KHB command

[Sofiya Volvakova]

Hi,

I am trying to apply the khb adjustment to make coefficients across nested logit models comparable so that I can report of the influence of IV on my DV even after taking account of all mediators. Here, W1ExcludeYP is DV, W1ethgrpYP is IV and concomitant are my confounders.

However, the error I am getting is invalid stub() option; stub name too long. Any solution on what that can be?

 khb logit W1ExcludeYP i.W1ethgrpYP || i.in_poverty W1hiqualgP W1SOCMajorP school_disengagement W1hwndayYP W1heposs9YP i.W1ssclubfYP, concomitant (W1truantYP substance_use delinquency)

[Felix Bittmann]

Unfortunately, the ado has a few bugs. Sometimes it is enough to rename the vars to shorter names (var1, var2, ....). If you have access to Stata 18, you can try the new causal mediation models in there.