The article introduces new methods to estimate panel data models with censored variables and selection bias. The researchers use a two-step approach, adding extra variables to account for hidden factors causing bias. They test for different types of bias and include dynamics and state dependence in their analysis. The study shows that their method can effectively handle these issues and test underlying assumptions.