Hello everyone,

I would like to know if there is a coefficient stability test between two models, one of which is a unified model (it takes into account all the observations) and the other model is a subsample of the unified model (I estimate the same model but for specific observations). So I do not have a "complex" and "reduced" model as required by a LR test but rather two models with a different number of observations (but the same IV and DV variables).

PS : my model is a multinomial logit model

Thank you for your help