Dear All,
For my master thesis, I am applying quantile regression and came across an issue I am struggling to solve.
I am currently using the qreg command for my estimations. Ideally, I wanted to use bootstrapped SEs (with the command bsqreg), however, this does not allow me to apply sample weights while qreg does.
To graph the results, I use grqreg but the command seems to use other standard errors as the qreg estimation, even though I run them together (the coefficient of a variable is significantly different from zero for some quantiles in the graph but not in the qreg estimations).
Thus, I have the following questions:
Best,
Valentina
For my master thesis, I am applying quantile regression and came across an issue I am struggling to solve.
I am currently using the qreg command for my estimations. Ideally, I wanted to use bootstrapped SEs (with the command bsqreg), however, this does not allow me to apply sample weights while qreg does.
To graph the results, I use grqreg but the command seems to use other standard errors as the qreg estimation, even though I run them together (the coefficient of a variable is significantly different from zero for some quantiles in the graph but not in the qreg estimations).
Thus, I have the following questions:
- Do you have an idea where the discrepancy between the graph and quantile regression table output could stem from and how I could try to solve the issue?
- Do you know whether I could somehow apply sample weights when using quantile regression with bootstrapped standard errors (bsqreg)?
Best,
Valentina
