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

Re: st: What form of data needed for ipf function?

From   Maarten buis <>
Subject   Re: st: What form of data needed for ipf function?
Date   Fri, 15 Aug 2008 20:24:56 +0100 (BST)

--- Richard Williams <> wrote:
> I don't use this stuff much but I think the glm 
> command can give you the same info as ipf:

True, a week ago I was at a conference diner where some people who
worked or studied with Leo Goodman (who played an important role in
developing log-linear models) discussed how he thought of these models
as ANOVA for count data. So just as you can estimate an ANOVA using
linear regression with a number of dummies, so you can estimate a
log-linear model using Poisson regression using a number of dummies.
The subject of the conversation was scientific intuition, and the role
of analogies as an important way  of obtaining it. 

The similarity between ANOVA and log-linear models also explains the
rather anoying tendency that people who use ANOVA and log-linear models
have in common: they tend to report the fit-statistics of a (large)
number of models, and derive their conclusions only on those, the
parameter estimates are often not even reported at all... 

-- Maarten

Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands

visiting address:
Buitenveldertselaan 3 (Metropolitan), room Z434

+31 20 5986715

Send instant messages to your online friends 
*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index