Back in May I posted a piece titled, "Gripe of the Day". With that post I provided some EViews code to run homoskedasticity tests for Logit and Probit models. Unfortunately, there was a small error in the code. This has now been fixed, and there is a note to this effect on the Code page for this blog, and in the original post.
The error affected the test results only at the second or third decimal places.
HT to eagle-eyed Sven Steinkamp at Universität Osnabrück for bringing the error to my attention.
© 2011, David E. Giles