Thanks @junpenglao . I know that the observations will always be greater than 0. However, due to the mu, and sigma values from the upper levels of the model passed into the Normal distribution, the output takes on some negative values.
I may be confused, but isn’t this different from the censored values and what is done in the example? Here, only my output needs to be censored, not my observations. Do I need a truncating model vs. a censored model? I think I want to modify the normal distribution such that all output less than 0 have no weight. Would I still need to rewrite the PDF/CDF? Sorry, if I totally missed the point: I appreciate your help!