Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: logit aggregation question

From   Tim Wade <[email protected]>
To   [email protected]
Subject   Re: st: logit aggregation question
Date   Tue, 2 Mar 2004 06:48:05 -0800 (PST)

--- "David J. Bernstein" <[email protected]>
> Hello mighty Stata'ers,
> I have transaction data for a corporation that has
> companies throughout the US.
> Using a logit model with dummies and continuous
> variables, I would like to study the race effect on
> promotions by company and then aggregate the results
> for an overall effect.  Like with a Mantel-Hanzel
> statistic that is used for combining 2x2's.
> Can this be done?  I'm a relatively new Stata user
> so
> I apologize if this is a beginners question.
> Thanks,
> David Jay Bernstein

This is very well summarized in Steve Selvin's book:
Statistical Analysis of Epidemiological Data, p.
238-240. Essentially, the Maentel Haenszel chi square
is equivalent to a likelihood ratio test of the model
with the dummy variables compared to the model without
the dummy variables (without any interaction terms).
The summary odds ratio for promotions is equivalent to
the exponentiated coefficient for promitions in the
model with the dummy variables for race (again with no
interaction terms). All of this can be easily done in
Stata, see commands: lrtest, xi, and logit/logistic.

Hope this helps, Tim

> __________________________________
> Do you Yahoo!?
> Get better spam protection with Yahoo! Mail.
> *
> *   For searches and help try:
> *  
> *
> *

[email protected]

Do you Yahoo!?
Yahoo! Search - Find what you�re looking for faster
*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index