Statalist


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

st: Non-normally distributed dependent variable for Random-Intercept Model and 2SLS


From   Christian Weiß <mail@cweiss.org>
To   statalist <statalist@hsphsun2.harvard.edu>
Subject   st: Non-normally distributed dependent variable for Random-Intercept Model and 2SLS
Date   Mon, 9 Nov 2009 17:22:13 -0500

Dear Statalist,

I am using a dependent variable which is almost, but according to the
respective tests, not normally distributed (even after transforming).

The dependent variable is used in in a Random-Intercept and a 2SLS
model and due to the non-normality I assume that the confidence
intervalls are biased.
Thus, I am wondering how to correct for the non-normality.

I heared about alternative methods to calculate the confidence
intervalls which will be valid also for non-normal distributed
variables (boostrapping and Tschebyscheff - I could not find an
implementation of the latter one in Stata though).

Also, I heared read somwhere in the statalist that the robus SE
options of xtmixed and ivreg2 might help to deal with this problem, is
that correct?

I am happy for any further suggestions and for tipps of any
implementation of the Tschebyscheff or Markov's method.

Best
Christian
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index