Dear All:

I use logistic regression analysis to model the likelihood that a drug-development project results in termination or approval. We include firm-specific fixed effects to control for unobserved firm heterogeneity. Our independent variables include failure and success experience of the firms. We assume that how many times you have failed or succeed in the past will affect your chances of future failure.

One of the reviewers at a journal suggested using Heckman selection model. After reading about it it seems to me that this would be redundant as our model already includes firm-specific fixed effects. I am just wondering if I am missing something in their suggestion.

Thanks
Antonio