|
|
|
|
|
|
|
|
|
|
|
|
|
|
| require(lme4)
|
| require(lfe)
|
| require(MatchIt)
|
| require(WeightIt)
|
| require(tjbal)
|
| require(optmatch)
|
| require(stargazer)
|
| require(cobalt)
|
| require(tidyverse)
|
| require(ggridges)
|
|
|
| rm(list = ls())
|
| gc()
|
|
|
| load('../Data/replication_data.RData')
|
|
|
|
|
|
|
| source('./helper_functions.R')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| D <- unlist(lapply(c("sc_","ln_"),function(x) paste0(x,paste0(c("county_","DMA_","state_"),"cases"))))
|
| FE <- c("0","DMA_CODE","date")
|
| covariates <- c("sc_CTY_LTHS","sc_CTY_CollUp",
|
| "sc_CTY_LT30yo","sc_CTY_60Up",
|
| "sc_CTY_Below_poverty_level_AGE_18_64","sc_CTY_Female_hher_no_husbandhh",
|
| "sc_CTY_Unem_rate_pop_16_over","sc_CTY_Labor_Force_Part_Rate_pop_16_over",
|
| "sc_CTY_Manufactur","sc_CTY_Md_inc_hhs",
|
| "sc_CTY_POPPCT_RURAL","sc_CTY_Speak_only_English","sc_CTY_White","sc_CTY_Black_or_African_American",
|
| "ln_CTY_tot_pop","sc_turnout_pct_20")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| weeks <- c("2020-02-22","2020-02-29",
|
| "2020-03-03","2020-03-10","2020-03-17","2020-04-07")
|
|
|
| resWkly <- list()
|
| for(y in c("pcttw_sanders")) {
|
| for(d in c("DMA_March10Cases",
|
| "DMA_March17Cases")) {
|
| for(fe in "0") {
|
|
|
| wkly <- wklyBin <- wklyMatch <- wklyWeight <- wklyMatchBin <- wklyWeightBin <- NULL
|
| if(grepl("March",d)) {
|
| wks <- weeks
|
| } else {
|
| wks <- weeks[4:6]
|
| }
|
| for(week in wks) {
|
| if(fe == 'stab' & week == '2020-04-07') { next }
|
| cat(paste0(week,'/',y,'/',d,'/',fe,'\n'))
|
| if(grepl("county",d) & week == "2020-03-01") { next }
|
| if(week < as.Date("2020-03-01")) {
|
| tmp <- finalDat %>% select(y,matches(gsub("sc_|ln_","",d)),DMA_CODE,stab,date,c(covariates,paste0(y,"16"))) %>% filter(date < as.Date(week)) %>% filter(complete.cases(.))
|
| } else {
|
| tmp <- finalDat %>% select(y,matches(gsub("sc_|ln_","",d)),DMA_CODE,stab,date,c(covariates,paste0(y,"16"))) %>% filter(date == as.Date(week)) %>% filter(complete.cases(.))
|
| }
|
| tmp$treatBin <- ifelse(tmp[[gsub("sc_|ln_","",d)]] > 0,1,0)
|
|
|
| wkly <- bind_rows(wkly,data.frame(t(summary(felm(as.formula(paste0("scale(",y,") ~ scale(",gsub("sc_","",d),") + ",
|
| paste(unique(c(covariates,paste0(y,"16"))),collapse = "+"),
|
| "| ",fe," | 0 | 0")),tmp))$coefficients[paste0("scale(",gsub("sc_","",d),")"),]),date = week))
|
|
|
| wklyBin <- bind_rows(wklyBin,data.frame(t(summary(felm(as.formula(paste0("scale(",y,") ~ treatBin + ",
|
| paste(unique(c(covariates,paste0(y,"16"))),collapse = "+"),
|
| "| ",fe," | 0 | 0")),tmp))$coefficients["treatBin",]),date = week))
|
|
|
| if(length(unique(tmp$treatBin)) == 2) {
|
| m.out <- try(matchit(formula = as.formula(paste("treatBin ~ ",paste(c(covariates,paste0(y,"16")),collapse = " + "))),
|
| data = tmp,
|
| method = "nearest",
|
| distance = "mahalanobis"))
|
| if(class(m.out) != "try-error") {
|
| m.data <- match.data(m.out)
|
| wklyMatch <- bind_rows(wklyMatch,data.frame(t(summary(felm(as.formula(paste0("scale(",y,") ~ scale(",gsub("sc_","",d),") + ",
|
| paste(c(covariates,paste0(y,"16")),collapse = " + ")," | ",fe," | 0 | ",fe)),
|
| m.data,weights = m.data$weights))$coefficients[paste0("scale(",gsub("sc_","",d),")"),]),date = week))
|
| wklyMatchBin <- bind_rows(wklyMatchBin,data.frame(t(summary(felm(as.formula(paste0("scale(",y,") ~ treatBin + ",
|
| paste(c(covariates,paste0(y,"16")),collapse = " + ")," | ",fe," | 0 | ",fe)),
|
| m.data,weights = m.data$weights))$coefficients["treatBin",]),date = week))
|
| }
|
|
|
| W.out <- try(weightit(formula(paste0("treatBin ~ ",paste(c(covariates,paste0(y,"16")),collapse = " + "))),
|
| data = tmp, estimand = "ATT", method = "cbps"))
|
|
|
|
|
| if(class(W.out) != "try_error") {
|
| wklyWeight <- bind_rows(wklyWeight,data.frame(t(summary(felm(as.formula(paste0("scale(",y,") ~ scale(",gsub("sc_","",d),") + ",
|
| paste(c(covariates,paste0(y,"16")),collapse = " + ")," | ",fe," | 0 | ",fe)),
|
| tmp,weights = W.out$weights))$coefficients[paste0("scale(",gsub("sc_","",d),")"),]),date = week))
|
|
|
| wklyWeightBin <- bind_rows(wklyWeightBin,data.frame(t(summary(felm(as.formula(paste0("scale(",y,") ~ treatBin + ",
|
| paste(c(covariates,paste0(y,"16")),collapse = " + ")," | ",fe," | 0 | ",fe)),
|
| tmp,weights = W.out$weights))$coefficients["treatBin",]),date = week))
|
| }
|
| }
|
| }
|
| resWkly[[y]][[d]][[fe]]$basic$cont <- wkly
|
| resWkly[[y]][[d]][[fe]]$basic$bin <- wklyBin
|
| resWkly[[y]][[d]][[fe]]$matching$cont <- wklyMatch
|
| resWkly[[y]][[d]][[fe]]$matching$bin <- wklyMatchBin
|
| resWkly[[y]][[d]][[fe]]$weighting$cont <- wklyWeight
|
| resWkly[[y]][[d]][[fe]]$weighting$bin <- wklyWeightBin
|
| }
|
| }
|
| }
|
|
|
|
|
| pdf('../Figures/figure5.pdf',width = 7,height = 5)
|
| bind_rows(resWkly$pcttw_sanders$DMA_March17Cases$`0`$basic$cont %>%
|
| mutate(type = 'basic',
|
| date = as.Date(date) - .25),
|
| resWkly$pcttw_sanders$DMA_March17Cases$`0`$weighting$cont %>%
|
| mutate(type = 'weighting',
|
| date = as.Date(date) + .25)) %>%
|
| ggplot(aes(x = date,y = Estimate,color = type,fill = type)) +
|
| geom_hline(yintercept = 0,linetype = 'dashed') +
|
| geom_point(size = 1.5) +
|
| geom_rect(aes(xmin = date-.25,ymin = Estimate - 1.96*`Std..Error`,
|
| xmax = date+.25,ymax = Estimate + 1.96*`Std..Error`),alpha = .5,color = NA) +
|
| theme_ridges() +
|
| scale_color_manual(values = c('basic' = 'grey50','weighting' = 'grey40'),
|
| guide = F) +
|
| scale_fill_manual(name = 'Method',values = c('basic' = 'grey70','weighting' = 'grey40'),
|
| labels = c('basic' = 'Unweighted OLS','weighting' = 'CBPS Weights')) +
|
| xlab('Date') + ylab('Coefficient on March 17 Cases') +
|
| theme(legend.position = 'bottom')
|
| dev.off() |
