[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: group size needed for mixed models (binary response)

From   "Verkuilen, Jay" <>
To   <>
Subject   RE: st: group size needed for mixed models (binary response)
Date   Tue, 27 Nov 2007 13:33:45 -0500

Anders Alexandersson wrote (replying to several people):

>>>I suggest a third alternative, one with a more Bayesian flavor (a la
Bartels or Gelman): estimate the multilevel model, which uses partial
pooling, but also the classical models with no or complete pooling.<<<

Actually the Burnham and Anderson volume discusses multimodel inference.
However, I want to be careful about providing advice to a "client" (in
the broad sense of the word) that is too complex. 

>>>In response to Jeph, the clustering effect is determined not only by
the size of the intraclass correlation coefficient but rather by the
design effect. A brief reference is Linda Muthen's posting at<<<

Thanks, that's a very nice reference. 

>>>In response to Jay, the real issue is whether a multilevel model is
justified. For example, I think that only a multilevel model can
separately estimate the predictive effects of an individual predictor
and its group-level mean, which sometimes are interpreted as "direct"
and "contextual" effects of the predictor.<<<

True enough. In this case, it seems to me---ignorant that I am of the
details of the study---that the answer is "no" and that the single fawn
vs. twin fawns is more of a nuisance than an important study question in
its own right. Even if it's justified, one has to question whether
fitting it will be possible. It's all well and good to require the
"right" model but (paraphrasing Tukey), often the best we get is a good
approximate answer. Nonetheless, I do think that trying the mixed models
will be worthwhile but remain skeptical that they will work here. 

J. Verkuilen
Assistant Professor of Educational Psychology
City University of New York-Graduate Center
365 Fifth Ave.
New York, NY 10016
Office: (212) 817-8286 
FAX: (212) 817-1516
Cell: (217) 390-4609

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index