rdplot is giving a different result from rdrobust?

[badcode goodcoffee]

I am running a fuzzy regression discontinuity design - the direction of the coefficients I am getting from OLS, IV, and rdplot are positive and rdrobust is negative. I don't understand why the direction of my coefficients is changing.

[Sofia Hernandez]

Hi, please provide the output from rdrobust and rdplot so we can help you with this

[badcode goodcoffee]

The code I wrote is

reghdfe dep_var indep_var $controls2 [aweight = weight], absorb(state) vce(cluster district_num)

ivreghdfe dep_var (indep_var =treatment) $controls2 [aw = weight], absorb(state) vce(cluster district_num)

rdrobust dep_var indep_var , p(1) masspoints(adjust) vce(cluster district_num) kernel(uni) c(0) weights(weight) fuzzy(fuzzy_score) covs($controls2)

The result is the following (I omitted the controls part):

My question is why is the direction of the coefficient of IV and OLS different from rdd.