Dear statalist community,

I am conducting a multilevel negative binomial regression analysis in stata (menbreg command) and was wondering how the coefficients are correctly interpreted. Related to this question, which defaults does stata use? I checked the manual, however I am not sure whether stata uses the log link and therefore estimates the log odds? Are conditional or marginal estimates calculated?

Please find below an example. In the syntax, the dependent variable is times of support/ help of an individual enacted towards their peers. Support is regressed on two independent variables: performance score (0-100%) of the individual and performance score of its peer. The dependent variable - times of support - is nested in the individual. That is, for each peer in their team we count how often the individual helped this peer. Since there are many zeros (no support), we use the menbreg analysis. Please note, the IVs are standardised, therefore the coefficients are labeled as betas.

Could you please let me know how to correctly interpret the coefficients in the stata output below? All advice would be highly appreciated.

Thank you,
Lea

menbreg support c.performanceIndividual c.performancePeer || individual:
Array