Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: Stata journal paper on interaction terms in count models


From   Maarten Buis <maartenlbuis@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Stata journal paper on interaction terms in count models
Date   Tue, 14 May 2013 10:51:45 +0200

On Thu, May 9, 2013 at 4:46 PM, James Bernard wrote:
> The more I read the more confused I get.... I don't understand why a
> subject so close to mathematics (not exactly math) can be so confusing
> to every one
>
> Perhaps the problem is that statistics/econometrics is becoming too vast

I think that part of the problem is just conceptual: an effect is
simply a comparison of groups while an interaction is a comparison of
comparisons. The latter is as a consequence a step further removed
from the stuff we see (the data) and thus harder to visualize. Another
part of the problem is that non-linear models are non-linear. The way
to solve that is to take the non-linearity seriously and report
(interaction) effects in the natural metric of that model. In case of
a log link function (like in a Poisson model) you need to think in
terms of ratios rather than differences. Once you do that everything
becomes a lot simpler, as I showed earlier in this thread.

-- Maarten

---------------------------------
Maarten L. Buis
WZB
Reichpietschufer 50
10785 Berlin
Germany

http://www.maartenbuis.nl
---------------------------------
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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