+  lm_out2[[i]] <- summary(lm_l2[[i]])$coef
+  role_mat2[i, 1:2] <- summary(lm_l2[[i]])$coef["women_role", 1:2]
+}
+rownames(role_mat2) <- outcome_ref
+
+pdf("figure_D3_2.pdf", height = 6, width = 8)
+par(mfrow = c(1, 2), mar = c(4, 2, 4, 1), oma = c(1, 10, 2, 2))
+
+plot(rev(role_mat[,1]), seq(1:18), pch = 19, xlim = c(-0.3, 0.6), ylim = c(1, 18),
+     xlab = "Coefficients", ylab = "", yaxt = "n",
+     main = "Ordered logit", col = col_p)
+segments(rev(role_mat[,1]) - 1.96*rev(role_mat[,2]), seq(1:18),
+         rev(role_mat[,1]) + 1.96*rev(role_mat[,2]), seq(1:18), col = col_p)
+abline(v = 0, lty = 2)
+
+plot(rev(role_mat2[,1]), seq(1:18), pch = 19, xlim = c(-0.1, 0.15), ylim = c(1, 18),
+     xlab = "Coefficients", ylab = "", yaxt = "n",
+     main = "Linear regression", col = col_p)
+segments(rev(role_mat2[,1]) - 1.96*rev(role_mat2[,2]), seq(1:18),
+         rev(role_mat2[,1]) + 1.96*rev(role_mat2[,2]), seq(1:18), col = col_p)
+abline(v = 0, lty = 2)
+
+Axis(side = 2, at = seq(1:18), labels = rev(outcome_ref_name), las = 1, tick = 0,
+     outer  = TRUE, hadj = 0, line = 7.5)
+
+mtext(side = 3, at = 0.5, text = "Coefficients of Women's Role",
+      cex = 1.5, font = 2, outer = TRUE)
+dev.off()
+
+
+## ###################################
+## Appendix D4: Figure D.4.1 & D.4.2
+## ###################################
+# Replicate Figure 3 with Wave 1
+data.u1 <- dat[dat$wave == 1, ]
+
+data.u1$List.treat <- ifelse(data.u1$treatment_list == "Scenario 2", 1, 0)
+
+# Difference-in-Means (0.12618)
+# Message (hate_pol_message):
+# Attacks against refugee homes are sometimes necessary to make it clear to politicians that we have a refugee problem
+diff.in.means.results <- ictreg(outcome_list ~ 1, data = data.u1,
+                                treat = "List.treat", J = 3, method = "lm")
+summary(diff.in.means.results)
+
+data.u1$means_bin <- ifelse(data.u1$hate_violence_means >= 3, 1, 0)
+data.u1$condemn_bin <- ifelse(data.u1$hate_polcondemn >= 3, 1, 0)
+data.u1$justified_bin <- ifelse(data.u1$hate_justified >= 3, 1, 0)
+
+only.mean <- mean(data.u1$means_bin)
+condemn.mean <- mean(data.u1$condemn_bin)
+justified.mean <- mean(data.u1$justified_bin)
+
+only.se <- sd(data.u1$means_bin)/sqrt(length(data.u1$means_bin))
+condemn.se <- sd(data.u1$condemn_bin)/sqrt(length(data.u1$condemn_bin))
+justified.se <- sd(data.u1$justified_bin)/sqrt(length(data.u1$justified_bin))
+
+# plot different questions within the same wave
+point <- c(summary(diff.in.means.results)$par.treat, only.mean, condemn.mean, justified.mean)
+se_p  <- c(summary(diff.in.means.results)$se.treat,  only.se, condemn.se, justified.se)
+base <- barplot(point, ylim = c(0, 0.20))
+bar_name_u <- c("Message (List)", "Only Means", "Condemn", "Justified")
+bar_name <- rep("",4)
+
+# Figure D.4.1
+pdf("figure_D4_1.pdf", height = 4.5, width = 8)
+par(mar = c(4, 5, 2, 1))
+barplot(point, ylim = c(0, 0.3), names.arg = bar_name,
+        col = c(adjustcolor("red", 0.4), "gray", "gray", "gray"), cex.axis = 1.3)
+arrows(base[,1], point - 1.96*se_p, base[,1], point + 1.96*se_p,
+       lwd = 3, angle = 90, length = 0.05, code = 3,
+       col = c("red", "black", "black", "black"))
+mtext(bar_name_u[1], outer = FALSE, side = 1, at = base[1], cex = 1.2, line = 2.4)
+mtext(bar_name_u[2], outer = FALSE, side = 1, at = base[2], cex = 1.2, line = 2.4)
+mtext(bar_name_u[3], outer = FALSE, side = 1, at = base[3], cex = 1.2, line = 2.4)
+mtext(bar_name_u[4], outer = FALSE, side = 1, at = base[4], cex = 1.2, line = 2.4)
+text(x = base[1], y = 0.28, "Estimate from \nList Experiment", col = "red", font = 2)
+text(x = base[3], y = 0.28, "Direct Questions", font = 2)
+dev.off()
+
+## "Message" across Waves
+data.u1 <- dat[dat$wave == 1, ]
+data.u2 <- dat[dat$wave == 2, ]
+data.u3 <- dat[dat$wave == 3, ]
+data.u4 <- dat[dat$wave == 4, ]
+dat_all <- rbind(data.u1, data.u2, data.u3, data.u4)
+
+dat_all$hate_pol_message_bin <- ifelse(dat_all$hate_pol_message >=3, 1, 0)
+message_direct <- tapply(dat_all$hate_pol_message_bin, dat_all$wave, mean, na.rm = TRUE)[c(2,3,4)]
+message_direct_num <- tapply(dat_all$hate_pol_message_bin, dat_all$wave, function(x) sum(is.na(x)==FALSE))[c(2,3,4)]
+message_direct_se <- tapply(dat_all$hate_pol_message_bin, dat_all$wave, sd, na.rm = TRUE)[c(2,3,4)]/sqrt(message_direct_num)
+
+# plot The same question over time
+point <- c(summary(diff.in.means.results)$par.treat, message_direct)
+se_p  <- c(summary(diff.in.means.results)$se.treat, message_direct_se)
+base <- barplot(point, ylim = c(0, 0.20))
+bar_name_u <- c("Message \n(List)", "Message \n(Direct, Wave 2)",
+                "Message \n(Direct, Wave 3)", "Message \n(Direct, Wave 4)")
+bar_name <- rep("",4)
+
+# Figure D.4.2
+pdf("figure_D4_2.pdf", height = 4.5, width = 8)
+par(mar = c(4, 5, 2, 1))
+barplot(point, ylim = c(0, 0.25), names.arg = bar_name,
+        col = c(adjustcolor("red", 0.4), "gray", "gray", "gray"), cex.axis = 1.3,
+        ylab = "Proportion of respondents", cex.lab = 1.45)
+arrows(base[,1], point - 1.96*se_p, base[,1], point + 1.96*se_p,
+       lwd = 3, angle = 90, length = 0.05, code = 3,
+       col = c("red", "black", "black", "black"))
+mtext(bar_name_u[1], outer = FALSE, side = 1, at = base[1], cex = 1.2, line = 2.4)
+mtext(bar_name_u[2], outer = FALSE, side = 1, at = base[2], cex = 1.2, line = 2.4)
+mtext(bar_name_u[3], outer = FALSE, side = 1, at = base[3], cex = 1.2, line = 2.4)
+mtext(bar_name_u[4], outer = FALSE, side = 1, at = base[4], cex = 1.2, line = 2.4)
+text(x = base[1], y = 0.225, "Estimate from \nList Experiment", col = "red", font = 2)
+text(x = base[3], y = 0.225, "Direct Questions", font = 2)
+dev.off()
+
+
+# #############################
+# Appendix D5 Table D5
+# #############################
+formula.5 <-
+  as.character("hate_violence_means ~ MateComp.cont + JobComp.cont +
+               LifeSatis.cont +  factor(age_group) + factor(gender) +
+               factor(state) + factor(citizenship) + factor(marital) +
+               factor(religion) + eduyrs + factor(occupation) +
+               factor(income) + factor(household_size) + factor(self_econ) +
+               factor(ref_integrating) + factor(ref_citizenship) + factor(ref_reduce) +
+               factor(ref_moredone) + factor(ref_cultgiveup) +
+               factor(ref_economy) + factor(ref_crime) + factor(ref_terror)  +
+               factor(ref_loc_services) +  factor(ref_loc_economy) + factor(ref_loc_crime) +
+               factor(ref_loc_culture) + factor(ref_loc_islam) +
+               factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife)")
+
+formula.6 <- paste(formula.5, "factor(distance_ref) + factor(settle_ref)",
+                   "lrscale + afd + muslim_ind + afd_ind + contact_ind",
+                   sep="+", collapse="+")
+
+# with Difference Outcomes
+# hate_pol_message : "82. Support for Hate Crime_Attacks against refugee homes are somet"
+# hate_prevent_settlement : "82. Support for Hate Crime_Racist violence is defensible if it lea"
+# hate_polcondemn : "82. Support for Hate Crime_Politicians should condemn attacks agai"
+# hate_justified: "82. Support for Hate Crime_Hostility against foreigners is sometimes justified"
+
+formula.7.means   <- paste("hate_violence_means ~ ", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.message <- paste("hate_pol_message ~", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.prevent <- paste("hate_prevent_settlement ~", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.condemn <- paste("hate_polcondemn ~ ", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.justified <- paste("hate_justified ~ ", as.character(as.formula(formula.6))[3], sep = "")
+
+# output
+lm7.means <- lm(as.formula(formula.7.means), data=dat_use)
+lm7.justified <- lm(as.formula(formula.7.justified), data=dat_use)
+lm7.message <- lm(as.formula(formula.7.message), data=dat_use)
+lm7.prevent <- lm(as.formula(formula.7.prevent), data=dat_use)
+lm7.condemn <- lm(as.formula(formula.7.condemn), data=dat_use)
+
+## Table D.5 (in Appendix D.5)
+lm.list_d <- list(lm7.means, lm7.justified, lm7.message, lm7.prevent, lm7.condemn)
+star_out(stargazer(lm.list_d,
+                   covariate.labels = c("Mate Competition","Job Competition","Life Satisfaction"),
+                   keep=c("MateComp.cont","JobComp.cont","LifeSatis.cont")),
+         name = "table_D5_1.tex")
+
+## ##################################
+## Table D.5.2 (appendix) with East/West
+## ##################################
+rm(list=ls())
+# Set the directly appropriately
+
+dat <- read.dta13(file =  "survey.dta")
+source("Help.R")
+
+# Subset to wave 4
+dat_use <- dat[dat$wave == 4, ]
+{
+  dat_use$west <- 1 - dat_use$east
+
+  #  remove state
+  formula.5_int <-
+    as.character("hate_violence_means ~ MateComp.cont*west + JobComp.cont +
+               LifeSatis.cont +  factor(age_group) + factor(gender) +
+               factor(citizenship) + factor(marital) +
+               factor(religion) + eduyrs + factor(occupation) +
+               factor(income) + factor(household_size) + factor(self_econ) +
+               factor(ref_integrating) + factor(ref_citizenship) + factor(ref_reduce) +
+               factor(ref_moredone) + factor(ref_cultgiveup) +
+               factor(ref_economy) + factor(ref_crime) + factor(ref_terror)  +
+               factor(ref_loc_services) +  factor(ref_loc_economy) + factor(ref_loc_crime) +
+               factor(ref_loc_culture) + factor(ref_loc_islam) +
+               factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife)")
+
+  formula.6_int <- paste(formula.5_int, "factor(distance_ref) + factor(settle_ref)",
+                         "lrscale + afd + muslim_ind + afd_ind + contact_ind",
+                         sep="+", collapse="+")
+
+  ## Interaction with East/West
+  # with Difference Outcomes
+  # hate_pol_message: "82. Support for Hate Crime_Attacks against refugee homes are somet"
+  # hate_prevent_settlement: "82. Support for Hate Crime_Racist violence is defensible if it lea"
+  # hate_polcondemn: "82. Support for Hate Crime_Politicians should condemn attacks agai"
+  # hate_justified: "82. Support for Hate Crime_Hostility against foreigners is sometimes justified"
+
+  formula.7_int.means   <- paste("hate_violence_means ~ ",
+                                 as.character(as.formula(formula.6_int))[3], sep = "")
+  formula.7_int.message <- paste("hate_pol_message ~",
+                                 as.character(as.formula(formula.6_int))[3], sep = "")
+  formula.7_int.prevent <- paste("hate_prevent_settlement ~",
+                                 as.character(as.formula(formula.6_int))[3], sep = "")
+  formula.7_int.condemn <- paste("hate_polcondemn ~ ",
+                                 as.character(as.formula(formula.6_int))[3], sep = "")
+  formula.7_int.justified <- paste("hate_justified ~ ",
+                                   as.character(as.formula(formula.6_int))[3], sep = "")
+
+  # output
+  lm7_int.means <- lm(as.formula(formula.7_int.means), data = dat_use)
+  lm7_int.justified <- lm(as.formula(formula.7_int.justified), data=dat_use)
+  lm7_int.message <- lm(as.formula(formula.7_int.message), data=dat_use)
+  lm7_int.prevent <- lm(as.formula(formula.7_int.prevent), data=dat_use)
+  lm7_int.condemn <- lm(as.formula(formula.7_int.condemn), data=dat_use)
+
+  ## Table D.5.2 (in Appendix D.5)
+  lm.list_int <- list(lm7_int.means, lm7_int.justified, lm7_int.message, lm7_int.prevent, lm7_int.condemn)
+  star_out(stargazer(lm.list_int,
+                     covariate.labels = c("Mate Competition",
+                                          "West",
+                                          "Job Competition","Life Satisfaction",
+                                          "Mate Competition x West"),
+                     keep=c("MateComp.cont",  "west",
+                            "JobComp.cont","LifeSatis.cont",
+                            "MateComp.cont:west")),
+           name  = "table_D5_2.tex")
+}
+
+# ##########################################
+# Appendix D6: Replcate Results with Men
+# ##########################################
+rm(list=ls())
+# Set the directly appropriately
+
+dat <- read.dta13(file =  "survey.dta")
+source("Help.R")
+
+# Subset to wave 4
+dat_use <- dat[dat$wave == 4, ]
+dat_male <- dat_use[dat_use$gender == "Male",]
+dat_female <- dat_use[dat_use$gender == "Female",]
+
+# ##########################################
+# Table D.6.1: Replicate Main Models
+# ##########################################
+{
+  
+  lm1 <- lm(hate_violence_means ~ MateComp.cont, data=dat_male)
+  
+  lm2 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont, data=dat_male)
+  
+  lm3 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont + 
+              factor(age_group) +     # age group
+              factor(state) +     # state  
+              factor(citizenship) +    # german citizen
+              factor(marital) +    # marital status
+              factor(religion) +    # religious affiliation
+              eduyrs +    # education
+              factor(occupation) +    # main activity
+              factor(income) +   # income
+              factor(household_size) +   # household size
+              factor(self_econ),    # subjective social status
+            data=dat_male)
+  
+  lm4 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont + 
+              factor(age_group) +   # age group
+              factor(state) +   # state  
+              factor(citizenship) +  # german citizen
+              factor(marital) +  # marital status
+              factor(religion) +  # religious affiliation
+              eduyrs +  # education
+              factor(occupation) +  # main activity
+              factor(income) + # income
+              factor(household_size) + # household size
+              factor(self_econ) + # subjective social status
+              factor(ref_integrating) + # Refugee Index (National-level; Q73) 8 in total
+              factor(ref_citizenship) + factor(ref_reduce) + factor(ref_moredone) + factor(ref_cultgiveup) + 
+              factor(ref_economy) + factor(ref_crime) + factor(ref_terror),
+            data=dat_male)
+  
+  lm5 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont + 
+              factor(age_group) +      # age group
+              factor(state) +      # state  
+              factor(citizenship) +     # german citizen
+              factor(marital) +     # marital status
+              factor(religion) +     # religious affiliation
+              eduyrs + # education
+              factor(occupation) +     # main activity
+              factor(income) +    # income
+              factor(household_size) +    # household size
+              factor(self_econ) +    # subjective social status
+              factor(ref_integrating) + # Refugee Index (National-level; Q73) 8 in total
+              factor(ref_citizenship) + factor(ref_reduce) + factor(ref_moredone) + factor(ref_cultgiveup) + 
+              factor(ref_economy) + factor(ref_crime) + factor(ref_terror) + 
+              factor(ref_loc_services) +    # Refugee Index (Local, Q75)
+              factor(ref_loc_economy) + factor(ref_loc_crime) + factor(ref_loc_culture) + factor(ref_loc_islam) + 
+              factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife), ## end
+            data=dat_male)
+  
+  
+  # Add More Variables 
+  # lrscale  Q21  Left-Right Scale
+  # afd, Q23  Closeness to AfD
+  # muslim_ind, afd_ind, contact_ind
+  # distance_ref Q71. Distance to refugee reception centers
+  # settle_ref Q72. Settlement of refugees living in area
+  
+  formula.5 <- 
+    as.character("hate_violence_means ~ MateComp.cont + JobComp.cont + 
+               LifeSatis.cont +  factor(age_group) + 
+               factor(state) + factor(citizenship) + factor(marital) + 
+               factor(religion) + eduyrs + factor(occupation) + 
+               factor(income) + factor(household_size) + factor(self_econ) + 
+               factor(ref_integrating) + factor(ref_citizenship) + factor(ref_reduce) + 
+               factor(ref_moredone) + factor(ref_cultgiveup) + 
+               factor(ref_economy) + factor(ref_crime) + factor(ref_terror)  + 
+               factor(ref_loc_services) +  factor(ref_loc_economy) + factor(ref_loc_crime) + 
+               factor(ref_loc_culture) + factor(ref_loc_islam) + 
+               factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife)")
+  
+  formula.6 <- paste(formula.5, "factor(distance_ref) + factor(settle_ref)", 
+                     "lrscale + afd + muslim_ind + afd_ind + contact_ind", 
+                     sep="+", collapse="+") 
+  
+  lm6 <- lm(as.formula(formula.6), data=dat_male)
+}
+lm.list.table1 <- list(lm1, lm2, lm3, lm4, lm5, lm6)
+
+# Table D.6.1
+star_out(stargazer(lm.list.table1,
+                   covariate.labels = c("Mate Competition","Job Competition","Life Satisfaction"),
+                   keep=c("MateComp.cont","JobComp.cont","LifeSatis.cont")), 
+         name = "table_D6_1.tex")
+
+## ###############################################
+## Figure D.6.2: Replicating Figure 4 (with Male)
+## ###############################################
+# with Difference Outcomes
+# hate_pol_message: "82. Support for Hate Crime_Attacks against refugee homes are somet"
+# hate_prevent_settlement: "82. Support for Hate Crime_Racist violence is defensible if it lea"
+# hate_polcondemn: "82. Support for Hate Crime_Politicians should condemn attacks agai"
+# hate_justified: "82. Support for Hate Crime_Hostility against foreigners is sometimes justified"
+
+# without gender
+formula.5 <- 
+  as.character("hate_violence_means ~ MateComp.cont + JobComp.cont + 
+               LifeSatis.cont +  factor(age_group) + 
+               factor(state) + factor(citizenship) + factor(marital) + 
+               factor(religion) + eduyrs + factor(occupation) + 
+               factor(income) + factor(household_size) + factor(self_econ) + 
+               factor(ref_integrating) + factor(ref_citizenship) + factor(ref_reduce) + 
+               factor(ref_moredone) + factor(ref_cultgiveup) + 
+               factor(ref_economy) + factor(ref_crime) + factor(ref_terror)  + 
+               factor(ref_loc_services) +  factor(ref_loc_economy) + factor(ref_loc_crime) + 
+               factor(ref_loc_culture) + factor(ref_loc_islam) + 
+               factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife)")
+
+formula.6 <- paste(formula.5, "factor(distance_ref) + factor(settle_ref)", 
+                   "lrscale + afd + muslim_ind + afd_ind + contact_ind", 
+                   sep="+", collapse="+") 
+
+formula.7.means   <- paste("hate_violence_means ~ ", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.message <- paste("hate_pol_message ~", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.prevent <- paste("hate_prevent_settlement ~", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.condemn <- paste("hate_polcondemn ~ ", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.justified <- paste("hate_justified ~ ", as.character(as.formula(formula.6))[3], sep = "")
+
+# output
+lm7.means <- lm(as.formula(formula.7.means), data=dat_male)
+lm7.justified <- lm(as.formula(formula.7.justified), data=dat_male)
+lm7.message <- lm(as.formula(formula.7.message), data=dat_male)
+lm7.prevent <- lm(as.formula(formula.7.prevent), data=dat_male)
+lm7.condemn <- lm(as.formula(formula.7.condemn), data=dat_male)
+
+point <- c(coef(lm7.means)["MateComp.cont"],
+           coef(lm7.justified)["MateComp.cont"], coef(lm7.message)["MateComp.cont"],
+           coef(lm7.prevent)["MateComp.cont"], coef(lm7.condemn)["MateComp.cont"])
+
+se <- c(summary(lm7.means)$coef["MateComp.cont", 2],
+        summary(lm7.justified)$coef["MateComp.cont", 2], summary(lm7.message)$coef["MateComp.cont", 2],
+        summary(lm7.prevent)$coef["MateComp.cont", 2], summary(lm7.condemn)$coef["MateComp.cont", 2])
+
+
+pdf("figure_D6_2.pdf", height = 4, width = 8)
+par(mar = c(2,4,4,1))
+plot(seq(1:5), point, pch = 19, ylim = c(-0.05, 0.25), xlim = c(0.5, 5.5),
+     xlab = "", xaxt = "n", ylab = "Estimated Effects",
+     main = "Estimated Effects of Mate Competition (among male)", cex.lab = 1.25, cex.axis = 1.25, cex.main = 1.5)
+segments(seq(1:5), point - 1.96*se,
+         seq(1:5), point + 1.96*se, lwd = 2)
+Axis(side=1, at = seq(1:5), labels = c("Only Means", "Justified", "Message",
+                                       "Prevent", "Condemn"), cex.axis = 1.25)
+abline(h =0, lty = 2)
+dev.off()
+
+## Table D.6.3
+lm.list_d_m <- list(lm7.means, lm7.justified, lm7.message, lm7.prevent, lm7.condemn)
+star_out(stargazer(lm.list_d_m,
+                   covariate.labels = c("Mate Competition","Job Competition","Life Satisfaction"),
+                   keep=c("MateComp.cont","JobComp.cont","LifeSatis.cont")),
+         name  = "table_D6_3.tex")
+
+# ##########################################
+# Appendix D8, Table D8: YouGov analysis
+# ##########################################
+rm(list=ls())
+you_data <- read.dta13(file  = "YouGov.dta")
+source("Help.R")
+
+## (1) Main Regression
+lm1 <- lm(hate_cont ~ mate_compete +
+            age + # age
+            gender +  # gender
+            factor(sta) +  #state
+            factor(mstat) +  # Marital Status
+            reli + # religion
+            educ_aggr_rec  + # education
+            hinc +  # income
+            housz + # household size
+            pol_leftright, # leftright scale
+          data = you_data)
+summary(lm1)
+
+## (2) + Aggression Score
+lm2 <- lm(hate_cont ~
+            mate_compete +
+            age + # age
+            gender +  # gender
+            factor(sta) +  #state
+            factor(mstat) +  # Marital Status
+            reli + # religion
+            educ_aggr_rec  + # education
+            hinc +  # income
+            housz + # household size
+            pol_leftright + # leftright scale
+            angry_mean, # aggression score
+          data = you_data)
+summary(lm2)
+
+## (3) + Refugee Index
+lm3 <- lm(hate_cont ~
+            mate_compete +
+            age + # age
+            gender +  # gender
+            factor(sta) +  #state
+            factor(mstat) +  # Marital Status
+            reli + # religion
+            educ_aggr_rec  + # education
+            hinc +  # income
+            housz + # household size
+            pol_leftright + # leftright scale
+            angry_mean + # aggression score
+            ref_loc_services + ref_loc_economy + ref_loc_crime + ## Refugee Questions (9 Questions)
+            ref_loc_culture + ref_loc_islam + ref_local_job +
+            ref_loc_schools + ref_loc_housing + ref_loc_wayoflife,
+          data = you_data)
+summary(lm3)
+
+## (4) + Refugee Contact
+lm4 <- lm(hate_cont ~
+            mate_compete +
+            age + # age
+            gender +  # gender
+            factor(sta) +  #state
+            factor(mstat) +  # Marital Status
+            reli + # religion
+            educ_aggr_rec  + # education
+            hinc +  # income
+            housz + # household size
+            pol_leftright + # leftright scale
+            angry_mean + # aggression score
+            ref_loc_services + ref_loc_economy + ref_loc_crime + ## Refugee Questions (9 Questions)
+            ref_loc_culture + ref_loc_islam + ref_local_job +
+            ref_loc_schools + ref_loc_housing + ref_loc_wayoflife +
+            see_ref_road + see_ref_store + see_ref_center + ## Refugee Contact (5 Questions)
+            see_ref_school + see_ref_work,
+          data = you_data)
+summary(lm4)
+
+## (5) + AfD Score
+lm5 <- lm(hate_cont ~
+            mate_compete +
+            age + # age
+            gender +  # gender
+            factor(sta) +  #state
+            factor(mstat) +  # Marital Status
+            reli + # religion
+            educ_aggr_rec  + # education
+            hinc +  # income
+            housz + # household size
+            pol_leftright + # leftright scale
+            angry_mean + # aggression score
+            ref_loc_services + ref_loc_economy + ref_loc_crime + ## Refugee Questions (9 Questions)
+            ref_loc_culture + ref_loc_islam + ref_local_job +
+            ref_loc_schools + ref_loc_housing + ref_loc_wayoflife +
+            see_ref_road + see_ref_store + see_ref_center + ## Refugee Contact (5 Questions)
+            see_ref_school + see_ref_work +
+            afd.score, # Closeness to AfD
+          data = you_data)
+summary(lm5)
+
+star_out(stargazer(list(lm1, lm2, lm3, lm4, lm5),
+          covariate.labels = c("Mate Competition", "Aggressiveness"), keep=c("mate_compete", "angry_mean")),
+         name = "table_D8_1.tex")
+
+
+rm(list=ls())
+you_data <- read.dta13(file  = "YouGov.dta")
+you_male <- you_data[you_data$gender == levels(you_data$gender)[1], ]
+source("Help.R")
+
+{
+  ## (1) Main Regression 
+  lm1 <- lm(hate_cont ~ mate_compete + 
+              age + # age 
+              factor(sta) +  #state 
+              factor(mstat) +  # Marital Status
+              reli + # religion 
+              educ_aggr_rec  + # education 
+              hinc +  # income 
+              housz + # household size 
+              pol_leftright, # leftright scale
+            data = you_male)
+  
+  ## (2) + Aggression Score 
+  lm2 <- lm(hate_cont ~ 
+              mate_compete + 
+              age + # age 
+              factor(sta) +  #state 
+              factor(mstat) +  # Marital Status
+              reli + # religion 
+              educ_aggr_rec  + # education 
+              hinc +  # income 
+              housz + # household size 
+              pol_leftright + # leftright scale
+              angry_mean, # aggression score 
+            data = you_male)
+  
+  ## (3) + Refugee Index
+  lm3 <- lm(hate_cont ~ 
+              mate_compete + 
+              age + # age 
+              factor(sta) +  #state 
+              factor(mstat) +  # Marital Status
+              reli + # religion 
+              educ_aggr_rec  + # education 
+              hinc +  # income 
+              housz + # household size 
+              pol_leftright + # leftright scale
+              angry_mean + # aggression score 
+              ref_loc_services + ref_loc_economy + ref_loc_crime + ## Refugee Questions (9 Questions)
+              ref_loc_culture + ref_loc_islam + ref_local_job +
+              ref_loc_schools + ref_loc_housing + ref_loc_wayoflife, 
+            data = you_male)
+  summary(lm3)
+  
+  ## (4) + Refugee Contact 
+  lm4 <- lm(hate_cont ~ 
+              mate_compete + 
+              age + # age 
+              # gender +  # gender 
+              factor(sta) +  #state 
+              factor(mstat) +  # Marital Status
+              reli + # religion 
+              educ_aggr_rec  + # education 
+              hinc +  # income 
+              housz + # household size 
+              pol_leftright + # leftright scale
+              angry_mean + # aggression score 
+              ref_loc_services + ref_loc_economy + ref_loc_crime + ## Refugee Questions (9 Questions)
+              ref_loc_culture + ref_loc_islam + ref_local_job +
+              ref_loc_schools + ref_loc_housing + ref_loc_wayoflife +
+              see_ref_road + see_ref_store + see_ref_center + ## Refugee Contact (5 Questions)
+              see_ref_school + see_ref_work, 
+            data = you_male)
+  summary(lm4)
+  
+  ## (5) + AfD Score 
+  lm5 <- lm(hate_cont ~ 
+              mate_compete + 
+              age + # age 
+              # gender +  # gender 
+              factor(sta) +  #state 
+              factor(mstat) +  # Marital Status
+              reli + # religion 
+              educ_aggr_rec  + # education 
+              hinc +  # income 
+              housz + # household size 
+              pol_leftright + # leftright scale
+              angry_mean + # aggression score 
+              ref_loc_services + ref_loc_economy + ref_loc_crime + ## Refugee Questions (9 Questions)
+              ref_loc_culture + ref_loc_islam + ref_local_job +
+              ref_loc_schools + ref_loc_housing + ref_loc_wayoflife +
+              see_ref_road + see_ref_store + see_ref_center + ## Refugee Contact (5 Questions)
+              see_ref_school + see_ref_work + 
+              afd.score, # Closeness to AfD
+            data = you_male)
+  summary(lm5)
+}
+
+star_out(stargazer(list(lm1, lm2, lm3, lm4, lm5),
+                   covariate.labels = c("Mate Competition", "Aggressiveness"), 
+                   keep=c("mate_compete", "angry_mean")),
+         name = "table_D8_2.tex")

32/replication_package/SurveyAnalysis_Main.R

@@ -0,0 +1,378 @@
+# Replication File for Survey Analysis  
+# Figure 2: Individuals Living in Municipalities with a Higher Degree ofExcess MalesPerceiveMore Mate Competition
+# Figure 3: List Experiment
+# Table 1: Mate Competition Predicts Support for Hate Crime
+# Figure 4: Estimated Effects of Mate Competition on Support for Hate Crimes
+
+# R version 4.0.2 (2020-06-22)
+
+# ##################
+# Data Preparation  
+# ##################
+rm(list=ls())
+# Set the directly appropriately
+
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+# install.packages("pBrackets") # pBrackets_1.0  
+# install.packages("stargazer") # stargazer_5.2.2
+
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+require(pBrackets) # pBrackets_1.0  
+require(stargazer) # stargazer_5.2.2
+source("Help.R")
+
+dat <- read.dta13(file =  "survey.dta")
+
+# Subset to people in the wave 4
+dat_use <- dat[dat$wave == 4, ]
+
+# #######################
+# Figure 2
+# #######################
+# Prepare Two data sets
+dat_male <- dat_use[dat_use$gender == "Male" & dat_use$age <= 44 & dat_use$age >= 18, ]
+dat_male_y <- dat_use[dat_use$gender == "Male" & dat_use$age <= 40 & dat_use$age >= 30, ]
+
+# Overall Samples 
+dat_use$MateComp.cont_bin <- ifelse(dat_use$MateComp.cont >= 3, 1, 0)
+dat_use$excess_c <- ifelse(dat_use$pop_15_44_muni_gendergap_2015 < 1.04, "1", 
+                           ifelse(dat_use$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+mean_all <- tapply(dat_use$MateComp.cont_bin, dat_use$excess_c, mean)
+se_all <- tapply(dat_use$MateComp.cont_bin, dat_use$excess_c, sd)/sqrt(table(dat_use$excess_c))
+
+# Male (18 - 44)
+dat_male$MateComp.cont_bin <- ifelse(dat_male$MateComp.cont >= 3, 1, 0)
+dat_male$excess_c <- ifelse(dat_male$pop_15_44_muni_gendergap_2015 < 1.04, "1", 
+                            ifelse(dat_male$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+mean_all_m <- tapply(dat_male$MateComp.cont_bin, dat_male$excess_c, mean)
+se_all_m <- tapply(dat_male$MateComp.cont_bin, dat_male$excess_c, sd)/sqrt(table(dat_male$excess_c))
+
+# Male (30 - 40)
+dat_male_y$MateComp.cont_bin <- ifelse(dat_male_y$MateComp.cont >= 3, 1, 0)
+dat_male_y$excess_c <- ifelse(dat_male_y$pop_15_44_muni_gendergap_2015 < 1.04, "1", 
+                              ifelse(dat_male_y$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+mean_all_y <- tapply(dat_male_y$MateComp.cont_bin, dat_male_y$excess_c, mean)
+se_all_y <- tapply(dat_male_y$MateComp.cont_bin, dat_male_y$excess_c, sd)/sqrt(table(dat_male_y$excess_c))
+
+
+mean_all ## 0.1835004 0.1970803 0.2244489 
+mean_all_m ## 0.2282609 0.2745902 0.3750000 
+mean_all_y ## 0.1743119 0.2818182 0.4705882 
+
+{
+  diff <- c(mean_all[2] - mean_all[1], 
+            mean_all[3] - mean_all[2],  
+            mean_all[3] - mean_all[1])
+  sd_d <- c(sqrt(se_all[2]^2 + se_all[1]^2), 
+            sqrt(se_all[3]^2 + se_all[2]^2),  
+            sqrt(se_all[3]^2 + se_all[1]^2))
+  diff_m <- c(mean_all_m[2] - mean_all_m[1], 
+              mean_all_m[3] - mean_all_m[2],  
+              mean_all_m[3] - mean_all_m[1])
+  sd_d_m <- c(sqrt(se_all_m[2]^2 + se_all_m[1]^2), 
+              sqrt(se_all_m[3]^2 + se_all_m[2]^2),  
+              sqrt(se_all_m[3]^2 + se_all_m[1]^2))
+  diff_y <- c(mean_all_y[2] - mean_all_y[1], 
+              mean_all_y[3] - mean_all_y[2],  
+              mean_all_y[3] - mean_all_y[1])
+  sd_d_y <- c(sqrt(se_all_y[2]^2 + se_all_y[1]^2), 
+              sqrt(se_all_y[3]^2 + se_all_y[2]^2),  
+              sqrt(se_all_y[3]^2 + se_all_y[1]^2))
+  
+  
+  diff_l <- c(diff, diff_m, diff_y)
+  se_l <- c(sd_d, sd_d_m, sd_d_y)
+  p_value <- 2*(1 - pnorm(abs(diff_l/se_l)))
+  diff_table <- cbind(diff_l, se_l, p_value)
+}
+
+pdf("figure_2.pdf", height= 15.5, width = 6.5)
+par(mfrow = c(3, 1), mar = c(6,5,5,2), oma = c(0,4,0,0))
+plot(seq(1:3), mean_all, pch = 19, ylim = c(0.1,0.4), 
+     xlim = c(0.5, 3.5),
+     main = "All", xaxt = "n", xlab = "", ylab = "", 
+     cex.axis = 2.25, cex.main = 2.5, yaxt  = "n",
+     cex = 2.25, cex.lab = 2.5)
+segments(seq(1:3), mean_all - 1.96*se_all, 
+         seq(1:3), mean_all + 1.96*se_all, pch = 19, lwd = 3)
+Axis(side = 1, at = c(1,2,3), labels = c("1st tercile", "2nd tercile", "3rd tercile"), cex.axis = 2.25)
+Axis(side = 2, at = c(0.1,0.2,0.3, 0.4), labels = c("0.1", "0.2", "0.3", "0.4"), cex.axis = 2.25)
+brackets(x1 = 1.1, y1  = 0.3, x2 = 1.9, y2 =  0.3, h = 0.01, type = 4)
+brackets(x1 = 2.1, y1  = 0.3, x2 = 2.9, y2 =  0.3, h = 0.01, type =  4)
+brackets(x1 = 1, y1  = 0.37, x2 = 3, y2 =  0.37, h = 0.01, type =  4)
+# text(x = 1.5, y = 0.33, paste0("pv = ", round(p_value[1],digits=3)), cex = 1.95)
+text(x = 1.5, y = 0.33, paste0("pv = 0.40"), cex = 1.95)
+text(x = 2.5, y = 0.33, paste0("pv = ", round(p_value[2],2)), cex = 1.95)
+text(x = 2, y = 0.40, paste0("pv = ", round(p_value[3],2)), cex = 1.95)
+mtext("Excess Males", side = 1, cex = 1.75, line = 3.75)
+mtext("Proportion Perceiving\nMate Competition", side = 2, cex = 1.75, line = 3.75)
+
+plot(seq(1:3), mean_all_m, pch = 19, ylim = c(0.1,0.6),
+     xlim = c(0.5, 3.5),
+     main = "Male (18-44)", xaxt = "n", xlab = "", ylab = "", cex.axis = 2.25, cex.main = 2.5,
+     cex = 2.25, cex.lab = 2.5)
+segments(seq(1:3), mean_all_m - 1.96*se_all_m, 
+         seq(1:3), mean_all_m + 1.96*se_all_m, pch = 19, lwd = 3)
+Axis(side = 1, at = c(1,2,3), labels = c("1st tercile", "2nd tercile", "3rd tercile"), cex.axis = 2.25)
+brackets(x1 = 1.1, y1  = 0.48, x2 = 1.9, y2 =  0.48, h = 0.01, type = 4)
+brackets(x1 = 2.1, y1  = 0.48, x2 = 2.9, y2 =  0.48, h = 0.01, type =  4)
+brackets(x1 = 1, y1  = 0.53, x2 = 3, y2 =  0.53, h = 0.03, type =  4)
+text(x = 1.5, y = 0.51, paste0("pv = ", round(p_value[4],2)), cex = 1.95)
+text(x = 2.5, y = 0.51, paste0("pv = ", round(p_value[5],2)), cex = 1.95)
+# text(x = 2, y = 0.58, paste0("pv = ", round(p_value[6],2)), cex = 1.95)
+text(x = 2, y = 0.58, paste0("pv = 0.00"), cex = 1.95)
+mtext("Excess Males", side = 1, cex = 1.75, line = 3.75)
+mtext("Proportion Perceiving\nMate Competition", side = 2, cex = 1.75, line = 3.75)
+
+
+plot(seq(1:3), mean_all_y, pch = 19, ylim = c(0.1,0.75),
+     xlim = c(0.5, 3.5),
+     main = "Male (30 - 40)", xaxt = "n", xlab = "", ylab = "", cex.axis = 2.25, cex.main = 2.5,
+     cex = 2.25, cex.lab = 2.5)
+segments(seq(1:3), mean_all_y - 1.96*se_all_y, 
+         seq(1:3), mean_all_y + 1.96*se_all_y, pch = 19, lwd = 3)
+Axis(side = 1, at = c(1,2,3), labels = c("1st tercile", "2nd tercile", "3rd tercile"), cex.axis = 2.25)
+brackets(x1 = 1.1, y1  = 0.62, x2 = 1.9, y2 =  0.62, h = 0.03, type = 4)
+brackets(x1 = 2.1, y1  = 0.62, x2 = 2.9, y2 =  0.62, h = 0.03, type =  4)
+brackets(x1 = 1, y1  = 0.7, x2 = 3, y2 =  0.7, h = 0.03, type =  4)
+text(x = 1.5, y = 0.67, paste0("pv = ", round(p_value[7],2)),cex = 1.95)
+text(x = 2.5, y = 0.67, paste0("pv = ", round(p_value[8],2)), cex = 1.95)
+text(x = 2, y = 0.75, paste0("pv = 0.00"), cex = 1.95)
+# text(x = 2, y = 0.75, paste0("pv = ", round(p_value[9],3)))
+mtext("Excess Males", side = 1, cex = 1.75, line = 3.75)
+mtext("Proportion Perceiving\nMate Competition", side = 2, cex = 1.75, line = 3.75)
+
+dev.off()
+
+# ############################
+# Main Models (Table 1)
+# ############################
+lm1 <- lm(hate_violence_means ~ MateComp.cont, data=dat_use)
+summary(lm1)
+
+lm2 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont, data=dat_use)
+summary(lm2)
+
+lm3 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont + 
+            factor(age_group) +     # age group
+            factor(gender) +     # gender 
+            factor(state) +     # state  
+            factor(citizenship) +    # german citizen
+            factor(marital) +    # marital status
+            factor(religion) +    # religious affiliation
+            eduyrs +    # education
+            factor(occupation) +    # main activity
+            factor(income) +   # income
+            factor(household_size) +   # household size
+            factor(self_econ),    # subjective social status
+          data=dat_use)
+summary(lm3)
+
+lm4 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont + 
+            factor(age_group) +   # age group
+            factor(gender) +   # gender 
+            factor(state) +   # state  
+            factor(citizenship) +  # german citizen
+            factor(marital) +  # marital status
+            factor(religion) +  # religious affiliation
+            eduyrs +  # education
+            factor(occupation) +  # main activity
+            factor(income) + # income
+            factor(household_size) + # household size
+            factor(self_econ) + # subjective social status
+            factor(ref_integrating) + # Refugee Index (National-level; Q73) 8 in total
+            factor(ref_citizenship) + factor(ref_reduce) + factor(ref_moredone) + factor(ref_cultgiveup) + 
+            factor(ref_economy) + factor(ref_crime) + factor(ref_terror),
+          data=dat_use)
+summary(lm4)
+
+lm5 <- lm(hate_violence_means ~ MateComp.cont + JobComp.cont + LifeSatis.cont + 
+            factor(age_group) +      # age group
+            factor(gender) +      # gender 
+            factor(state) +      # state  
+            factor(citizenship) +     # german citizen
+            factor(marital) +     # marital status
+            factor(religion) +     # religious affiliation
+            eduyrs + # education
+            factor(occupation) +     # main activity
+            factor(income) +    # income
+            factor(household_size) +    # household size
+            factor(self_econ) +    # subjective social status
+            factor(ref_integrating) + # Refugee Index (National-level; Q73) 8 in total
+            factor(ref_citizenship) + factor(ref_reduce) + factor(ref_moredone) + factor(ref_cultgiveup) + 
+            factor(ref_economy) + factor(ref_crime) + factor(ref_terror) + 
+            factor(ref_loc_services) +    # Refugee Index (Local, Q75)
+            factor(ref_loc_economy) + factor(ref_loc_crime) + factor(ref_loc_culture) + factor(ref_loc_islam) + 
+            factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife), ## end
+          data=dat_use)
+summary(lm5)
+
+
+# Add More Variables 
+# lrscale  Q21  Left-Right Scale
+# afd, Q23  Closeness to AfD
+# muslim_ind, afd_ind, contact_ind
+# distance_ref Q71. Distance to refugee reception centers
+# settle_ref Q72. Settlement of refugees living in area
+
+formula.5 <- 
+  as.character("hate_violence_means ~ MateComp.cont + JobComp.cont + 
+               LifeSatis.cont +  factor(age_group) + factor(gender) + 
+               factor(state) + factor(citizenship) + factor(marital) + 
+               factor(religion) + eduyrs + factor(occupation) + 
+               factor(income) + factor(household_size) + factor(self_econ) + 
+               factor(ref_integrating) + factor(ref_citizenship) + factor(ref_reduce) + 
+               factor(ref_moredone) + factor(ref_cultgiveup) + 
+               factor(ref_economy) + factor(ref_crime) + factor(ref_terror)  + 
+               factor(ref_loc_services) +  factor(ref_loc_economy) + factor(ref_loc_crime) + 
+               factor(ref_loc_culture) + factor(ref_loc_islam) + 
+               factor(ref_loc_schools) + factor(ref_loc_housing) + factor(ref_loc_wayoflife)")
+
+formula.6 <- paste(formula.5, "factor(distance_ref) + factor(settle_ref)", 
+                   "lrscale + afd + muslim_ind + afd_ind + contact_ind", 
+                   sep="+", collapse="+") 
+
+lm6 <- lm(as.formula(formula.6), data=dat_use)
+summary(lm6)
+
+lm.list.table1 <- list(lm1, lm2, lm3, lm4, lm5, lm6)
+
+# Table 1
+star_out(stargazer(lm.list.table1,
+                   covariate.labels = c("Mate Competition","Job Competition","Life Satisfaction"),
+                   keep=c("MateComp.cont","JobComp.cont","LifeSatis.cont"),
+                   star.char = c("\\dagger", "*", "**"),
+                   notes = c("$^{\\dagger}$ p$<$0.1; $^{*}$ p$<$0.05; $^{**}$ p$<$0.01"), notes.append = FALSE), 
+         name = "table1.tex")
+
+## #################
+## Figure 4
+## #################
+# with Difference Outcomes 
+# hate_pol_message (v_320): "82. Support for Hate Crime_Attacks against refugee homes are somet" 
+# hate_prevent_settlement (v_319): "82. Support for Hate Crime_Racist violence is defensible if it lea"
+# hate_polcondemn (v_316): "82. Support for Hate Crime_Politicians should condemn attacks agai" 
+# hate_justified (v_315): "82. Support for Hate Crime_Hostility against foreigners is sometimes justified" 
+
+formula.7.means   <- paste("hate_violence_means ~ ", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.message <- paste("hate_pol_message ~", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.prevent <- paste("hate_prevent_settlement ~", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.condemn <- paste("hate_polcondemn ~ ", as.character(as.formula(formula.6))[3], sep = "")
+formula.7.justified <- paste("hate_justified ~ ", as.character(as.formula(formula.6))[3], sep = "")
+
+# output
+lm7.means <- lm(as.formula(formula.7.means), data=dat_use)
+lm7.justified <- lm(as.formula(formula.7.justified), data=dat_use)
+lm7.message <- lm(as.formula(formula.7.message), data=dat_use)
+lm7.prevent <- lm(as.formula(formula.7.prevent), data=dat_use)
+lm7.condemn <- lm(as.formula(formula.7.condemn), data=dat_use)
+
+# Figure 5
+point <- c(coef(lm7.means)["MateComp.cont"],
+           coef(lm7.justified)["MateComp.cont"], coef(lm7.message)["MateComp.cont"], 
+           coef(lm7.prevent)["MateComp.cont"], coef(lm7.condemn)["MateComp.cont"])
+
+se <- c(summary(lm7.means)$coef["MateComp.cont", 2],
+        summary(lm7.justified)$coef["MateComp.cont", 2], summary(lm7.message)$coef["MateComp.cont", 2],
+        summary(lm7.prevent)$coef["MateComp.cont", 2], summary(lm7.condemn)$coef["MateComp.cont", 2])
+
+
+pdf("figure_4.pdf", height = 4, width = 8)
+par(mar = c(2,4,4,1))
+plot(seq(1:5), point, pch = 19, ylim = c(-0.05, 0.25), xlim = c(0.5, 5.5), 
+     xlab = "", xaxt = "n", ylab = "Estimated Effects", 
+     main = "Estimated Effects of Mate Competition", cex.lab = 1.25, cex.axis = 1.25, cex.main = 1.5)
+segments(seq(1:5), point - 1.96*se, 
+         seq(1:5), point + 1.96*se, lwd = 2)
+Axis(side=1, at = seq(1:5), labels = c("Only Means", "Justified", "Message", 
+                                       "Prevent", "Condemn"), cex.axis = 1.25)
+abline(h =0, lty = 2)
+dev.off()
+
+## Table C.5 (in Appendix C.5)
+# lm.list_d <- list(lm7.means, lm7.justified, lm7.message, lm7.prevent, lm7.condemn)
+# stargazer(lm.list_d,
+#           covariate.labels = c("Mate Competition","Job Competition","Life Satisfaction"),
+#           keep=c("MateComp.cont","JobComp.cont","LifeSatis.cont"))
+
+
+## #############################
+## Figure 3: List Experiment
+## #############################
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+# install.packages("list") # list_9.2 
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+require(list) # list_9.2 
+
+dat <- read.dta13(file =  "survey.dta")
+data.u2 <- dat[dat$wave == 2, ]
+
+# Means: When it comes to the refugee problem, violence is sometimes the only means that citizens have to get the attention of German politicians
+data.list.u2   <- data.u2[data.u2$list == "1",]
+data.direct.u2 <- data.u2[data.u2$list == "2",]
+data.list.u2 <- data.list.u2[is.na(data.list.u2$treatment_list)==FALSE,]
+data.list.u2$List.treat <- ifelse(data.list.u2$treatment_list == "Scenario 2", 1, 0)
+
+## Difference-in-Means
+## with Mean = 0.15401 sd = 0.03358
+diff.in.means.results2 <- ictreg(outcome_list ~ 1, data = data.list.u2,
+                                 treat = "List.treat", J=3, method = "lm")
+summary(diff.in.means.results2)
+
+## Compare to All People who answered Direct Question (n = 2170)
+data.u2.all.direct <- data.u2[is.na(data.u2$hate_violence_means) == FALSE, ]
+data.u2.all.direct$hate.direct.bin <- ifelse(data.u2.all.direct$hate_violence_means >=3, 1, 0)
+point_dir2 <- mean(data.u2.all.direct$hate.direct.bin) ## 0.181 
+se_dir2 <- sd(data.u2.all.direct$hate.direct.bin)/sqrt(length(data.u2.all.direct$hate.direct.bin)) # 0.0083
+
+# Compare Questions within Wave 2 
+# Direct Questions
+data.u2$message.bin <- ifelse(data.u2$hate_pol_message >= 3, 1, 0)
+data.u2$condemn.bin <- ifelse(data.u2$hate_polcondemn >= 3, 1, 0)
+data.u2$justified.bin <- ifelse(data.u2$hate_justified >= 3, 1, 0)
+
+message.mean2 <- mean(data.u2$message.bin)
+condemn.mean2 <- mean(data.u2$condemn.bin)
+justified.mean2 <- mean(data.u2$justified.bin)
+message.se2 <- sd(data.u2$message.bin)/sqrt(length(data.u2$message.bin)) # 0.0070
+condemn.se2 <- sd(data.u2$condemn.bin)/sqrt(length(data.u2$condemn.bin)) # 0.0079
+justified.se2 <- sd(data.u2$justified.bin)/sqrt(length(data.u2$justified.bin)) # 0.0074
+
+# plot
+point <- c(summary(diff.in.means.results2)$par.treat, point_dir2, justified.mean2, message.mean2, condemn.mean2)
+se_p  <- c(summary(diff.in.means.results2)$se.treat,  se_dir2, justified.se2, message.se2, condemn.se2)
+base <- barplot(point, ylim = c(0, 0.20))
+bar_name_u <- c("Only Means\n(List)","Only Means\n(Direct)", "Justified", "Message", "Condemn")
+bar_name <- rep("",5)
+
+pdf("figure_3.pdf", height = 4.5, width = 8)
+par(mar = c(4, 5, 2, 1))
+barplot(point, ylim = c(0, 0.3), names.arg = bar_name, 
+        col = c(adjustcolor("red", 0.4), "gray", "gray", "gray", "gray"), cex.axis = 1.3,
+        ylab = "Proportion of respondents", cex.lab = 1.45)
+arrows(base[,1], point - 1.96*se_p, base[,1], point + 1.96*se_p, 
+       lwd = 3, angle = 90, length = 0.05, code = 3,
+       col = c("red", "black", "black", "black", "black"))
+mtext(bar_name_u[1], outer = FALSE, side = 1, at = base[1], cex = 1.2, line = 2.4)
+mtext(bar_name_u[2], outer = FALSE, side = 1, at = base[2], cex = 1.2, line = 2.4)
+mtext(bar_name_u[3], outer = FALSE, side = 1, at = base[3], cex = 1.2, line = 2.4)
+mtext(bar_name_u[4], outer = FALSE, side = 1, at = base[4], cex = 1.2, line = 2.4)
+mtext(bar_name_u[5], outer = FALSE, side = 1, at = base[5], cex = 1.2, line = 2.4)
+text(x = base[1], y = 0.275, "Estimate from \nList Experiment", col = "red", font = 2)
+text(x = (base[3] + base[4])/2, y = 0.275, "Direct Questions", font = 2)
+dev.off()

32/replication_package/YouGov.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68bc13dffdedff3ad30211644a050d0e85e24693ecaf6b2e8c95f6e37cfa9f5d
+size 212090

32/replication_package/codebook_YouGov.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:97b900343580543bff2743c8a158d93e1d6a9aedc0c8215911fe84c02c27c2d1
+size 139433

32/replication_package/codebook_context.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ca51e10c3186ba504bf83bcea63385813cfb51ce1e354d47832e9cbb916d60a
+size 140823

32/replication_package/codebook_context_placebo.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e04f05c86b1ad29b2e004537221a1ec4d610285d84a539227473f84dd938e7d9
+size 164836

32/replication_package/codebook_survey.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99bb13e7d6936c4f57eaba442ae1d453db5368b1984dd8ec4e47023d732ae542
+size 146756

32/replication_package/context.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe201f5a75c7d014e63a155ddbf44a8a144e0b93c366d1d58c7898bf411c6009
+size 7408421

32/replication_package/context_placebo.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c300be8df5d3bf094c2cfe4fa5e7c70d55dd9902cd3369d789b434704a7abfc0
+size 4064170

32/replication_package/master.R

@@ -0,0 +1,10 @@
+# Main Analysis in the paper
+source("ContextAnalysis_Main.R")
+source("SurveyAnalysis_Main.R")
+
+# Numbers we report in the paper 
+source("number_in_texts.R")
+
+# Results in Appendix
+source("ContextAnalysis_Appendix.R")
+source("SurveyAnalysis_Appendix.R")

32/replication_package/merge_context.R

@@ -0,0 +1,217 @@
+# Internal File for Merging datasets for producing "context.dta"
+# R version 4.0.2 (2020-06-22)
+
+rm(list=ls())
+# install.packages("readstata13") # readstata13_0.9.2
+
+require(readstata13)  #  readstata13_0.9.2
+
+setwd("source_data")
+
+# 0. Base data that contains AGS identifies and Year 
+base <- read.dta13("base.dta")
+
+# Note: 
+# Every data source we use below is fully described in "source_context.pdf" 
+
+# 1. Hate Crime Data 
+hate <- read.dta13("hate.dta")
+hate$Housing_all_muni <- hate$Arson_muni + hate$Other_muni
+hate$Hate_all_muni <- hate$Housing_all_muni + hate$Physical_muni
+
+context0 <- merge(base, hate[, c("ags_muni", "year", 
+                                 "Hate_all_muni", 
+                                 "Physical_muni")], by = c("ags_muni", "year"), all.x = TRUE)
+
+# 2. Population Data 
+pop_dat <- read.dta13("pop_gemeinde_2008_2018.dta")
+colnames(pop_dat)[colnames(pop_dat) == "ags"] <- "ags_muni"
+
+pop_dat$pop_m_25_44 <- pop_dat$pop_m_25_29 + pop_dat$pop_m_30_34 +  pop_dat$pop_m_35_39 + pop_dat$pop_m_40_44
+pop_dat$pop_f_25_44 <- pop_dat$pop_f_25_29 + pop_dat$pop_f_30_34 +  pop_dat$pop_f_35_39 + pop_dat$pop_f_40_44
+
+pop_dat$pop_m_15_44 <- pop_dat$pop_m_15_17 +  pop_dat$pop_m_18_19 +  pop_dat$pop_m_20_24 + pop_dat$pop_m_25_44
+pop_dat$pop_f_15_44 <- pop_dat$pop_f_15_17 +  pop_dat$pop_f_18_19 +  pop_dat$pop_f_20_24 + pop_dat$pop_f_25_44
+
+pop_dat$population_muni <- pop_dat$pop_mf_total
+
+pop_dat$pop_25_44_muni_gendergap <- pop_dat$pop_m_25_44/pop_dat$pop_f_25_44
+pop_dat$pop_15_44_muni_gendergap <- pop_dat$pop_m_15_44/pop_dat$pop_f_15_44
+pop_dat$pop_25_44_muni_gendergap[is.infinite(pop_dat$pop_25_44_muni_gendergap)] <- NA
+pop_dat$pop_15_44_muni_gendergap[is.infinite(pop_dat$pop_15_44_muni_gendergap)] <- NA
+
+pop_dat_2015 <- subset(pop_dat, year ==  2015)
+pop_dat_2015$pop_25_44_muni_gendergap_2015 <- pop_dat_2015$pop_m_25_44/pop_dat_2015$pop_f_25_44
+pop_dat_2015$pop_15_44_muni_gendergap_2015 <- pop_dat_2015$pop_m_15_44/pop_dat_2015$pop_f_15_44
+pop_dat_2015$pop_25_44_muni_gendergap_2015[is.infinite(pop_dat_2015$pop_25_44_muni_gendergap_2015)] <- NA
+pop_dat_2015$pop_15_44_muni_gendergap_2015[is.infinite(pop_dat_2015$pop_15_44_muni_gendergap_2015)] <- NA
+
+pop_dat_2015$population_muni_2015 <- pop_dat_2015$pop_mf_total
+
+
+#  3. area  
+area <- read.dta13("area_mun.dta")
+area_use <- area[area$ags %in% context0$ags_muni, ]
+colnames(area_use)[colnames(area_use) == "ags"] <- "ags_muni"
+
+pop_dat_2015 <- merge(pop_dat_2015, area_use[,c("ags_muni", "area_sqk")], all.x =  TRUE)
+pop_dat_2015$popdens_muni_2015 <- pop_dat_2015$population_muni_2015/pop_dat_2015$area_sqk
+
+pop_dat <- merge(pop_dat, area_use[,c("ags_muni", "area_sqk")], all.x =  TRUE)
+pop_dat$popdens_muni <- pop_dat$population_muni/pop_dat$area_sqk
+
+context0 <- merge(context0, pop_dat_2015[, c("ags_muni", 
+                                             "pop_25_44_muni_gendergap_2015",
+                                             "pop_15_44_muni_gendergap_2015",
+                                             "population_muni_2015",
+                                             "popdens_muni_2015")], by = c("ags_muni"), all.x = TRUE)
+
+# 4. Unemployment
+pop_dat <- read.dta13("pop_gemeinde_2008_2018.dta")
+colnames(pop_dat)[colnames(pop_dat) == "ags"] <- "ags_muni"
+pop_dat$pop_m_25_44 <- pop_dat$pop_m_25_29 + pop_dat$pop_m_30_34 +  pop_dat$pop_m_35_39 + pop_dat$pop_m_40_44
+pop_dat$pop_f_25_44 <- pop_dat$pop_f_25_29 + pop_dat$pop_f_30_34 +  pop_dat$pop_f_35_39 + pop_dat$pop_f_40_44
+pop_dat$pop_m_15_44 <- pop_dat$pop_m_15_17 +  pop_dat$pop_m_18_19 +  pop_dat$pop_m_20_24 + pop_dat$pop_m_25_44
+pop_dat$pop_f_15_44 <- pop_dat$pop_f_15_17 +  pop_dat$pop_f_18_19 +  pop_dat$pop_f_20_24 + pop_dat$pop_f_25_44
+unemp_dat <-  read.dta13("unempl_gemeinde_2008_2017.dta")
+colnames(unemp_dat)[colnames(unemp_dat) == "ags"] <- "ags_muni"
+colnames(unemp_dat)[colnames(unemp_dat) == "ags_dist"] <- "ags_county"
+
+## unemployed as share of working age population (age 15-64)
+pop_dat$pop_mf_15_64 <- pop_dat$pop_mf_15_17 +  pop_dat$pop_mf_18_19 +  pop_dat$pop_mf_20_24 + 
+  pop_dat$pop_mf_25_29 + pop_dat$pop_mf_30_34 +  pop_dat$pop_mf_35_39 + pop_dat$pop_mf_40_44 + 
+  pop_dat$pop_mf_45_49 + pop_dat$pop_mf_50_54 + pop_dat$pop_mf_55_59 + pop_dat$pop_mf_60_64
+
+pop_dat$pop_m_15_64 <- pop_dat$pop_m_15_44 + pop_dat$pop_m_45_49 + pop_dat$pop_m_50_54 + 
+  pop_dat$pop_m_55_59 + pop_dat$pop_m_60_64
+
+pop_dat$pop_f_15_64 <- pop_dat$pop_f_15_44 + pop_dat$pop_f_45_49 + pop_dat$pop_f_50_54 + 
+  pop_dat$pop_f_55_59 + pop_dat$pop_f_60_64
+
+unemp_dat_use <- unemp_dat[, c("ags_muni", "ags_county",
+                               "year", 
+                               "unempl_all_total", 
+                               "unempl_all_male_total", 
+                               "unempl_all_fem_total")]
+pop_dat_m <- pop_dat[pop_dat$year >= 2011, c("ags_muni", "year", "pop_mf_15_64", "pop_m_15_64", "pop_f_15_64")]
+unemp_merge  <- merge(pop_dat_m, unemp_dat_use, by = c("ags_muni", "year"), all.x = TRUE, all.y = FALSE)
+unemp_merge$unemp_all_muni  <- (unemp_merge$unempl_all_total/unemp_merge$pop_mf_15_64)*100
+
+unemp_2015 <- unemp_merge[unemp_merge$year == 2015, ]
+unemp_2015$unemp_all_muni_2015 <- unemp_2015$unemp_all_muni
+unemp_2015$log_unemp_all_muni_2015  <- log(unemp_2015$unemp_all_muni_2015 + 1)
+
+context0 <- merge(context0, unemp_2015[, c("ags_muni", 
+                                           "log_unemp_all_muni_2015")], by = c("ags_muni"), 
+                  all.x = TRUE, all.y = FALSE)
+
+# 5. Unemployment Gender Gap
+d2 <- read.dta13("merged_context_2.dta")  # we created this data set in "produce_context_data.do"
+d2_15 <- d2[d2$year == 2015, ]
+d2_15$unemp_gendergap_2015 <- round(d2_15$unemp_gendergap, 6)
+
+context0 <- merge(context0, d2_15[, c("ags_county", "unemp_gendergap_2015")], 
+                  all.x = TRUE, all.y = FALSE, by = "ags_county")
+
+# 6. Population Change
+pop_dat <- read.dta13("pop_gemeinde_2008_2018.dta")
+colnames(pop_dat)[colnames(pop_dat) == "ags"] <- "ags_muni"
+pop_dat$population_muni <- pop_dat$pop_mf_total
+pop_dat_2015 <- subset(pop_dat, year ==  2015)
+pop_dat_2011 <- subset(pop_dat, year ==  2011)
+pop_dat_2015$d_pop1511_muni <- 
+  (pop_dat_2015$population_muni - pop_dat_2011$population_muni)/pop_dat_2011$population_muni
+
+context0 <- merge(context0, pop_dat_2015[, c("ags_muni", 
+                                             "d_pop1511_muni")], by = c("ags_muni"), all.x = TRUE)
+
+# 7. Voting
+voting <- read.dta13("voting.dta")
+
+context0 <- merge(context0, voting[, c("ags_muni", 
+                                       "vote_afd_2013_muni")], by = c("ags_muni"), all.x = TRUE)
+
+# 8. Refugee Data
+ref_dat <- read.dta13("refugees_2008_2017.dta")
+ref_2014 <- subset(ref_dat, year == 2014)
+ref_2015 <- subset(ref_dat, year == 2015)
+table(ref_2014$ags_county  == ref_2015$ags_county)
+
+ref_2014$ref_inflow_1514 <- ref_2015$pop_ref - ref_2014$pop_ref
+ref_2014$log_ref_inflow_1514 <- log(1500 + ref_2014$ref_inflow_1514)
+
+ref_2014$pop_ref_2014 <- ref_2014$pop_ref
+ref_2014$pop_ref_2015 <- ref_2015$pop_ref
+
+# Proportion of male refugees
+ref_prop <- read.dta13("merged_context_2.dta") # we created this data set in "produce_context_data.do"
+
+context0 <- merge(context0, ref_2014[, c("ags_county", 
+                                         "log_ref_inflow_1514",
+                                         "pop_ref_2014")], by = c("ags_county"), all.x = TRUE)
+context0 <- merge(context0, ref_prop[, c("year", "ags_county", 
+                                         "pc_ref_male")], by = c("year", "ags_county"), all.x = TRUE)
+
+#  9. Violence
+crime <- read.dta13("crime.dta")
+crime <- crime[crime$year == 2015, ]
+
+pop_county <- read.dta13("pop_kreise_2015_2017.dta")
+pop_county1 <- subset(pop_county, year == 2015)
+
+crime2 <- merge(crime[, c("ags_county", "violence_num_cases")],
+                pop_county1[, c("ags_county", "population")],
+                by = "ags_county", all.x = TRUE, all.y = FALSE)
+crime2$violence_percap_2015 <- crime2$violence_num_cases/crime2$population
+
+context0 <- merge(context0, crime2[, c("ags_county", 
+                                       "violence_percap_2015")], by = c("ags_county"), all.x = TRUE)
+
+# 10. Education
+edu <- read.dta13("merged_context_2.dta") # we created this data set in "produce_context_data.do"
+edu  <- edu[edu$year ==  2011, ]
+edu <- edu[, c("ags_county", "pc_hidegree_all2011")]
+
+context0 <- merge(context0, edu[, c("ags_county", 
+                                    "pc_hidegree_all2011")], by = c("ags_county"), all.x = TRUE)
+
+
+#  10. Industry 
+manu0 <- read.dta13("merged_context_2.dta") # we created this data set in "produce_context_data.do"
+manu0 <- manu0[, c("year", "ags_county", "pc_manufacturing")]
+manu <- manu0[manu0$year >= 2011 & manu0$year <= 2015, ]
+rownames(manu) <- NULL
+manu_orig <- manu
+
+manu  <- manu_orig[manu_orig$year ==  2015, ]
+manu <- manu[, c("ags_county", "pc_manufacturing")]
+manu <- manu[is.element(manu$ags_county, unique(context0$ags_county)),]
+manu$pc_manufacturing_2015 <- manu$pc_manufacturing
+
+manu2011  <- manu_orig[manu_orig$year ==  2011, ]
+manu2011 <- manu2011[, c("ags_county", "pc_manufacturing")]
+manu2011 <- manu2011[is.element(manu2011$ags_county, unique(context0$ags_county)),]
+manu2011$pc_manufacturing_2011 <- manu2011$pc_manufacturing
+
+# d_manuf1115
+manu$d_manuf1115 <- manu$pc_manufacturing_2015 - manu2011$pc_manufacturing_2011
+
+context0 <- merge(context0, manu[, c("ags_county", 
+                                     "pc_manufacturing_2015",
+                                     "d_manuf1115")], by = c("ags_county"), all.x = TRUE)
+
+# 11. East
+context0$east <- 0
+context0$east[context0$ags_state %in% c("11","12","13","14","15","16")]  <-  1
+
+# 12. Create additional variables 
+context0$Hate_all_muni_bin <- as.numeric(context0$Hate_all_muni >  0)
+context0$Physical_muni_bin <- as.numeric(context0$Physical_muni >  0)
+
+context0$log_population_muni_2015 <- log(context0$population_muni_2015)
+context0$log_popdens_muni_2015 <- log(context0$popdens_muni_2015)
+context0$log_pop_ref_2014 <- log(context0$pop_ref_2014)
+context0$log_violence_percap_2015 <- log(context0$violence_percap_2015)
+
+context0 <- context0[order(context0$year, context0$ags_muni), ]
+save.dta13(context0, file = "context.dta")

32/replication_package/merge_context_placebo.R

@@ -0,0 +1,231 @@
+rm(list = ls())
+# install.packages("readstata13") # readstata13_0.9.2
+# install.packages("plm") # plm_2.2-3  
+
+library(readstata13) # readstata13_0.9.2
+library(plm) # plm_2.2-3  
+
+setwd("source_data")
+
+# 0. Base data that contains AGS identifies and Year 
+base_pl <- read.dta13("base_pl.dta")
+
+# Note: 
+# Every data source we use below is fully described in "source_context.pdf" 
+
+# 1. Hate Crime Data
+hate <- read.dta13("hate.dta")
+hate$Housing_all_muni <- hate$Arson_muni + hate$Other_muni
+hate$Hate_all_muni <- hate$Housing_all_muni + hate$Physical_muni
+
+
+context0 <- merge(base_pl, hate[, c("ags_muni", "year", 
+                                    "Hate_all_muni")], by = c("ags_muni", "year"), all.x = TRUE)
+
+# 2. Population Data 
+pop_dat <- read.dta13("pop_gemeinde_2008_2018.dta")
+pop_dat <- pop_dat[pop_dat$year >= 2011, ]
+
+pop_dat$pop_m_25_44 <- pop_dat$pop_m_25_29 + pop_dat$pop_m_30_34 +  pop_dat$pop_m_35_39 + pop_dat$pop_m_40_44
+pop_dat$pop_f_25_44 <- pop_dat$pop_f_25_29 + pop_dat$pop_f_30_34 +  pop_dat$pop_f_35_39 + pop_dat$pop_f_40_44
+
+pop_dat$pop_m_15_44 <- pop_dat$pop_m_15_17 +  pop_dat$pop_m_18_19 +  pop_dat$pop_m_20_24 + pop_dat$pop_m_25_44
+pop_dat$pop_f_15_44 <- pop_dat$pop_f_15_17 +  pop_dat$pop_f_18_19 +  pop_dat$pop_f_20_24 + pop_dat$pop_f_25_44
+
+pop_dat$population_muni_anu <- pop_dat$pop_mf_total
+
+pop_dat$pop_25_44_muni_gendergap_anu <- pop_dat$pop_m_25_44/pop_dat$pop_f_25_44
+pop_dat$pop_15_44_muni_gendergap_anu <- pop_dat$pop_m_15_44/pop_dat$pop_f_15_44
+pop_dat$pop_25_44_muni_gendergap_anu[is.infinite(pop_dat$pop_25_44_muni_gendergap_anu)] <- NA
+pop_dat$pop_15_44_muni_gendergap_anu[is.infinite(pop_dat$pop_15_44_muni_gendergap_anu)] <- NA
+
+pop_dat_pl <-  pop_dat[, c("year", "ags_muni","ags_county", "ags_state",
+                           "population_muni_anu", "pop_15_44_muni_gendergap_anu")]
+
+# creating pop_15_44_muni_gendergap_future
+pop_dat_pl_p <- pdata.frame(pop_dat_pl, index = c("ags_muni", "year"))
+pop_dat_pl_p$pop_15_44_muni_gendergap_future <- as.numeric(lead(pop_dat_pl_p$pop_15_44_muni_gendergap_anu, k = 1))
+class(pop_dat_pl_p) <- "data.frame"
+pop_dat_pl <- pop_dat_pl_p
+rownames(pop_dat_pl) <- NULL
+
+#  3. area  
+area <- read.dta13("area_mun.dta")
+area_use <- area[area$ags %in% base_pl$ags_muni, ]
+colnames(area_use)[colnames(area_use) == "ags"] <- "ags_muni"
+
+pop_dat_pl <- merge(pop_dat_pl, area_use[,c("ags_muni", "area_sqk")], all.x =  TRUE, all.y = FALSE)
+pop_dat_pl$popdens_muni_anu <- pop_dat_pl$population_muni_anu/pop_dat_pl$area_sqk
+
+pop_dat_2015 <- subset(pop_dat_pl, year ==  2015)
+pop_dat_2011 <- subset(pop_dat_pl, year ==  2011)
+pop_dat_2011 <- pop_dat_2011[match(pop_dat_2015$ags_muni, pop_dat_2011$ags_muni),]
+pop_dat_2015$d_pop_muni_anu <- 
+  (pop_dat_2015$population_muni - pop_dat_2011$population_muni)/pop_dat_2011$population_muni
+
+pop_dat_2016 <- subset(pop_dat_pl, year ==  2016)
+pop_dat_2012 <- subset(pop_dat_pl, year ==  2012)
+pop_dat_2012 <- pop_dat_2012[match(pop_dat_2016$ags_muni, pop_dat_2012$ags_muni),]
+pop_dat_2016$d_pop_muni_anu <- 
+  (pop_dat_2016$population_muni - pop_dat_2012$population_muni)/pop_dat_2012$population_muni
+
+pop_dat_2017 <- subset(pop_dat_pl, year ==  2017)
+pop_dat_2013 <- subset(pop_dat_pl, year ==  2013)
+pop_dat_2013 <- pop_dat_2013[match(pop_dat_2017$ags_muni, pop_dat_2013$ags_muni),]
+pop_dat_2017$d_pop_muni_anu <- 
+  (pop_dat_2017$population_muni - pop_dat_2013$population_muni)/pop_dat_2013$population_muni
+
+pop_dat_d <- rbind(pop_dat_2015, pop_dat_2016, pop_dat_2017)
+
+context0 <- merge(context0, pop_dat_d[, c("year", "ags_muni", 
+                                          "pop_15_44_muni_gendergap_anu",
+                                          "pop_15_44_muni_gendergap_future",
+                                          "population_muni_anu",
+                                          "popdens_muni_anu",
+                                          "d_pop_muni_anu")], by = c("year", "ags_muni"), 
+                  all.x = TRUE, all.y = FALSE)
+
+# 4. Unemployment
+pop_dat <- read.dta13("pop_gemeinde_2008_2018.dta")
+colnames(pop_dat)[colnames(pop_dat) == "ags"] <- "ags_muni"
+pop_dat$pop_m_25_44 <- pop_dat$pop_m_25_29 + pop_dat$pop_m_30_34 +  pop_dat$pop_m_35_39 + pop_dat$pop_m_40_44
+pop_dat$pop_f_25_44 <- pop_dat$pop_f_25_29 + pop_dat$pop_f_30_34 +  pop_dat$pop_f_35_39 + pop_dat$pop_f_40_44
+pop_dat$pop_m_15_44 <- pop_dat$pop_m_15_17 +  pop_dat$pop_m_18_19 +  pop_dat$pop_m_20_24 + pop_dat$pop_m_25_44
+pop_dat$pop_f_15_44 <- pop_dat$pop_f_15_17 +  pop_dat$pop_f_18_19 +  pop_dat$pop_f_20_24 + pop_dat$pop_f_25_44
+unemp_dat <- read.dta13("unempl_gemeinde_2008_2017.dta")
+colnames(unemp_dat)[colnames(unemp_dat) == "ags"] <- "ags_muni"
+colnames(unemp_dat)[colnames(unemp_dat) == "ags_dist"] <- "ags_county"
+
+## unemployed as share of working age population (age 15-64)
+pop_dat$pop_mf_15_64 <- pop_dat$pop_mf_15_17 +  pop_dat$pop_mf_18_19 +  pop_dat$pop_mf_20_24 + 
+  pop_dat$pop_mf_25_29 + pop_dat$pop_mf_30_34 +  pop_dat$pop_mf_35_39 + pop_dat$pop_mf_40_44 + 
+  pop_dat$pop_mf_45_49 + pop_dat$pop_mf_50_54 + pop_dat$pop_mf_55_59 + pop_dat$pop_mf_60_64
+
+pop_dat$pop_m_15_64 <- pop_dat$pop_m_15_44 + pop_dat$pop_m_45_49 + pop_dat$pop_m_50_54 + 
+  pop_dat$pop_m_55_59 + pop_dat$pop_m_60_64
+
+pop_dat$pop_f_15_64 <- pop_dat$pop_f_15_44 + pop_dat$pop_f_45_49 + pop_dat$pop_f_50_54 + 
+  pop_dat$pop_f_55_59 + pop_dat$pop_f_60_64
+
+unemp_dat_use <- unemp_dat[, c("ags_muni", "ags_county",
+                               "year", 
+                               "unempl_all_total", 
+                               "unempl_all_male_total", "unempl_all_fem_total")]
+pop_dat_m <- pop_dat[pop_dat$year >= 2011, c("ags_muni", "year", "pop_mf_15_64", "pop_m_15_64", "pop_f_15_64")]
+unemp_merge  <- merge(pop_dat_m, unemp_dat_use, by = c("ags_muni", "year"), all.x = TRUE, all.y = FALSE)
+
+unemp_merge$unemp_all_muni  <- (unemp_merge$unempl_all_total/unemp_merge$pop_mf_15_64)*100
+
+dat_2015 <- unemp_merge[unemp_merge$year == 2015, ]
+dat_2016 <- unemp_merge[unemp_merge$year == 2016, ]
+dat_2017 <- unemp_merge[unemp_merge$year == 2017, ]
+
+dat_2015$log_unemp_all_muni_anu <- log(dat_2015$unemp_all_muni  + 1) # constants are chosen to make sure all values are positive
+dat_2016$log_unemp_all_muni_anu <- log(dat_2016$unemp_all_muni  + 0.3)
+dat_2017$log_unemp_all_muni_anu <- log(dat_2017$unemp_all_muni  + 0.3)
+unemp_u <- rbind(dat_2015, dat_2016, dat_2017)
+
+context0  <- merge(context0, unemp_u[, c("ags_muni", "year",
+                                         "log_unemp_all_muni_anu")], 
+                   by = c("ags_muni", "year"), all.x = TRUE, all.y = FALSE)
+
+# 5 Unemployment Rate Gap
+d20 <- read.dta13("merged_context_2.dta")
+d2 <- unique(d20[, c("year", "ags_county", "unemp_gendergap")])
+d2$unemp_gendergap_anu <- d2$unemp_gendergap
+
+context0 <- merge(context0, d2[, c("year", "ags_county", "unemp_gendergap_anu")], 
+                  all.x = TRUE, all.y = FALSE, by = c("year", "ags_county"))
+
+# 6. Voting
+voting <- read.dta13("voting.dta")
+
+context0 <- merge(context0, voting[, c("ags_muni", 
+                                       "vote_afd_2013_muni")], by = c("ags_muni"), 
+                  all.x = TRUE, all.y = FALSE)
+
+# 7. Refugee Data
+ref_dat <- read.dta13("refugees_2008_2017.dta")
+ref_dat <- ref_dat[ref_dat$year >= 2011, ]
+
+# Creating pop_ref_anu
+ref_dat_p <- pdata.frame(ref_dat, index = c("ags_county", "year"))
+ref_dat_p$pop_ref_anu <- as.numeric(lag(ref_dat_p$pop_ref, k = 1))
+class(ref_dat_p) <- "data.frame"
+ref_dat <- ref_dat_p
+
+ref_dat <- ref_dat[ref_dat$ags_county %in% unique(ref_dat$ags_county[ref_dat$year == 2014]),]
+ref_2014 <- ref_dat[ref_dat$year == 2014, ]
+ref_2015 <- ref_dat[ref_dat$year == 2015, ]
+ref_2016 <- ref_dat[ref_dat$year == 2016, ]
+ref_2017 <- ref_dat[ref_dat$year == 2017, ]
+
+ref_2015$ref_inflow_1514 <- ref_2015$pop_ref - ref_2014$pop_ref
+ref_2016$ref_inflow_1615 <- ref_2016$pop_ref - ref_2015$pop_ref
+ref_2017$ref_inflow_1716 <- ref_2017$pop_ref - ref_2016$pop_ref
+
+ref_2015$log_ref_inflow_anu <- log(1500 + ref_2015$ref_inflow_1514) # constants are chosen such that all values are positive
+suppressWarnings(ref_2016$log_ref_inflow_anu <-  log(ref_2016$ref_inflow_1615 + 649))
+suppressWarnings(ref_2017$log_ref_inflow_anu <-  log(ref_2017$ref_inflow_1716 + 1261))
+
+ref_2015 <- ref_2015[, c("year", "ags_county", "pop_ref_anu", "log_ref_inflow_anu")]
+ref_2016 <- ref_2016[, c("year", "ags_county", "pop_ref_anu", "log_ref_inflow_anu")]
+ref_2017 <- ref_2017[, c("year", "ags_county", "pop_ref_anu", "log_ref_inflow_anu")]
+ref_data_c <- rbind(ref_2015, ref_2016, ref_2017)
+
+context0 <- merge(context0, ref_data_c[, c("year", "ags_county", 
+                                           "pop_ref_anu",
+                                           "log_ref_inflow_anu")], 
+                  by = c("year", "ags_county"), all.x = TRUE, all.y = FALSE)
+
+#  8. Violence
+crime <- read.dta13("crime.dta")
+pop <- read.dta13("pop_kreise_2015_2017.dta")
+
+crime2 <- merge(crime[, c("year", "ags_county", "violence_num_cases")],
+                pop[, c("year", "ags_county", "population")],
+                by = c("year", "ags_county"), all.x = TRUE, all.y = FALSE)
+crime2$violence_percap_anu <- crime2$violence_num_cases/crime2$population
+
+context0 <- merge(context0, crime2[, c("year", "ags_county", 
+                                       "violence_percap_anu")], by = c("year", "ags_county"), 
+                  all.x = TRUE, all.y = FALSE)
+
+# 9. Education
+edu <- read.dta13("merged_context_2.dta")
+edu  <- edu[edu$year ==  2011, ]
+edu <- edu[, c("ags_county", "pc_hidegree_all2011")]
+
+
+context0 <- merge(context0, edu[, c("ags_county", "pc_hidegree_all2011")], by = c("ags_county"), all.x = TRUE)
+
+#  10. Industry 
+manu0 <- read.dta13("merged_context_2.dta")
+manu0 <- manu0[, c("year", "ags_county", "pc_manufacturing")]
+manu <- manu0[manu0$year >= 2011 & manu0$year <= 2015, ]
+rownames(manu) <- NULL
+manu_orig <- manu
+
+manu  <- manu_orig[manu_orig$year ==  2015, ]
+manu <- manu[, c("ags_county", "pc_manufacturing")]
+manu <- manu[is.element(manu$ags_county, unique(base_pl$ags_county)),]
+manu$pc_manufacturing_2015 <- manu$pc_manufacturing
+
+manu2011  <- manu_orig[manu_orig$year ==  2011, ]
+manu2011 <- manu2011[, c("ags_county", "pc_manufacturing")]
+manu2011 <- manu2011[is.element(manu2011$ags_county, unique(base_pl$ags_county)),]
+manu2011$pc_manufacturing_2011 <- manu2011$pc_manufacturing
+
+# d_manuf1115
+manu$d_manuf1115 <- manu$pc_manufacturing_2015 - manu2011$pc_manufacturing_2011
+
+context0 <- merge(context0, manu[, c("ags_county", 
+                                     "pc_manufacturing_2015",
+                                     "d_manuf1115")], by = c("ags_county"), all.x = TRUE)
+
+
+# 11. Create additional variables 
+context0$Hate_all_muni_bin <- as.numeric(context0$Hate_all_muni >  0)
+
+context0 <- context0[order(context0$year, context0$ags_muni), ]
+save.dta13(context0, file = "context_placebo.dta")

32/replication_package/number_in_texts.R

@@ -0,0 +1,223 @@
+# Replication  File for numbers we mention in the main text of the paper. 
+# Please see "ContextAnalysis_Main.R", "SurveyAnalysis_Main.R", "ContextAnalysis_Appendix.R", and "SurveyAnalysis_Appendix.R"
+# to reproduce Tables and Figures in the paper and in the appendix. 
+
+# R version 4.0.2 (2020-06-22)
+
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+source("Help.R")
+
+# ###############
+# Section: Existing Explanations and Mate Competition
+# ###############
+you_data <- read.dta13(file  = "YouGov.dta")
+
+# The number of people who think marriages between a German woman and a non-German man is common and very common
+prop.table(table(you_data$int_marriage))[3:4]
+
+# ###############
+# Section: Refugees and Mate Competition: A Topic of Debate
+# ###############
+# The number of hate crimes in 2015 and 2016 
+dat <- read.dta13("context.dta")
+tapply(dat$Hate_all_muni, dat$year, sum)[1:2]
+
+
+# ###############
+# Section: Empirical Analyses
+# ###############
+# ###############
+# Sub Section: Mate Competition and the Incidence of Anti-Refugee Hate Crime
+# ###############
+# The proportion of municipalities that witnessed at least one hate crime in each year
+round(tapply(dat$Hate_all_muni_bin, dat$year, mean)*100, 1)
+
+# The proportion of municipalities that witnessed at least one hate crime in three years
+dat_2015 <- dat[dat$year == 2015, ]
+dat_2016 <- dat[dat$year == 2016, ]
+dat_2017 <- dat[dat$year == 2017, ]
+dat_2015$Hate_all_muni_1517 <- dat_2015$Hate_all_muni + dat_2016$Hate_all_muni + dat_2017$Hate_all_muni
+dat_2015$Hate_all_muni_1517_bin <- ifelse(dat_2015$Hate_all_muni_1517 > 0, 1, 0)
+round(mean(dat_2015$Hate_all_muni_1517_bin)*100)
+
+# Point Estimates that we mention when discussing Figure 1
+# Remove Extreme Value of Excess Males 
+range_x <- quantile(dat_2015$pop_15_44_muni_gendergap_2015, c(0.025, 0.975), na.rm = TRUE)
+dat_2015_s <- dat_2015[dat_2015$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+                         dat_2015$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+dat_s <- dat[dat$pop_15_44_muni_gendergap_2015 >= range_x[1] & 
+               dat$pop_15_44_muni_gendergap_2015 <= range_x[2], ]
+
+# The size of population for exxluded municipalities
+dat_exc <- dat[dat$pop_15_44_muni_gendergap_2015 < range_x[1] | 
+                 dat$pop_15_44_muni_gendergap_2015 > range_x[2], ]
+
+ceiling(median(dat_exc$population_muni_2015, na.rm = TRUE)) # 247 
+
+# sum
+bin_1_sum <- bin.summary(Hate_all_muni_1517_bin ~ 
+                           pop_15_44_muni_gendergap_2015 + 
+                           log_population_muni_2015 + log_popdens_muni_2015 + 
+                           log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                           log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                           pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                           unemp_gendergap_2015 + as.factor(ags_state),
+                         id  = "ags_county", data = dat_2015_s)
+
+# annual
+bin_1_p <- bin.summary(Hate_all_muni_bin ~ 
+                         pop_15_44_muni_gendergap_2015 + 
+                         log_population_muni_2015 + log_popdens_muni_2015 + 
+                         log_unemp_all_muni_2015 + d_pop1511_muni + vote_afd_2013_muni + 
+                         log_ref_inflow_1514  + log_pop_ref_2014 + log_violence_percap_2015 + ## county level
+                         pc_hidegree_all2011 + d_manuf1115 + pc_manufacturing_2015 + ## county level
+                         unemp_gendergap_2015 + as.factor(ags_state) + as.factor(year),
+                       id  = "ags_county", data = dat_s)
+
+# Excess Males
+# Effect Estimation 
+bin_1_sum_effect <- marginal_effect(bin_1_sum,
+                                    newdata = dat_2015_s, family = "logit",
+                                    main_var = "pop_15_44_muni_gendergap_2015",
+                                    difference = TRUE, 
+                                    treat_range = c(1, 1.2))
+
+bin_1_p_effect <- marginal_effect(bin_1_p,
+                                  newdata = dat_s, family = "logit",
+                                  main_var = "pop_15_44_muni_gendergap_2015",
+                                  difference = TRUE, 
+                                  treat_range = c(1, 1.2))
+
+# Point estimate
+round(bin_1_sum_effect$out_main[1:3]*100, 2)
+# 2.5%       97.5% 
+# 0.88  2.60  4.19 
+round(bin_1_p_effect$out_main[1:3]*100, 2)
+# 2.5%       97.5% 
+# 0.76  1.71  2.55 
+
+
+## Comparing to round(bin_1_sum_effect$out_main[1:3]*100, 2)
+## The Effect of Unemployment 
+effect_unemp <- marginal_effect(bin_1_sum,
+                                newdata = dat_2015_s, family = "logit",
+                                main_var = "log_unemp_all_muni_2015",
+                                difference = TRUE, 
+                                treat_range = quantile(dat_2015_s$log_unemp_all_muni_2015, prob = c(0.2, 0.8), 
+                                                       na.rm = TRUE))
+round(effect_unemp$out_main[1:3]*100, 2) # 2.60 is more than half of 4.24 
+
+## The Effect of Education
+effect_educ <- marginal_effect(bin_1_sum,
+                               newdata = dat_2015_s, family = "logit",
+                               main_var = "pc_hidegree_all2011",
+                               difference = TRUE, 
+                               treat_range = quantile(dat_2015_s$pc_hidegree_all2011, prob = c(0.8, 0.2), 
+                                                      na.rm = TRUE))
+round(effect_educ$out_main[1:3]*100, 2) 
+# 2.60 is more than twice of 1.20
+
+
+# Correlation between Excess Males and Male Disadvantage
+round(cor(dat_2015_s$pop_15_44_muni_gendergap_2015, dat_2015_s$unemp_gendergap_2015, use = "complete.obs"), 3)
+
+
+# ###############
+# Sub Section: Mate Competition and Support for Anti-Refugee Hate Crime
+# ###############
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+# install.packages("pBrackets") # pBrackets_1.0  
+# install.packages("stargazer") # stargazer_5.2.2
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+require(pBrackets) # pBrackets_1.0  
+require(stargazer) # stargazer_5.2.2
+source("Help.R")
+
+dat <- read.dta13(file =  "survey.dta")
+
+# Subset to people in the wave 4
+dat_use <- dat[dat$wave == 4, ]
+
+# Prepare Two data sets
+dat_male <- dat_use[dat_use$gender == "Male" & dat_use$age <= 44 & dat_use$age >= 18, ]
+dat_male_y <- dat_use[dat_use$gender == "Male" & dat_use$age <= 40 & dat_use$age >= 30, ]
+
+# Overall Samples 
+dat_use$MateComp.cont_bin <- ifelse(dat_use$MateComp.cont >= 3, 1, 0)
+dat_use$excess_c <- ifelse(dat_use$pop_15_44_muni_gendergap_2015 < 1.04, "1", 
+                           ifelse(dat_use$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+mean_all <- tapply(dat_use$MateComp.cont_bin, dat_use$excess_c, mean)
+se_all <- tapply(dat_use$MateComp.cont_bin, dat_use$excess_c, sd)/sqrt(table(dat_use$excess_c))
+
+# Male (18 - 44)
+dat_male$MateComp.cont_bin <- ifelse(dat_male$MateComp.cont >= 3, 1, 0)
+dat_male$excess_c <- ifelse(dat_male$pop_15_44_muni_gendergap_2015 < 1.04, "1", 
+                            ifelse(dat_male$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+mean_all_m <- tapply(dat_male$MateComp.cont_bin, dat_male$excess_c, mean)
+se_all_m <- tapply(dat_male$MateComp.cont_bin, dat_male$excess_c, sd)/sqrt(table(dat_male$excess_c))
+
+# Male (30 - 40)
+dat_male_y$MateComp.cont_bin <- ifelse(dat_male_y$MateComp.cont >= 3, 1, 0)
+dat_male_y$excess_c <- ifelse(dat_male_y$pop_15_44_muni_gendergap_2015 < 1.04, "1", 
+                              ifelse(dat_male_y$pop_15_44_muni_gendergap_2015 < 1.12, "2", "3"))
+mean_all_y <- tapply(dat_male_y$MateComp.cont_bin, dat_male_y$excess_c, mean)
+se_all_y <- tapply(dat_male_y$MateComp.cont_bin, dat_male_y$excess_c, sd)/sqrt(table(dat_male_y$excess_c))
+
+
+round(mean_all,2)[c(1,3)] ## 0.18 0.22 
+round(mean_all_m,2)[c(1,3)] ## 0.23 0.38 
+round(mean_all_y,2)[c(1,3)] ## 0.17 0.47 
+
+# ################
+# List Experiment
+# ###############
+rm(list=ls())
+# install.packages("readstata13")  #  readstata13_0.9.2
+# install.packages("MASS") # MASS_7.3-51.6  
+# install.packages("sandwich")  # sandwich_2.5-1 
+# install.packages("lmtest") # lmtest_0.9-37 
+# install.packages("list") # list_9.2 
+
+require(readstata13)  #  readstata13_0.9.2
+require(MASS) # MASS_7.3-51.6  
+require(sandwich)  # sandwich_2.5-1 
+require(lmtest) # lmtest_0.9-37  
+require(list) # list_9.2 
+
+dat <- read.dta13(file =  "survey.dta")
+data.u2 <- dat[dat$wave == 2, ]
+
+# Means: When it comes to the refugee problem, violence is sometimes the only means that citizens have to get the attention of German politicians
+data.list.u2   <- data.u2[data.u2$list == "1",]
+data.direct.u2 <- data.u2[data.u2$list == "2",]
+data.list.u2 <- data.list.u2[is.na(data.list.u2$treatment_list)==FALSE,]
+data.list.u2$List.treat <- ifelse(data.list.u2$treatment_list == "Scenario 2", 1, 0)
+
+# The mean for Control Group
+round(mean(data.list.u2$outcome_list[data.list.u2$List.treat == 0]), 2)
+# The mean for Treatment Group
+round(mean(data.list.u2$outcome_list[data.list.u2$List.treat == 1]), 2)
+# Please see "SurveyAnalysis_Main.R" for complete analysis, which was used to create Figure 3. 
+# Here, we only reproduce the code for numbers we mention in the paper. 
+
+## Compare to All People who answered Direct Question (n = 2170)
+data.u2.all.direct <- data.u2[is.na(data.u2$hate_violence_means) == FALSE, ]
+data.u2.all.direct$hate.direct.bin <- ifelse(data.u2.all.direct$hate_violence_means >=3, 1, 0)
+round(mean(data.u2.all.direct$hate.direct.bin)*100) ## 18

32/replication_package/out_count_table.rdata

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1ccd234462ece80c73745d90e9e27c76c646e09715a4f23e4f875725a6006dd5
+size 10955242

32/replication_package/produce_context_data.do

@@ -0,0 +1,89 @@
+*************************************************************
+*Purpose: Merge context data and generate additional vars
+*Stata Version: 16
+*************************************************************
+
+*-------------------------------
+*generate merged_context_1.dta
+*-------------------------------
+use "source_data/unemployment.dta", clear
+
+*unemployment gendergap
+*----------------------
+gen unemp_gendergap = unemp_men/unemp_female
+label var unemp_gendergap "male unemployment rate / female unemployment rate"
+
+rename ags_dist ags_county // "county" and "district" used as synonyms
+label var ags_c " Identifier for County"
+save "source_data/merged_context_1.dta",replace
+
+*-----------------------------
+*generate merged_context_2.dta
+*-----------------------------
+
+*unemployment data
+*-------------------
+use "source_data/unemployment.dta", clear
+
+gen unemp_gendergap = unemp_men/unemp_female
+label var unemp_gendergap "male unemployment rate / female unemployment rate"
+
+save "source_data/temp1.dta", replace
+
+*refugees by gender
+*-------------------
+use "source_data/refugee_gender.dta", clear
+
+egen ref_male = rowtotal(all_male0_3 all_male3_6 all_male6_15 all_male15_18 all_male18_25 all_male25_30 all_male30_40 all_male40_50 all_male50_65 all_male65_75 all_male75_up)
+gen pc_ref_male = ref_male*100/all_totref
+label var pc_ref_male "% male refugees, of all refugees"
+label var ref_male "total number of male refugees (all ages)"
+
+save "source_data/temp2.dta", replace
+
+*education
+*----------
+use "source_data/education.dta", clear
+
+*high degree
+gen pc_hidegree_all2011 = pop15_high_degree*100/pop15_total
+label var pc_hidegree_all2011 "% population with university entrance exam, incl. still in school" // (census 2011)
+
+save "source_data/temp3.dta", replace
+
+
+* sector
+*--------
+use "source_data/sectors.dta", clear
+
+gen pc_manufacturing = no_manufacturing/no_employed
+label var pc_manufacturing "pc_manufacturing"
+
+save "source_data/temp4.dta", replace
+
+*merge data
+*----------
+/*
+this is a more comprehensive dataset - start by using a master data file
+that includes the ags year combinations we need
+*/
+
+use  "source_data/population.dta"
+
+merge 1:1 ags year using "source_data/temp1.dta" // unemployment
+drop _m 
+merge 1:1 ags year using "source_data/temp2.dta" // refugees
+drop _m
+merge m:1 ags using "source_data/temp3.dta" // education in 2011 this m:1 merge
+drop _m
+merge 1:1 ags year using "source_data/temp4.dta" //sectors
+drop _m
+
+rename ags_dist ags_county // "county" and "district" used as synonyms
+label var ags_c " Identifier for County"
+
+save "source_data/merged_context_2.dta",replace
+erase "source_data/temp1.dta"
+erase "source_data/temp2.dta"
+erase "source_data/temp3.dta"
+erase "source_data/temp4.dta"

32/replication_package/source_context.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f795cd86bfb2e74eef85962eca903471790639cd848c998934406a6208d6eed3
+size 157393

32/replication_package/source_context_placebo.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9a93779e37456c5055ae0b1386564fbba9616c1c8669176130110fd177dedc88
+size 180992

32/replication_package/source_data/area_mun.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fbbc3f1c2ef170246e40c6b32ce440b9d60aaae466c89f51a8d81ba278988b2e
+size 765299

32/replication_package/source_data/base.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5af5c1de4f9362caa1eb1fe0233241b34f721512416bf8daa79a69fd79ec66a
+size 844133

32/replication_package/source_data/base_pl.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da54848f03aae66e23beef62ef0edd781f03869355c09408b72c31d002b65fba
+size 694698

32/replication_package/source_data/crime.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3d93d6b20b821852be117c8715e33636ac432aeaf2314d20eedca14107b66573
+size 21664

32/replication_package/source_data/education.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fee3371a6aa2fbdb92bf1bfd861f0d460f2eb55daa9d9f5f08e1dc47cac50865
+size 39650

32/replication_package/source_data/hate.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17d13a62e1abead6692c79bb476766e41e6fd64701be85c14549f77afeb6130d
+size 2143165

32/replication_package/source_data/merged_context_1.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2df0c90bf9b8e38d80bab08f85de2b3f87df05861233b2deb276c3ed9162e7b9
+size 241539

32/replication_package/source_data/merged_context_2.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf108efd04187720950b11bb72d30c929a26ae6ffa4ba009bafc37694f2d3cad
+size 1323816

32/replication_package/source_data/pop_gemeinde_2008_2018.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d80379610f6d011650aee5a4e8d5e289a626460ae568dd7feae5f6257bbb27d5
+size 77180609

32/replication_package/source_data/pop_kreise_2015_2017.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea3561218aa950bd9c682d3d242ab979e9d5bc23177e7ea6d268468a0ce7c958
+size 26488

32/replication_package/source_data/population.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7fe4cba1be98e195be54a00e5a827d5b1e2fda56093054bcafa22f2ec909a93
+size 46987

32/replication_package/source_data/refugee_gender.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b2126acb78af60c555ccaebbe1c024401569df114540b0a9ab47d61da06f264
+size 127827

32/replication_package/source_data/refugees_2008_2017.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8fc45783e4a4f51ba663d22f766b5fe56639a075ae9c5f9b9e635602632f4118
+size 1380624

32/replication_package/source_data/sectors.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5bca2f44da65dee4c54c9ff2804b69b6a911715ae96a59198a0f6a367a93f456
+size 488307

32/replication_package/source_data/unempl_gemeinde_2008_2017.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4c2ba9d7c5964f1671530e94d565bc029b2355dbd0e1dc78ef4a82ab0b948be
+size 8082440

32/replication_package/source_data/unemployment.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:714cc77c4840d0fa99a155e2dd7498b2b0c6a63a4effe64381adf44b82a59513
+size 231439

32/replication_package/source_data/voting.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7be35304a013490ee92c5fbb6dedc75032dc731f70eaab99833448f8b0186fb9
+size 1705216

32/replication_package/survey.dta

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40d6fa1133cfe63c006fc7d37c89a0980f1cf2559fa83537d0ac4339b314fc6e
+size 4702645

32/replication_package/table1.tex

@@ -0,0 +1,34 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:30:28
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{6}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-7} 
+\\[-1.8ex] & \multicolumn{5}{c}{hate\_violence\_means} & formula.6 \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5) & (6)\\ 
+\hline \\[-1.8ex] 
+ Mate Competition & 0.437$^{***}$ & 0.263$^{***}$ & 0.236$^{***}$ & 0.206$^{***}$ & 0.185$^{***}$ & 0.155$^{***}$ \\ 
+  & (0.016) & (0.020) & (0.021) & (0.019) & (0.019) & (0.019) \\ 
+  & & & & & & \\ 
+ Job Competition &  & 0.250$^{***}$ & 0.236$^{***}$ & 0.077$^{***}$ & 0.065$^{***}$ & 0.056$^{***}$ \\ 
+  &  & (0.019) & (0.019) & (0.020) & (0.020) & (0.019) \\ 
+  & & & & & & \\ 
+ Life Satisfaction &  & $-$0.015$^{**}$ & $-$0.014$^{*}$ & $-$0.003 & $-$0.002 & $-$0.0001 \\ 
+  &  & (0.006) & (0.007) & (0.006) & (0.006) & (0.006) \\ 
+  & & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 3,019 & 3,019 & 3,008 & 3,008 & 3,008 & 3,008 \\ 
+R$^{2}$ & 0.191 & 0.240 & 0.288 & 0.394 & 0.410 & 0.459 \\ 
+Adjusted R$^{2}$ & 0.191 & 0.240 & 0.267 & 0.371 & 0.382 & 0.431 \\ 
+Residual Std. Error & 0.799 (df = 3017) & 0.775 (df = 3015) & 0.760 (df = 2921) & 0.704 (df = 2897) & 0.698 (df = 2873) & 0.670 (df = 2857) \\ 
+F Statistic & 714.588$^{***}$ (df = 1; 3017) & 317.891$^{***}$ (df = 3; 3015) & 13.756$^{***}$ (df = 86; 2921) & 17.141$^{***}$ (df = 110; 2897) & 14.878$^{***}$ (df = 134; 2873) & 16.170$^{***}$ (df = 150; 2857) \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{6}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C1.tex

@@ -0,0 +1,62 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:32:51
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{6}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-7} 
+\\[-1.8ex] & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5) & (6)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 15 - 44) & 2.127$^{***}$ & 2.132$^{***}$ & 1.598$^{***}$ & 1.549$^{***}$ & 1.406$^{***}$ & 1.537$^{***}$ \\ 
+  & (0.491) & (0.432) & (0.491) & (0.411) & (0.463) & (0.412) \\ 
+  & & & & & & \\ 
+ Log (Population) & 1.730$^{***}$ & 1.519$^{***}$ & 1.619$^{***}$ & 1.392$^{***}$ & 1.522$^{***}$ & 1.372$^{***}$ \\ 
+  & (0.056) & (0.041) & (0.060) & (0.044) & (0.055) & (0.042) \\ 
+  & & & & & & \\ 
+ Log (Population Density) & 0.098 & 0.087$^{*}$ & 0.052 & 0.039 & 0.017 & $-$0.002 \\ 
+  & (0.070) & (0.052) & (0.070) & (0.051) & (0.066) & (0.049) \\ 
+  & & & & & & \\ 
+ Log (Unemployment Rate) &  &  & 1.087$^{***}$ & 1.028$^{***}$ & 0.634$^{***}$ & 0.715$^{***}$ \\ 
+  &  &  & (0.180) & (0.142) & (0.156) & (0.130) \\ 
+  & & & & & & \\ 
+ % of population change (2011 vs 2015) &  &  & $-$0.617 & $-$0.235 & $-$0.299 & 0.020 \\ 
+  &  &  & (0.975) & (0.608) & (0.851) & (0.541) \\ 
+  & & & & & & \\ 
+ Vote share for AfD (2013) &  &  & 5.618$^{*}$ & 4.822$^{**}$ & 5.358 & 3.178 \\ 
+  &  &  & (3.116) & (2.406) & (3.384) & (2.718) \\ 
+  & & & & & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) &  &  &  &  & 0.857$^{***}$ & 0.734$^{***}$ \\ 
+  &  &  &  &  & (0.324) & (0.261) \\ 
+  & & & & & & \\ 
+ Log (Refugee Size) (2014) &  &  &  &  & $-$0.210$^{**}$ & $-$0.192$^{**}$ \\ 
+  &  &  &  &  & (0.104) & (0.084) \\ 
+  & & & & & & \\ 
+ Log (General Violence per capita) &  &  &  &  & 0.136 & 0.022 \\ 
+  &  &  &  &  & (0.189) & (0.151) \\ 
+  & & & & & & \\ 
+ % of High Education &  &  &  &  & $-$0.022$^{*}$ & $-$0.018$^{*}$ \\ 
+  &  &  &  &  & (0.013) & (0.011) \\ 
+  & & & & & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) &  &  &  &  & 8.177$^{**}$ & 9.588$^{***}$ \\ 
+  &  &  &  &  & (4.062) & (3.167) \\ 
+  & & & & & & \\ 
+ Share of Manufacturing &  &  &  &  & 0.057 & $-$0.306 \\ 
+  &  &  &  &  & (0.750) & (0.606) \\ 
+  & & & & & & \\ 
+ Male Disadvantage &  &  &  &  & 0.920$^{**}$ & 0.846$^{***}$ \\ 
+  &  &  &  &  & (0.371) & (0.319) \\ 
+  & & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 10,307 & 30,921 & 10,029 & 30,087 & 9,282 & 27,846 \\ 
+Log Likelihood & $-$2,813.561 & $-$5,645.573 & $-$2,771.915 & $-$5,487.582 & $-$2,776.250 & $-$5,290.740 \\ 
+Akaike Inf. Crit. & 6,433.122 & 12,101.150 & 6,141.830 & 11,577.160 & 5,600.500 & 10,633.480 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{6}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C2.tex

@@ -0,0 +1,62 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:35:11
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{6}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-7} 
+\\[-1.8ex] & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5) & (6)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 25 - 44) & 1.595$^{***}$ & 1.794$^{***}$ & 0.958$^{**}$ & 1.066$^{***}$ & 0.932$^{**}$ & 1.192$^{***}$ \\ 
+  & (0.454) & (0.400) & (0.456) & (0.373) & (0.430) & (0.364) \\ 
+  & & & & & & \\ 
+ Log (Population) & 1.726$^{***}$ & 1.511$^{***}$ & 1.614$^{***}$ & 1.387$^{***}$ & 1.518$^{***}$ & 1.365$^{***}$ \\ 
+  & (0.056) & (0.041) & (0.059) & (0.043) & (0.054) & (0.042) \\ 
+  & & & & & & \\ 
+ Log (Population Density) & 0.082 & 0.073 & 0.030 & 0.019 & 0.001 & $-$0.014 \\ 
+  & (0.070) & (0.052) & (0.070) & (0.051) & (0.066) & (0.049) \\ 
+  & & & & & & \\ 
+ Log (Unemployment Rate) &  &  & 1.120$^{***}$ & 1.039$^{***}$ & 0.642$^{***}$ & 0.704$^{***}$ \\ 
+  &  &  & (0.184) & (0.145) & (0.157) & (0.131) \\ 
+  & & & & & & \\ 
+ % of population change (2011 vs 2015) &  &  & $-$0.463 & $-$0.023 & $-$0.207 & 0.156 \\ 
+  &  &  & (0.905) & (0.554) & (0.816) & (0.511) \\ 
+  & & & & & & \\ 
+ Vote share for AfD (2013) &  &  & 5.974$^{*}$ & 5.282$^{**}$ & 5.776$^{*}$ & 3.771 \\ 
+  &  &  & (3.088) & (2.400) & (3.366) & (2.699) \\ 
+  & & & & & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) &  &  &  &  & 0.859$^{***}$ & 0.705$^{***}$ \\ 
+  &  &  &  &  & (0.324) & (0.263) \\ 
+  & & & & & & \\ 
+ Log (Refugee Size) (2014) &  &  &  &  & $-$0.211$^{**}$ & $-$0.191$^{**}$ \\ 
+  &  &  &  &  & (0.103) & (0.084) \\ 
+  & & & & & & \\ 
+ Log (General Violence per capita) &  &  &  &  & 0.130 & 0.019 \\ 
+  &  &  &  &  & (0.187) & (0.149) \\ 
+  & & & & & & \\ 
+ % of High Education &  &  &  &  & $-$0.022$^{*}$ & $-$0.018$^{*}$ \\ 
+  &  &  &  &  & (0.013) & (0.010) \\ 
+  & & & & & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) &  &  &  &  & 8.213$^{**}$ & 9.464$^{***}$ \\ 
+  &  &  &  &  & (4.029) & (3.139) \\ 
+  & & & & & & \\ 
+ Share of Manufacturing &  &  &  &  & 0.076 & $-$0.318 \\ 
+  &  &  &  &  & (0.748) & (0.603) \\ 
+  & & & & & & \\ 
+ Male Disadvantage &  &  &  &  & 0.910$^{**}$ & 0.856$^{***}$ \\ 
+  &  &  &  &  & (0.369) & (0.316) \\ 
+  & & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 10,378 & 31,134 & 10,097 & 30,291 & 9,288 & 27,864 \\ 
+Log Likelihood & $-$2,823.656 & $-$5,669.987 & $-$2,781.008 & $-$5,512.023 & $-$2,782.381 & $-$5,312.211 \\ 
+Akaike Inf. Crit. & 6,453.312 & 12,149.970 & 6,160.016 & 11,626.050 & 5,612.761 & 10,676.420 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{6}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C3.tex

@@ -0,0 +1,64 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:35:27
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{6}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-7} 
+\\[-1.8ex] & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5) & (6)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 15 - 44) & 0.116$^{***}$ & 0.069$^{***}$ & 0.089$^{***}$ & 0.051$^{***}$ & 0.095$^{***}$ & 0.057$^{***}$ \\ 
+  & (0.027) & (0.015) & (0.025) & (0.013) & (0.026) & (0.014) \\ 
+  & & & & & & \\ 
+ Log (Population) & 0.162$^{***}$ & 0.086$^{***}$ & 0.156$^{***}$ & 0.081$^{***}$ & 0.152$^{***}$ & 0.078$^{***}$ \\ 
+  & (0.009) & (0.005) & (0.008) & (0.004) & (0.008) & (0.004) \\ 
+  & & & & & & \\ 
+ Log (Population Density) & 0.011 & 0.011$^{***}$ & 0.010 & 0.011$^{***}$ & 0.008 & 0.008$^{**}$ \\ 
+  & (0.007) & (0.004) & (0.007) & (0.003) & (0.007) & (0.003) \\ 
+  & & & & & & \\ 
+ Log (Unemployment Rate) &  &  & 0.137$^{***}$ & 0.090$^{***}$ & 0.093$^{***}$ & 0.066$^{***}$ \\ 
+  &  &  & (0.015) & (0.009) & (0.014) & (0.008) \\ 
+  & & & & & & \\ 
+ % of population change (2011 vs 2015) &  &  & $-$0.059 & $-$0.031 & 0.024 & 0.025 \\ 
+  &  &  & (0.086) & (0.041) & (0.100) & (0.050) \\ 
+  & & & & & & \\ 
+ Vote share for AfD (2013) &  &  & 0.307$^{**}$ & 0.120$^{*}$ & 0.370$^{**}$ & 0.106 \\ 
+  &  &  & (0.147) & (0.069) & (0.182) & (0.089) \\ 
+  & & & & & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) &  &  &  &  & 0.075 & 0.046 \\ 
+  &  &  &  &  & (0.051) & (0.029) \\ 
+  & & & & & & \\ 
+ Log (Refugee Size) (2014) &  &  &  &  & $-$0.019 & $-$0.011 \\ 
+  &  &  &  &  & (0.014) & (0.007) \\ 
+  & & & & & & \\ 
+ Log (General Violence per capita) &  &  &  &  & 0.010 & 0.007 \\ 
+  &  &  &  &  & (0.021) & (0.011) \\ 
+  & & & & & & \\ 
+ % of High Education &  &  &  &  & $-$0.003$^{*}$ & $-$0.002$^{**}$ \\ 
+  &  &  &  &  & (0.002) & (0.001) \\ 
+  & & & & & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) &  &  &  &  & 0.725 & 0.583$^{**}$ \\ 
+  &  &  &  &  & (0.489) & (0.243) \\ 
+  & & & & & & \\ 
+ Share of Manufacturing &  &  &  &  & 0.021 & $-$0.023 \\ 
+  &  &  &  &  & (0.092) & (0.048) \\ 
+  & & & & & & \\ 
+ Male Disadvantage &  &  &  &  & 0.114$^{**}$ & 0.063$^{**}$ \\ 
+  &  &  &  &  & (0.047) & (0.025) \\ 
+  & & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 10,307 & 30,921 & 10,029 & 30,087 & 9,282 & 27,846 \\ 
+R$^{2}$ & 0.366 & 0.252 & 0.347 & 0.204 & 0.302 & 0.180 \\ 
+Adjusted R$^{2}$ & 0.340 & 0.242 & 0.327 & 0.196 & 0.301 & 0.179 \\ 
+Residual Std. Error & 0.316 (df = 9904) & 0.247 (df = 30516) & 0.314 (df = 9730) & 0.246 (df = 29786) & 0.320 (df = 9258) & 0.248 (df = 27820) \\ 
+F Statistic & 14.206$^{***}$ (df = 402; 9904) & 25.480$^{***}$ (df = 404; 30516) & 17.368$^{***}$ (df = 298; 9730) & 25.406$^{***}$ (df = 300; 29786) & 174.490$^{***}$ (df = 23; 9258) & 244.318$^{***}$ (df = 25; 27820) \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{6}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C4.tex

@@ -0,0 +1,62 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:38:34
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{6}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-7} 
+\\[-1.8ex] & Physical\_muni\_1517\_bin & Physical\_muni\_bin & Physical\_muni\_1517\_bin & Physical\_muni\_bin & Physical\_muni\_1517\_bin & Physical\_muni\_bin \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5) & (6)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 15 - 44) & 3.773$^{***}$ & 3.038$^{***}$ & 2.624$^{***}$ & 1.906$^{**}$ & 2.719$^{***}$ & 1.634$^{*}$ \\ 
+  & (0.938) & (0.818) & (0.910) & (0.744) & (0.944) & (0.902) \\ 
+  & & & & & & \\ 
+ Log (Population) & 1.775$^{***}$ & 1.640$^{***}$ & 1.562$^{***}$ & 1.429$^{***}$ & 1.442$^{***}$ & 1.366$^{***}$ \\ 
+  & (0.120) & (0.090) & (0.124) & (0.100) & (0.106) & (0.101) \\ 
+  & & & & & & \\ 
+ Log (Population Density) & 0.096 & 0.071 & 0.080 & 0.045 & 0.046 & $-$0.018 \\ 
+  & (0.134) & (0.109) & (0.129) & (0.103) & (0.113) & (0.105) \\ 
+  & & & & & & \\ 
+ Log (Unemployment Rate) &  &  & 1.591$^{***}$ & 1.515$^{***}$ & 1.179$^{***}$ & 1.309$^{***}$ \\ 
+  &  &  & (0.346) & (0.304) & (0.282) & (0.256) \\ 
+  & & & & & & \\ 
+ % of population change (2011 vs 2015) &  &  & $-$0.837 & $-$0.368 & 0.023 & 0.154 \\ 
+  &  &  & (0.772) & (0.721) & (0.580) & (0.555) \\ 
+  & & & & & & \\ 
+ Vote share for AfD (2013) &  &  & 2.058 & 0.875 & 0.588 & 0.763 \\ 
+  &  &  & (6.640) & (6.184) & (5.209) & (5.080) \\ 
+  & & & & & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) &  &  &  &  & 0.262 & $-$0.091 \\ 
+  &  &  &  &  & (0.686) & (0.646) \\ 
+  & & & & & & \\ 
+ Log (Refugee Size) (2014) &  &  &  &  & $-$0.429$^{**}$ & $-$0.288 \\ 
+  &  &  &  &  & (0.210) & (0.202) \\ 
+  & & & & & & \\ 
+ Log (General Violence per capita) &  &  &  &  & $-$0.303 & $-$0.288 \\ 
+  &  &  &  &  & (0.383) & (0.383) \\ 
+  & & & & & & \\ 
+ % of High Education &  &  &  &  & $-$0.011 & $-$0.018 \\ 
+  &  &  &  &  & (0.024) & (0.023) \\ 
+  & & & & & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) &  &  &  &  & 4.548 & $-$0.633 \\ 
+  &  &  &  &  & (8.376) & (8.393) \\ 
+  & & & & & & \\ 
+ Share of Manufacturing &  &  &  &  & 1.859 & 2.050 \\ 
+  &  &  &  &  & (1.742) & (1.632) \\ 
+  & & & & & & \\ 
+ Male Disadvantage &  &  &  &  & 1.258 & 1.217 \\ 
+  &  &  &  &  & (0.800) & (0.786) \\ 
+  & & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 10,307 & 30,921 & 10,029 & 30,087 & 9,282 & 27,846 \\ 
+Log Likelihood & $-$842.859 & $-$1,447.433 & $-$820.949 & $-$1,339.173 & $-$884.639 & $-$1,339.369 \\ 
+Akaike Inf. Crit. & 2,491.718 & 3,704.866 & 2,239.899 & 3,280.346 & 1,817.279 & 2,730.737 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{6}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C5.tex

@@ -0,0 +1,63 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:38:36
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{2}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-3} 
+\\[-1.8ex] & Hate\_all\_muni\_1517 & Hate\_all\_muni \\ 
+\\[-1.8ex] & (1) & (2)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 15 - 44) & 1.150$^{***}$ & 1.246$^{***}$ \\ 
+  & (0.366) & (0.363) \\ 
+  & & \\ 
+ Log (Population) & 1.264$^{***}$ & 1.249$^{***}$ \\ 
+  & (0.039) & (0.038) \\ 
+  & & \\ 
+ Log (Population Density) & $-$0.032 & $-$0.030 \\ 
+  & (0.045) & (0.045) \\ 
+  & & \\ 
+ Log (Unemployment Rate) & 0.703$^{***}$ & 0.729$^{***}$ \\ 
+  & (0.131) & (0.129) \\ 
+  & & \\ 
+ % of population change (2011 vs 2015) & 0.072 & 0.096 \\ 
+  & (0.508) & (0.457) \\ 
+  & & \\ 
+ Vote share for AfD (2013) & 4.464$^{*}$ & 4.088 \\ 
+  & (2.670) & (2.670) \\ 
+  & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) & 0.588$^{**}$ & 0.541$^{**}$ \\ 
+  & (0.280) & (0.272) \\ 
+  & & \\ 
+ Log (Refugee Size) (2014) & $-$0.189$^{**}$ & $-$0.180$^{*}$ \\ 
+  & (0.093) & (0.093) \\ 
+  & & \\ 
+ Log (General Violence per capita) & 0.034 & $-$0.016 \\ 
+  & (0.161) & (0.165) \\ 
+  & & \\ 
+ % of High Education & $-$0.017 & $-$0.017$^{*}$ \\ 
+  & (0.011) & (0.011) \\ 
+  & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) & 10.081$^{***}$ & 10.432$^{***}$ \\ 
+  & (3.333) & (3.452) \\ 
+  & & \\ 
+ Share of Manufacturing & $-$0.028 & $-$0.035 \\ 
+  & (0.651) & (0.632) \\ 
+  & & \\ 
+ Male Disadvantage & 0.617$^{*}$ & 0.656$^{*}$ \\ 
+  & (0.337) & (0.345) \\ 
+  & & \\ 
+\hline \\[-1.8ex] 
+Observations & 9,282 & 27,846 \\ 
+Log Likelihood & $-$5,125.037 & $-$7,759.071 \\ 
+$\theta$ & 1.247$^{***}$  (0.088) & 0.842$^{***}$  (0.053) \\ 
+Akaike Inf. Crit. & 10,298.070 & 15,570.140 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{2}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C6.tex

@@ -0,0 +1,68 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:38:37
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{4}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-5} 
+\\[-1.8ex] & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 15 - 44) & 1.291$^{***}$ & 1.442$^{***}$ & 1.588$^{**}$ & 1.445$^{**}$ \\ 
+  & (0.462) & (0.410) & (0.791) & (0.701) \\ 
+  & & & & \\ 
+ West & 0.075 & $-$0.124 & 0.542 & $-$0.119 \\ 
+  & (0.161) & (0.118) & (1.063) & (0.913) \\ 
+  & & & & \\ 
+ Log (Population) & 1.484$^{***}$ & 1.340$^{***}$ & 1.485$^{***}$ & 1.340$^{***}$ \\ 
+  & (0.046) & (0.039) & (0.046) & (0.039) \\ 
+  & & & & \\ 
+ Log (Population Density) & 0.046 & 0.030 & 0.046 & 0.030 \\ 
+  & (0.061) & (0.048) & (0.061) & (0.048) \\ 
+  & & & & \\ 
+ Log (Unemployment Rate) & 0.585$^{***}$ & 0.653$^{***}$ & 0.577$^{***}$ & 0.653$^{***}$ \\ 
+  & (0.148) & (0.124) & (0.151) & (0.128) \\ 
+  & & & & \\ 
+ % of population change (2011 vs 2015) & 0.050 & 0.308 & 0.072 & 0.308 \\ 
+  & (0.772) & (0.499) & (0.765) & (0.497) \\ 
+  & & & & \\ 
+ Vote share for AfD (2013) & 3.865 & 3.183 & 3.860 & 3.183 \\ 
+  & (3.223) & (2.624) & (3.223) & (2.625) \\ 
+  & & & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) & 1.475$^{***}$ & 1.282$^{***}$ & 1.471$^{***}$ & 1.282$^{***}$ \\ 
+  & (0.326) & (0.247) & (0.328) & (0.247) \\ 
+  & & & & \\ 
+ Log (Refugee Size) (2014) & $-$0.397$^{***}$ & $-$0.386$^{***}$ & $-$0.396$^{***}$ & $-$0.386$^{***}$ \\ 
+  & (0.094) & (0.079) & (0.094) & (0.079) \\ 
+  & & & & \\ 
+ Log (General Violence per capita) & 0.083 & $-$0.092 & 0.086 & $-$0.092 \\ 
+  & (0.166) & (0.138) & (0.166) & (0.139) \\ 
+  & & & & \\ 
+ % of High Education & $-$0.027$^{**}$ & $-$0.021$^{**}$ & $-$0.027$^{**}$ & $-$0.021$^{**}$ \\ 
+  & (0.012) & (0.010) & (0.012) & (0.010) \\ 
+  & & & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) & 7.554$^{*}$ & 8.568$^{**}$ & 7.560$^{*}$ & 8.568$^{**}$ \\ 
+  & (4.241) & (3.509) & (4.247) & (3.509) \\ 
+  & & & & \\ 
+ Share of Manufacturing & $-$0.332 & $-$0.611 & $-$0.337 & $-$0.611 \\ 
+  & (0.666) & (0.504) & (0.666) & (0.505) \\ 
+  & & & & \\ 
+ Male Disadvantage & 1.077$^{***}$ & 1.015$^{***}$ & 1.084$^{***}$ & 1.016$^{***}$ \\ 
+  & (0.369) & (0.317) & (0.370) & (0.318) \\ 
+  & & & & \\ 
+ Excess Males x West &  &  & $-$0.419 & $-$0.004 \\ 
+  &  &  & (0.938) & (0.812) \\ 
+  & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 9,282 & 27,846 & 9,282 & 27,846 \\ 
+Log Likelihood & $-$2,790.647 & $-$5,310.629 & $-$2,790.555 & $-$5,310.629 \\ 
+Akaike Inf. Crit. & 5,611.295 & 10,655.260 & 5,613.110 & 10,657.260 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{4}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C7.tex

@@ -0,0 +1,65 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:38:38
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lcc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{2}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-3} 
+\\[-1.8ex] & Hate\_all\_muni\_1517\_bin & Hate\_all\_muni\_bin \\ 
+\\[-1.8ex] & (1) & (2)\\ 
+\hline \\[-1.8ex] 
+ Excess Males (Age 15 - 44) & $-$8.441 & $-$17.754 \\ 
+  & (18.923) & (15.168) \\ 
+  & & \\ 
+ Log (Refugee Inflow) (2014 vs 2015) & $-$0.544 & $-$2.018 \\ 
+  & (2.711) & (2.185) \\ 
+  & & \\ 
+ Log (Population) & 1.522$^{***}$ & 1.372$^{***}$ \\ 
+  & (0.055) & (0.042) \\ 
+  & & \\ 
+ Log (Population Density) & 0.019 & 0.001 \\ 
+  & (0.066) & (0.049) \\ 
+  & & \\ 
+ Log (Unemployment Rate) & 0.630$^{***}$ & 0.708$^{***}$ \\ 
+  & (0.155) & (0.130) \\ 
+  & & \\ 
+ % of population change (2011 vs 2015) & $-$0.299 & 0.017 \\ 
+  & (0.856) & (0.548) \\ 
+  & & \\ 
+ Vote share for AfD (2013) & 5.358 & 3.192 \\ 
+  & (3.386) & (2.726) \\ 
+  & & \\ 
+ Log (Refugee Size) (2014) & $-$0.209$^{**}$ & $-$0.190$^{**}$ \\ 
+  & (0.104) & (0.085) \\ 
+  & & \\ 
+ Log (General Violence per capita) & 0.141 & 0.035 \\ 
+  & (0.190) & (0.152) \\ 
+  & & \\ 
+ % of High Education & $-$0.022$^{*}$ & $-$0.019$^{*}$ \\ 
+  & (0.013) & (0.011) \\ 
+  & & \\ 
+ Change in Manufacturing Share (2011 vs 2015) & 8.200$^{**}$ & 9.647$^{***}$ \\ 
+  & (4.077) & (3.200) \\ 
+  & & \\ 
+ Share of Manufacturing & 0.044 & $-$0.335 \\ 
+  & (0.750) & (0.606) \\ 
+  & & \\ 
+ Male Disadvantage & 0.929$^{**}$ & 0.865$^{***}$ \\ 
+  & (0.372) & (0.321) \\ 
+  & & \\ 
+ Excess Males × Log (Refugee Inflow) & 1.290 & 2.524 \\ 
+  & (2.477) & (1.987) \\ 
+  & & \\ 
+\hline \\[-1.8ex] 
+Observations & 9,282 & 27,846 \\ 
+Log Likelihood & $-$2,776.125 & $-$5,289.906 \\ 
+Akaike Inf. Crit. & 5,602.249 & 10,633.810 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{2}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_C9.tex

@@ -0,0 +1,26 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:38:39
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{5}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-6} 
+\\[-1.8ex] & \multicolumn{5}{c}{Hate\_all\_muni\_bin} \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5)\\ 
+\hline \\[-1.8ex] 
+ Future-Treatment & $-$1.326 & 0.480 & 0.544 & $-$0.277 & $-$0.243 \\ 
+  & (1.551) & (1.061) & (1.592) & (0.773) & (1.043) \\ 
+  & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 8,939 & 8,681 & 8,679 & 26,299 & 13,147 \\ 
+Log Likelihood & $-$1,366.904 & $-$2,312.448 & $-$1,546.793 & $-$5,284.379 & $-$4,558.570 \\ 
+Akaike Inf. Crit. & 2,783.807 & 4,674.896 & 3,143.587 & 10,622.760 & 9,171.140 \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{5}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table} 

32/replication_package/table_D5_1.tex

@@ -0,0 +1,34 @@
+
+% Table created by stargazer v.5.2.2 by Marek Hlavac, Harvard University. E-mail: hlavac at fas.harvard.edu
+% Date and time: Wed, Sep 09, 2020 - 23:38:46
+\begin{table}[!htbp] \centering 
+  \caption{} 
+  \label{} 
+\begin{tabular}{@{\extracolsep{5pt}}lccccc} 
+\\[-1.8ex]\hline 
+\hline \\[-1.8ex] 
+ & \multicolumn{5}{c}{\textit{Dependent variable:}} \\ 
+\cline{2-6} 
+\\[-1.8ex] & formula.7.means & formula.7.justified & formula.7.message & formula.7.prevent & formula.7.condemn \\ 
+\\[-1.8ex] & (1) & (2) & (3) & (4) & (5)\\ 
+\hline \\[-1.8ex] 
+ Mate Competition & 0.155$^{***}$ & 0.173$^{***}$ & 0.192$^{***}$ & 0.204$^{***}$ & 0.027 \\ 
+  & (0.019) & (0.019) & (0.019) & (0.018) & (0.021) \\ 
+  & & & & & \\ 
+ Job Competition & 0.056$^{***}$ & 0.050$^{***}$ & 0.097$^{***}$ & 0.087$^{***}$ & 0.017 \\ 
+  & (0.019) & (0.019) & (0.019) & (0.018) & (0.021) \\ 
+  & & & & & \\ 
+ Life Satisfaction & $-$0.0001 & $-$0.009 & $-$0.004 & $-$0.011$^{*}$ & $-$0.007 \\ 
+  & (0.006) & (0.006) & (0.006) & (0.006) & (0.007) \\ 
+  & & & & & \\ 
+\hline \\[-1.8ex] 
+Observations & 3,008 & 3,008 & 3,008 & 3,008 & 3,008 \\ 
+R$^{2}$ & 0.459 & 0.453 & 0.448 & 0.469 & 0.347 \\ 
+Adjusted R$^{2}$ & 0.431 & 0.424 & 0.419 & 0.441 & 0.313 \\ 
+Residual Std. Error (df = 2857) & 0.670 & 0.683 & 0.679 & 0.643 & 0.745 \\ 
+F Statistic (df = 150; 2857) & 16.170$^{***}$ & 15.746$^{***}$ & 15.462$^{***}$ & 16.814$^{***}$ & 10.136$^{***}$ \\ 
+\hline 
+\hline \\[-1.8ex] 
+\textit{Note:}  & \multicolumn{5}{r}{$^{*}$p$<$0.1; $^{**}$p$<$0.05; $^{***}$p$<$0.01} \\ 
+\end{tabular} 
+\end{table}