[R] Covariates in fuzzy RDD with rddtools

Chiara Malavasi chiara.malavasi at live.it
Fri Feb 9 16:36:35 CET 2018


Hello!
I am having trouble including covariates in a fuzzy RDD model in R with rddtools. I run the following commands:

    library(rddtools)

    data.complete2 <- as.data.frame(cbind(GDPpc.rel, GR.rate, fitval.firstStage)) colnames(data.complete2) <- c("GDPpc.rel", "GR.rate", "FitTreat")

    data.RDD<-rdd_data(x=data.complete2$GDPpc.rel, y=data.complete2$GR.rate, z=data.complete2$FitTreat, covar=dummytime, cutpoint = 0.75)
    data.RDD1<-rdd_data(x=data.complete2$GDPpc.rel, y=data.complete2$GR.rate, z=data.complete2$FitTreat, covar=cbind(dummytime, dummyreg), cutpoint = 0.75)

    reg_para <- rdd_reg_lm(rdd_object=data.RDD, covar.opt = list(strategy= "include"), order = 4)
    reg_para1 <- rdd_reg_lm(rdd_object=data.RDD1,  covar.opt = list(strategy = "include"), order = 4)

where dummytime and dummyreg are two dummy variables considering time and regional effects.

The problem is that looking at the summary of the two regressions, they seem to be exactly the same and the coefficient which should refer to the dummy variables is not reported. My guess is that I did not actually implement the covariates in the regression, but I cannot understand where is the problem. Any suggestion?

Thank you in advance.

Best regards,

Chiara Malavasi


	[[alternative HTML version deleted]]



More information about the R-help mailing list