David Airey wrote:
. . . I wonder if reviewers will buy this kind of
supporting documentation for analysis choice down the road.
--------------------------------------------------------------------------------
One approach to using an asymptotic method here would be to reference the
resulting test statistic (-xtlogit-'s likelihood-ratio chi-square, for
instance) against its empirical distribution in simulation under the null
hypothesis using the study's sample sizes and with the observed rho instead
of against the chi-square distribution. Wouldn't this overcome objections
that an asymptotic method, such as -xtlogit-, doesn't maintain Type I error
rate with small samples? I wonder how well this approach would be received.
Joseph Coveney
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/