This note illustrates that the typical parameter, beta, in a censored regression model can be used to calculate an interesting marginal effect even when the errors in the model and the explanatory variables are not independent. The result is relevant for cross sectional models such at the ones...