Dear Statalist,
I want to estimate the endogenous multinomial treatment model proposed by Deb and Trivedi (2006) with the user written command mtreatreg that builds upon the paper.
Specifically, I have a continous outcome variable with an endogenous multinomial treatment variable. I specifyed the outcome as normal - therefore the assumption is that the errors are normally distributed. However, when looking at residuals they show excess kurtosis (but are symmetric) and a Shapiro Wilk test shows that they are not normally distributed. How will this impact my result?
Kind regards,
Hel
Related Posts with Joint ML estimation of normal and logit model (-mtreatreg- ,user written command), violation of gaussian noise assumption
Add significance stars to F-stat with esttabDear all, I was wondering how to add significance stars for F-stat in my tables with esttab. I trie…
ADF is WLS?Dear friends of STATA, I have version 13 of Stata and I wanted to ask you if the ADF method used by…
Writing memory data to word tableHello As part of my job the results I get from Stata need to put into tables in a word document to…
Summarize Variables With District WeightsMy data comes with sample weights to make it representative of a school district. I know to tabulate…
Testing if the difference between two rates is more than 5%.I need to test if the difference between male and female program coverage rates (number treated/targ…
Subscribe to:
Post Comments (Atom)
0 Response to Joint ML estimation of normal and logit model (-mtreatreg- ,user written command), violation of gaussian noise assumption
Post a Comment