Dear Statalisters,
I'm using the user estimation command
with the fuzzy option, and I obtain significant first-stage estimates and significant treatment effect estimates, but these last estimates are out of range. In particular, both outcomes are binary: in the first stage, the outcome is one if a student becomes a teen parent, and in the second stage, the outcome is one if the student's mother receives any type of welfare. The running variable is the date of birth. I have run the same specification with and without control variables, different kernels, and bandwidth, and the results remain out of range. I think it has to be related to the fact that the first stage estimates are very small, and that might be "inflating" the results. If that is the case, how do I interpret an out-of-range result?
I'm not able to paste the results directly, but I have a picture of the stata output and a picture of the table with relevant information (such as averages of welfare to the left and right of the cutoff)
First: stata output of one of the regressions:
And here is the table with the other regressions and other relevant information such as averages to the left and right of the cutoff
Thank so you much for any help you can provide, and apologizes for the format of both output and tables.
I'm using the user estimation command
Code:
rdrobust
I'm not able to paste the results directly, but I have a picture of the stata output and a picture of the table with relevant information (such as averages of welfare to the left and right of the cutoff)
First: stata output of one of the regressions:
Code:
rdrobust welfare ru, fuzzy(parent) all p(1) vce(cluster ru)
And here is the table with the other regressions and other relevant information such as averages to the left and right of the cutoff
Thank so you much for any help you can provide, and apologizes for the format of both output and tables.
