Test regression coefficients from two different models

[rfrancis]

Hi everyone, I'm working with a dataset that is cross-sectional and longitudinal in nature, which creates cluster issues for standard errors (clustered in subjects and clustered in time). The goal is to test for differences between regression coefficients from two different models. So, for example, say Model 1 is Y = X1 + X2, and Model 2 is Y' = X1' + X2'. Here the goal is a test of equality between the coefficients for X1 and X1' ... and between X2 and X2'. PROC SYSLIN seems like the perfect vehicle for this scenario. But the clustered data creates a problem for the standard errors used in SYSLIN.  Guess two-way SEs for SYSLIN is a solution. I have access to two-way SEs using SURVEYREG, So, guess I could use these two-way SEs but need a bridge to incorporate them into SYSLIN. And I'm ALL ears for other approaches. I am grateful for any ideas.  Thank you!  Rick

[ChrisNZ]

Moved post to stats community and fixed title.

[PaigeMiller]

There was extensive discussion about this matter in your other thread at https://communities.sas.com/t5/Statistical-Procedures/Compare-regression-coefficients-from-two-model.... Wasn't that discussion sufficient? What about this problem that you just posted is different than that problem?

[rfrancis]

Hi Paige, the first post is in the "SAS Programming" forum, which was probably not the best place for the issue (and accordingly, very little discussion). The second post is in the "Statistical Procedures" forum, where there is more discussion. The solution that I am pursuing is essentially a Chow test as mentioned in an earlier.response: 1) duplicate the sample where the duplicate sample uses the alternative measurements for the variables; 2) assign a dummy variable to distinguish observations between the two samples; 3) interact the dummy variable with the X1 and X2 variables such that the model becomes: Y = X1 + X2 + D*X1 + D*X2. The coefficients for the interactive variables pick up the differential response between the two samples. Using a duplicate sample raises the issue of independence/clustering of observations. So, I evaluate the statistical significance of the coefficients using two-way (by subject and by time) cluster robust standard errors. Please offer any suggestions or criticisms of this approach.  Thank you,  Rick