library(foreign)
x <- read.dta("http://www-personal.umich.edu/~jowei/fema/Chen.Individual.Data.dta")


############################### TABLE 1 ###############################
##Table 1, Model 1:
summary(glm(Turnout2004~Turnout2002+Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="DEM" & x$date<=41104,]))

##Table 1, Model 2:
summary(glm(Turnout2004~Turnout2002+Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="REP" & x$date<=41104,]))

##Table 1, Model 3:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+I(Aid_awarded*(Party=="REP"))+I(Party=="REP"), family=binomial(link='logit'), data=x[(x$Party=="REP"|x$Party=="DEM") & x$date<=41104,]))

##Table 1, Model 4:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+mph+mphSQUARED+age+ageSQUARED+medhomevalue10k+(Gender=="M")+black+medincome10k+as.factor(CountyCode), family=binomial(link='logit'), data=x[x$Party=="DEM" & x$date<=41104,]))

##Table 1, Model 5:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+mph+mphSQUARED+age+ageSQUARED+medhomevalue10k+(Gender=="M")+black+medincome10k+as.factor(CountyCode), family=binomial(link='logit'), data=x[x$Party=="REP" & x$date<=41104,]))

##Table 1, Model 6:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+I(Aid_awarded*(Party=="REP"))+I(Party=="REP")+mph+mphSQUARED+age+ageSQUARED+medhomevalue10k+(Gender=="M")+black+medincome10k+as.factor(CountyCode), family=binomial(link='logit'), data=x[(x$Party=="REP"|x$Party=="DEM") & x$date<=41104,]))



############################### TABLE 2 ###############################
##Table 2, Model 1:
summary(glm(Turnout2004~Turnout2002+Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="DEM" & x$date>41104,]))

##Table 2, Model 2:
summary(glm(Turnout2004~Turnout2002+Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="REP" & x$date>41104,]))

##Table 2, Model 3:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+I(Aid_awarded*(Party=="REP"))+I(Party=="REP"), family=binomial(link='logit'), data=x[(x$Party=="REP"|x$Party=="DEM") & x$date>41104,]))

##Table 2, Model 4:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+mph+mphSQUARED+age+ageSQUARED+medhomevalue10k+(Gender=="M")+black+medincome10k+as.factor(CountyCode), family=binomial(link='logit'), data=x[x$Party=="DEM" & x$date>41104,]))

##Table 2, Model 5:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+mph+mphSQUARED+age+ageSQUARED+medhomevalue10k+(Gender=="M")+black+medincome10k+as.factor(CountyCode), family=binomial(link='logit'), data=x[x$Party=="REP" & x$date>41104,]))

##Table 2, Model 6:
summary(glm(Turnout2004~Turnout2002+Aid_awarded+I(Aid_awarded*(Party=="REP"))+I(Party=="REP")+mph+mphSQUARED+age+ageSQUARED+medhomevalue10k+(Gender=="M")+black+medincome10k+as.factor(CountyCode), family=binomial(link='logit'), data=x[(x$Party=="REP"|x$Party=="DEM") & x$date>41104,]))



############################### TABLE 4 ###############################
##Table 4, Model 1:
summary(glm(Turnout2004~Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="DEM" & x$date<=41104 & x$Turnout2002==1,]))

##Table 4, Model 2:
summary(glm(Turnout2004~Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="REP" & x$date<=41104 & x$Turnout2002==1,]))

##Table 4, Model 3:
summary(glm(Turnout2004~Aid_awarded+I(Aid_awarded*(Party=="REP"))+I(Party=="REP"), family=binomial(link='logit'), data=x[(x$Party=="REP"|x$Party=="DEM") & x$date<=41104 & x$Turnout2002==1,]))

##Table 4, Model 4:
summary(glm(Turnout2004~Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="DEM" & x$date<=41104 & x$Turnout2002==0,]))

##Table 4, Model 5:
summary(glm(Turnout2004~Aid_awarded, family=binomial(link='logit'), data=x[x$Party=="REP" & x$date<=41104 & x$Turnout2002==0,]))

##Table 4, Model 6:
summary(glm(Turnout2004~Aid_awarded+I(Aid_awarded*(Party=="REP"))+I(Party=="REP"), family=binomial(link='logit'), data=x[(x$Party=="REP"|x$Party=="DEM") & x$date<=41104 & x$Turnout2002==0,]))




#####################Precinct results

x <- read.table("http://www-personal.umich.edu/~jowei/fema/Chen.Precinct.Data.txt", header=T)

############################### TABLE 5 ###############################
##Table 5, Model 1:
summary(lm(100*bushshare~jebshare+bush00share+I(log.total.dollars.pc) , weights=TVAP00 ,data=x))

##Table 5, Model 2:
summary(lm(100*bushshare~jebshare+bush00share+I(log.total.dollars.pc)+cMAXSFC_MPH  +fMAXSFC_MPH  +iMAXSFC_MPH  +jMAXSFC_MPH , weights=TVAP00 ,data=x))

##Table 5, Model 3:
summary(lm(100*bushshare~jebshare+bush00share+I(log.total.dollars.pc)+cMAXSFC_MPH  +fMAXSFC_MPH  +iMAXSFC_MPH  +jMAXSFC_MPH +I(medhouseholdincome/10000)+I(welfarepc/1000)+blackprop+owner , weights=TVAP00 ,data=x))

##Table 5, Model 4:
summary(lm(100*bushshare~jebshare+bush00share+I(log.total.dollars.pc)+cMAXSFC_MPH  +fMAXSFC_MPH  +iMAXSFC_MPH  +jMAXSFC_MPH +I(medhouseholdincome/10000)+I(welfarepc/1000)+blackprop+owner, weights=TVAP00 ,data=x[x$bush00share<=0.5,]))

##Table 5, Model 5:
summary(lm(100*bushshare~jebshare+bush00share+I(log.total.dollars.pc)+cMAXSFC_MPH  +fMAXSFC_MPH  +iMAXSFC_MPH  +jMAXSFC_MPH +I(medhouseholdincome/10000)+I(welfarepc/1000)+blackprop+owner , weights=TVAP00 ,data=x[x$bush00share>0.5,]))