kenya_dat_long<-kenya_dat %>% group_by(year) %>% 
  summarize(percent_fridge=mean(fridge_stat,na.rm=T),percent_phone=mean(phone,na.rm=T),percent_radio=mean(radio_stat,na.rm=T),percent_owns_car=mean(car_stat,na.rm=T),percent_tv=mean(tv_stat,na.rm=T),percent_bike=mean(bike_stat,na.rm=T),percent_motorcycle=mean(motorcycle_stat,na.rm=T),percent_newspaper=mean(literacy,na.rm=T))
kenya_dat_long<-data.frame(kenya_dat_long)
kenya_modern<-melt(kenya_dat_long,id="year")
ggplot(data=kenya_modern,aes(x=year,y=value,col=variable))+
  geom_line(size=2)+
  xlab("DHS Study Year")+
  ylab("Percent")+
  ggtitle("Modernity Related Descriptive Statistics",subtitle = "Kenya Contraception Project - Spring 2018")
## Warning: Removed 6 rows containing missing values (geom_path).

kenya_dat_long2<-kenya_dat %>% group_by(year) %>% 
  summarize(percent_never_married=mean(never_married,na.rm=T),percent_more1_union=mean(more_than1_union,na.rm=T),percent_husband_home=mean(husband_home,na.rm=T),percent_desires_kids=mean(kid_desire,na.rm=T),percent_female_working=mean(femwork,na.rm=T))
kenya_dat_long2<-data.frame(kenya_dat_long2)
kenya_demo<-melt(kenya_dat_long2,id="year")
ggplot(data=kenya_demo,aes(x=year,y=value,col=variable))+
  geom_line(size=2)+
  xlab("DHS Study Year")+
  ylab("Percent")+
  ggtitle("Demographic Related Descriptive Statistics",subtitle = "Kenya Contraception Project - Spring 2018")

kenya_dat_long3<-kenya_dat %>% group_by(year) %>% 
  summarize(mean_children_ever=mean(cheb,na.rm=T),mean_sons_home=mean(sonsathome,na.rm=T),mean_daus_home=mean(dausathome,na.rm=T),mean_sons_died=mean(sonsdied,na.rm=T),mean_daus_died=mean(dausdied,na.rm=T))
kenya_dat_long3<-data.frame(kenya_dat_long3)
kenya_children<-melt(kenya_dat_long3,id="year")
ggplot(data=kenya_children,aes(x=year,y=value,col=variable))+
  geom_line(size=2)+
  xlab("DHS Study Year")+
  ylab("Percent")+
  ggtitle("Children Related Descriptive Statistics",subtitle = "Kenya Contraception Project - Spring 2018")

kenya_dat_long4<-kenya_dat %>% group_by(year) %>% 
  summarize(mean_urban=mean(urban_stat,na.rm=T),mean_mig_rural_urb=mean(mig_rurl_urb,na.rm=T),mean_mig_urb_rural=mean(mig_urb_rurl,na.rm=T))
kenya_dat_long4<-data.frame(kenya_dat_long4)
kenya_urban_migration<-melt(kenya_dat_long4,id="year")
ggplot(data=kenya_urban_migration,aes(x=year,y=value,col=variable))+
  geom_line(size=2)+
  xlab("DHS Study Year")+
  ylab("Percent")+
  ggtitle("Urban / Migration Related Descriptive Statistics",subtitle = "Kenya Contraception Project - Spring 2018")

#urban only model
options(survey.lonely.psu = "adjust")
des1<-svydesign(ids=~caseid,strata=~strata,weights=~perweight,data=kenya_dat[!is.na(kenya_dat$strata),],nest=T)
fit1<-svyglm(contracept~urban_stat,design=des1,family=binomial)
## Warning in eval(family$initialize): non-integer #successes in a binomial
## glm!
#urban + demographic model
options(survey.lonely.psu = "adjust")
fit2<-svyglm(contracept~urban_stat+year+more_than1_union+husband_home+kid_desire+femwork,design=des1,family=binomial)
## Warning in eval(family$initialize): non-integer #successes in a binomial
## glm!
#urban + demographic + children model
options(survey.lonely.psu = "adjust")
fit3<-svyglm(contracept~urban_stat+year+more_than1_union+husband_home+kid_desire+femwork+cheb+sonsathome+dausathome+sonsdied+dausdied,design=des1,family=binomial)
## Warning in eval(family$initialize): non-integer #successes in a binomial
## glm!
#urban + demographic + children + modernity model
options(survey.lonely.psu = "adjust")
fit4<-svyglm(contracept~urban_stat+year+more_than1_union+husband_home+kid_desire+femwork+cheb+sonsathome+dausathome+sonsdied+dausdied+fridge_stat+phone+tv_stat+radio_stat+newspaper+car_stat+bike_stat+motorcycle_stat+literacy,design=des1,family=binomial)
## Warning in eval(family$initialize): non-integer #successes in a binomial
## glm!
#urban + demographic + children + modernity + migratory model
options(survey.lonely.psu = "adjust")
fit5<-svyglm(contracept~urban_stat+year+more_than1_union+husband_home+kid_desire+femwork+cheb+sonsathome+dausathome+sonsdied+dausdied+fridge_stat+phone+tv_stat+radio_stat+newspaper+car_stat+bike_stat+motorcycle_stat+literacy+mig_rurl_urb+mig_urb_rurl,design=des1,family=binomial)
## Warning in eval(family$initialize): non-integer #successes in a binomial
## glm!
stargazer(fit1,fit2,fit3,fit4,fit5,type="html")
Dependent variable:
contracept
(1) (2) (3) (4) (5)
urban_stat 0.397*** 0.554*** 0.468*** 0.151*** 0.136**
(0.025) (0.040) (0.042) (0.057) (0.059)
year 0.057*** 0.055*** 0.088*** 0.091***
(0.002) (0.002) (0.005) (0.005)
more_than1_union -0.528*** -0.361*** -0.159* -0.157*
(0.073) (0.075) (0.094) (0.094)
husband_home 0.105*** 0.104** 0.109** 0.113**
(0.040) (0.041) (0.053) (0.053)
kid_desire -0.476*** -0.649*** -0.475*** -0.474***
(0.035) (0.040) (0.052) (0.052)
femwork 0.443*** 0.476*** 0.423*** 0.422***
(0.036) (0.037) (0.049) (0.048)
cheb -0.170*** -0.169*** -0.169***
(0.017) (0.023) (0.023)
sonsathome 0.171*** 0.218*** 0.219***
(0.021) (0.029) (0.029)
dausathome 0.136*** 0.154*** 0.156***
(0.021) (0.029) (0.029)
sonsdied -0.183*** -0.110* -0.110*
(0.045) (0.058) (0.058)
dausdied -0.154*** -0.032 -0.032
(0.048) (0.062) (0.062)
fridge_stat -0.365*** -0.355***
(0.117) (0.118)
phone 0.553*** 0.551***
(0.107) (0.108)
tv_stat 0.363*** 0.362***
(0.057) (0.057)
radio_stat 0.343*** 0.341***
(0.057) (0.056)
newspaper 0.098 0.094
(0.070) (0.070)
car_stat 0.105 0.101
(0.111) (0.111)
bike_stat 0.030 0.029
(0.048) (0.048)
motorcycle_stat 0.005 0.003
(0.092) (0.093)
literacy 0.756*** 0.753***
(0.064) (0.064)
mig_rurl_urb 0.135
(0.139)
mig_urb_rurl 0.119
(0.091)
Constant -0.915*** -114.559*** -110.114*** -177.690*** -184.311***
(0.012) (4.433) (4.475) (10.004) (10.358)
Observations 63,139 26,326 26,326 17,294 17,294
Log Likelihood -37,999.030 -15,499.080 -15,241.040 -10,278.400 -10,276.560
Akaike Inf. Crit. 76,002.070 31,012.150 30,506.080 20,598.810 20,599.130
Note: p<0.1; p<0.05; p<0.01