Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: RE: Re: xtmelogit convergence issues and log transforming IVs


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: Re: xtmelogit convergence issues and log transforming IVs
Date   Thu, 26 Jan 2012 09:49:12 +0000

I said that

"Normality of any predictor is not an assumption of regression-type
models, although it does no harm."

However, I never said or meant to imply that the marginal distribution
of predictors is immaterial (or if I did, that was wrong). The
marginal distribution of any variable can reflect the existence of
outliers. Even a kurtosis of 7, and certainly a kurtosis of 520, says
"outliers!!!???" tot me.

You'll have read stories of miners taking canaries down the mine. The
canaries were more sensitive to noxious gases than the miners. If they
keeled over, it was a sign of danger. Skewness and kurtosis are
statistical people's canaries.

What to do about outliers will depend, and "nothing" can be an option.
Transformation is another.



© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index