I am running a truncated normal regression and a lognormal regression as the second part of a double hurdle model. The dependent variable is transaction revenue and I have several independent variables, some numerical and some categorical coded as dummy variables. The truncated normal regression is left-truncated at point 0.

I use the following line of code for the truncated normal regression:

amount_reg <- truncreg(transactionRevenue ~ . - logTransactionRevenue, data = data_modelling_omit)

and this for the lognormal regression:

amount_reg_log <- glm(logTransactionRevenue ~ . - transactionRevenue, data = data_modelling_omit, family = gaussian)

I want to compare the coefficients returned from the two models. How do I make coefficients comparable? Can I interpret the coefficients from the truncated regression directly, or do I have to calculate marginal effects?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.