Posted by
statisticsdoc on
Jan 02, 2012; 6:50pm
URL: http://spssx-discussion.165.s1.nabble.com/Multinomial-logistic-regression-r2-for-two-different-sets-of-variables-tp5114543p5115171.html
Andrea,
As Scott suggests, AIC and BIC can be used to compare non-nested models,
although some caution must be exercised as there are circumstances in which
the use of these indices can be misleading (e.g., if the dependent variable
is transformed in one model and not in the other). Nested models are less
problematic as they tend to resemble one another except for the exclusion of
certain predictors.
If the following suggesting makes sense theoretically and in application,
you might want to reframe your question in terms of a test of nested models.
For example, you might want to ask whether set 2 makes am incremental
contribution to a model that already contains set 1. The comparison here
would lie between a model with both sets and one with set 1 only.
In any case, please also note that AIC and BIC indicate relative support for
a model, and do not mean the same thing as variance accounted for. If you
want to say that one model has substantially more support from the data, the
use of AIC and BIC would be helpful. However, the claim that one model
accounts for X% more variance cannot be addressed in the OLS sense as
logistic regression is inherently different. Nonetheless, I suspect that
you can address your research questions by focusing on the question of
relative support from AIC and BIC.
Best Regards,
Stephen Brand
www.StatisticsDoc.com
-----Original Message-----
From: SPSSX(r) Discussion [mailto:
[hidden email]] On Behalf Of
Rieden246
Sent: Monday, January 02, 2012 6:49 AM
To:
[hidden email]
Subject: Multinomial logistic regression: r2 for two different sets of
variables
Hi
I'm running a multinomial logistic regression and I'd like to find out which
set of variable contributes more to the total R2 (i.e.: variables that were
assessed before entering school (1th set) and variables that were assessed
after a few months in school (2nd set)).
I can only think of running two seperat regression models and compare the
two R2, but I'm sure there must be a more elegant way of doing this. I would
appreciate your help a lot.
Thank you in advance and kind regards
Rieden
--
View this message in context:
http://spssx-discussion.1045642.n5.nabble.com/Multinomial-logistic-regression-r2-for-two-different-sets-of-variables-tp5114543p5114543.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD