Statalist The Stata Listserver


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

Re: st: RE: xi3: mlogit effect coding


From   Maarten buis <maartenbuis@yahoo.co.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: RE: xi3: mlogit effect coding
Date   Sun, 4 Feb 2007 11:39:12 +0000 (GMT)

--- Diego Bassani <diego.bassani@gmail.com> wrote:
> the top panel is actually a prevalence ratio (the ratio of the 
> prevalences among exposed and unexposed) and when I use effect coding
> I am subtracting this estimate from the 'mean' prevalence rate across
> all categories of the dependent variable (not dividing it), obtaining
> a prevalence ratio difference (or, a relative risk difference - 
> although i disagree with the use of this term in cross sectional 
> designs). Why do you disagree with it?

I begin to suspect that this is due to a difference in the use of
terminology between disciplines. In my particular sub sub branch of
sociology these are all called odds ratios, and for good reason because
they are ratios of odds. Disadvantage is that you have to be very
precise in describing the odds ratio, i.e. you have to describe both
the odds and the groups you are comparing. This doesn't make nice
prose. This use of the words odds ratio is very clear in for instance
the literature on analyzing mobility tables (a cross classification of
father's class and chid's class) with log linear analysis, e.g. (Hout
1983).  

Rereading Bobby Gutierrez's post
(http://www.stata.com/statalist/archive/2005-04/msg00678.html) I
suspect that there are disciplines that limit the word odds ratio to
describe ratios of odds for two collectively exhaustive categories,
e.g. you either go to college or you don't. But maybe he can better
comment on that.

So in Bobby's c.s. use, when specifying the odds in the odds ratio, you
can speak of the odds of going to college because this automatically
implies that the other outcome is don't go to college. In the way the
word odds ratios is used in my discipline I have to specify both
outcomes that are compared in the odds, and these oucomes don't have to
be collectively exhaustive, e.g. you can compare going to college
versus dropping out of highschool. 

The problem with Bobby's c.s. use of the word odds ratio is that now
you have to invent a new word to describe the outcome from -mlogit-. If
in your discipline they are called prevalence ratios then you should
use that word.

Hope this helps,
Maarten

Michael Hout (1983) "Mobility Tables". Thousand Oaks: Sage.

-----------------------------------------
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

http://home.fsw.vu.nl/m.buis/
-----------------------------------------


	
	
		
___________________________________________________________ 
New Yahoo! Mail is the ultimate force in competitive emailing. Find out more at the Yahoo! Mail Championships. Plus: play games and win prizes. 
http://uk.rd.yahoo.com/evt=44106/*http://mail.yahoo.net/uk 
*
*   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/



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