Marginal effects in the censored regression model

We find that a well known result for marginal effects in the censored regression model with normally distributed disturbances applies more generally to any censored regression model in which the disturbances have a continuous distribution. The result suggests a comparison of the coefficients and marginal effects in alternative models, e.g. normal vs. logistic, that is qualitatively different from the familiar counterpart in the probit and logit models for binary choice.