Extends the Heckman selection framework to panel data with individual random effects. The first stage models participation via a panel Probit specification, while the second stage can take a panel linear, Probit, Poisson, or Poisson log-normal form. Model details are provided in Bailey and Peng (2025)