Hi I’m using Stata/SE 15.1 and am able to run all example do files included in the rdplot/rdrobust/rdbwselect and rd on my computer.
My research project involves assessing the impact of a program which is age restricted (25 years old is my cutoff and my running variable is integer values for age 18-55). My dependent variables are a binary and a 0-12 integer values (nothing close to a normal distribution). While I’m interested in assessing discontinuity at the cutoff for program access, many other factors can have an impact on the outcome. My basic logistic regression model includes binary program access and multiple binary and categorical variables (I understand from trying that factor variables are not allowed in rdrobust, which is an issue for me).
While -rdplot- provides me with plausible visual results, -rdrobust- (without covariates) always sends me:
rdrobust() not able to compute the loc. poly. bandwidth (h) above the threshold.
> Please run rdbwselect() for more information}
r(125);
And
-rdbwselect- returns
Invertibility problem in the computation of preliminary bandwidth below the thre
> sholdInvertibility problem in the computation of bias bandwidth (b) below the
> thresholdInvertibility problem in the computation of bias bandwidth (b) above
> the thresholdInvertibility problem in the computation of loc. poly. bandwidth
> (h) below the thresholdInvertibility problem in the computation of loc. poly.
> bandwidth (h) above the threshold
I’ve tried with other continuous variables in my dataset as dependent variables and the programs are running fine and providing something that fits with what I can see from descriptive analysis.
So my questions are:
Can I use RD (and more specifically these packages) with a binary outcome?
I can’t find anything in the documentation to assess this invertibility problem?
Can I use an integer variable as a running variable?
How would I go about integrating binary and categorical variables as covariates?
Any help you be greatly appreciated.
My research project involves assessing the impact of a program which is age restricted (25 years old is my cutoff and my running variable is integer values for age 18-55). My dependent variables are a binary and a 0-12 integer values (nothing close to a normal distribution). While I’m interested in assessing discontinuity at the cutoff for program access, many other factors can have an impact on the outcome. My basic logistic regression model includes binary program access and multiple binary and categorical variables (I understand from trying that factor variables are not allowed in rdrobust, which is an issue for me).
While -rdplot- provides me with plausible visual results, -rdrobust- (without covariates) always sends me:
rdrobust() not able to compute the loc. poly. bandwidth (h) above the threshold.
> Please run rdbwselect() for more information}
r(125);
And
-rdbwselect- returns
Invertibility problem in the computation of preliminary bandwidth below the thre
> sholdInvertibility problem in the computation of bias bandwidth (b) below the
> thresholdInvertibility problem in the computation of bias bandwidth (b) above
> the thresholdInvertibility problem in the computation of loc. poly. bandwidth
> (h) below the thresholdInvertibility problem in the computation of loc. poly.
> bandwidth (h) above the threshold
I’ve tried with other continuous variables in my dataset as dependent variables and the programs are running fine and providing something that fits with what I can see from descriptive analysis.
So my questions are:
Can I use RD (and more specifically these packages) with a binary outcome?
I can’t find anything in the documentation to assess this invertibility problem?
Can I use an integer variable as a running variable?
How would I go about integrating binary and categorical variables as covariates?
Any help you be greatly appreciated.
