phd=read.table("phd_data.csv",sep=",",header=T)

phd$global_clustering_no_ego_2=phd$`global clustering without ego_2`

phd$constraint_2=phd$`constraint of ego_2`
phd$constraint_4=phd$`constraint of ego_4`
phd$constraint_6=phd$`constraint of ego_6`

hist(log(phd$degree_0))

nrow(phd[])

hist(log(phd$betweenness_0))

m1=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        #log(degree_2+1) + 
        #log(global_clustering_no_ego_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011),])
#summary(m1)

m2=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        log(degree_2+1) + 
        #log(global_clustering_no_ego_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011),])
#summary(m2)

m3=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        log(degree_2+1) + 
        log(constraint_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011),])
#summary(m3)

cor_d=phd[which(phd$def_year<2011),c("citations_cum_8", "citations_cum_2",
                                     "papers_cum_2", "degree_2", "constraint_2")]

cor(cor_d, use="pairwise.complete.obs", method="pearson")

length(unique(phd$discipline))

#library(stargazer)
capture.output(stargazer(m1, m2, m3, type="text", 
                         dep.var.labels="Citations 8 years post-defence",
                         column.labels = c("M1","M2", "M3"),
                                                #  covariate.labels=c("CIT_2","PAPERS_2", "PAPERS_8_2"), 
                                                                     # "DEG_0 ? CON_0",
                                                                     # "PAPERS_0", "PAPERS_8",
                                                                     # "DEG_0_2", "DEG_2_4", "DEG_4_6",
                                                                     # "CON_0_2", "CON_2_4", "CON_4_6"),
                         omit.stat=c("LL","ser","f"), ci=F,  single.row=F,
                         out="dec14_1.txt"))


###############################################xx


# non-live science
m4=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        log(degree_2+1) + 
        log(constraint_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011 & phd$Research_Area=="non-live science"),])

# live science
m5=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        log(degree_2+1) + 
        log(constraint_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011 & phd$Research_Area=="live science"),])

# engineering
m6=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        log(degree_2+1) + 
        log(constraint_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011 & phd$Research_Area=="engineering"),])

# social science
m7=lm(log(citations_cum_8+1) ~ 
        log(citations_cum_2+1) + 
        log(papers_cum_2+1) + 
        log(papers_cum_8-papers_cum_2+1) + 
        log(degree_2+1) + 
        log(constraint_2+1) +
        discipline +  as.factor(def_year)
      , data=phd[which(phd$def_year<2011 & phd$Research_Area=="social science"),])


capture.output(stargazer(m4, m5, m6, m7, type="text", 
                         dep.var.labels="Citations 8 years post-defence",
                         column.labels = c("Sciences","Life Sci", "Engineer", "Social Sci"),
                         #                         covariate.labels=c("HIV_0","DEG_0", "CON_0", 
                         #                                            "DEG_0 ? CON_0",
                         #                                            "PAPERS_0", "PAPERS_8",
                         #                                            "DEG_0_2", "DEG_2_4", "DEG_4_6",
                         #                                            "CON_0_2", "CON_2_4", "CON_4_6"),
                         omit.stat=c("LL","ser","f"), ci=F,  single.row=F,
                         out="dec14_2.txt"))