RE: st: question on chow test

[Richard Williams <Richard.A.Williams.5@nd.edu>]

At 03:06 PM 2/20/2005 -0500, Richard Williams wrote:
> I'd probably want to run the single regression.  I'd also probably want to 
> focus on the most critical contrasts, e.g. do the effects of x1-x4 differ 
> across groups or don't they?  If you do want to try out all the different 
> possibilities, you should go for more stringent significant levels.  But, 
> the more complicated you make this, the more difficult it will be to 
> interpret, and the more likely you are to come up with results that may be 
> ideosyncratic to your data set.  Just my opinion, others may disagree.
Just to follow up on my own message: for a single variable, you have 6 
potential contrasts: Group 4 v. 3, 4 v. 2, 4 v. 1, 3 v.2, 3 v. 1, 2 v. 
1.  And, you have 4 variables, so that is 24 potential contrasts.  Plus, 
you are thinking about making comparisons across variables, which 
potentially adds dozens more contrasts.  I would want to be careful that my 
discussion didn't wind up sounding like a random collection of significant 
results, e.g. x1 had a significantly larger effect on group 3 than it did 
on the others, but x2 had larger effects on groups 1 and 2 than it did on 3 
and 4, but x3 on the other hand...Much tidier is something like x1, x2, x3 
and x4 consistently had larger effects on group 3 than they did for the 
other groups.

Of course, a more scattered pattern might actually make sense if it was 
consistent with what you hypothesized; but personally I don't like 
seemingly random results and I often suspect they are more flukes than real 
differences.

*
*   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/