| Re: Weighting Combined Individual Data for Logistic Regression Analysis [message #26509 is a reply to message #3723] |
Mon, 27 March 2023 22:36  |
Alanood
Messages: 18
Registered: November 2022
|
Member |
|
|
Hello
Regarding your reply, " In general, the most conservative thing to do would be to report both weighted and unweighted estimates. They really shouldn't vary too much - if they do, there is probably something weird going on with either your model or your basic assumptions (and their relationship with reality)."
I have a difficulty with my Instrument variable method results (weighted and unweighted) , the sata command for weigted is as follow
ivregress 2sls ....(outcome variable) (controls) (fixed effect) ......... (endogenous variable = instrument)[pweight = sampwt], vce(cluster v021) first
unweighted
ivregress 2sls ....(outcome variable) (controls) (fixed effect) ......... (endogenous variable = instrument), robust first
1- the results are differed between the weighted and unweighted, in the unweighted option: the main independent variable is significant at 3 out of 4 regressions outcome variables , where in the weighted option , i only got one outcome variable out of 4 regressions is significant, what do you think the reason?
2- how we can include the strata option in the ivregress command ? "svy" is not alowed with "ivreg" and i tried to include "strata" in the regression but I got an error.
Thank you in advanced
[Updated on: Mon, 27 March 2023 22:39] Report message to a moderator |
|
|
|
The DHS Program User Forum
Members
Search
Help
Register
Login
Home
