Problem with rdrobust: control on the right of the cutoff, treatment on the left of the cutoff

[Paula Pereda]

I want to replicate an exercise from Handbook on Impact Evaluation: Quantitative Methods and Practices that uses data from Bangladesh's households in 1998 (hh_98.dta).

Impact assessment can be implemented with RD design using the Bangladesh data because participation in microcredit programs is officially determined by a household’s landholding; that is, a household is eligible to participate only if it has fewer than 50 decimals of land. Therefore, the cutoff point of 50 decimals in land assets fulfills the RD design criterion. Some eligible households (land asset < 50 decimals) do not participate, and some ineligible households (land asset >= 50 decimals) do participate.

The problem here is that I want to replicate the exercise of a Fuzzy RD with rdrobust but the control on the right of the cutoff, treatment on the left of the cutoff so I think I am estimating in the wrong idea as the default is left, control and right, treatment.

Any suggestions on how to deal with this?

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

This sounds like a class assignment. We don't do homework for folks.

With a user written procedure, help depends on someone using that specific procedure being active on the list. Apparently, no one is.

[Liu Qiang]

hi Paula, I think for your question, you can replace x=-x+50.