The commands "predict resid, residuals" and "estat dwatson" work fine in case of on OLS regression but give error when I try to check the residuals in case of a probit model. Can someone please shed some light on this? Thanks!
Hello Dipen,
What statistical software are you using (and which specific function or subprogram)?
I am using STATA. The command I had put in was -
"predict resid, residuals".
It works fine if I run an OLS regression but won't work on a Probit command. When I looked up online, I figured that some researchers are writing about Deviance Residuals. However, I'm not convinced about using them interchangeably.
Hello Dipen,
This link seemingly explains how to elicit predicted values and residuals in probit regression using STATA (specifically, "predict"); D-W would be pretty straightforward from there: https://www.stata.com/manuals/rprobitpostestimation.pdf
If you think that your current attempts don't jibe with the manual's explanation for available probit postestimation options, you might try contacting stata for help (the contact for which I believe still shows on stata splash/start-up screen).
Good luck with your work.
This pdf is referring to the Deviance Residuals.
Thanks for the suggestion! I'll contact STATA right away.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size