Logistic regression with grouped data PreviousNext
Mplus Discussion > Categorical Data Modeling >
Message/Author
 Anonymous posted on Thursday, March 22, 2012 - 9:41 am
Is it possible to perform logistic regression with grouped rather than individual data in M-plus?
 Bengt O. Muthen posted on Thursday, March 22, 2012 - 10:40 am
Yes. Use the FREQWEIGHT option.
 Anonymous posted on Saturday, March 24, 2012 - 5:56 am
How could I set this up in M-plus? For example, my data begins:

p=0.745; age = 25
p=0.372; age = 52

I want to perform a logistic regression of p on age.

Thanks
 Linda K. Muthen posted on Saturday, March 24, 2012 - 6:27 am
See Example 3.5 in the Mplus User's Guide.
 Bengt O. Muthen posted on Saturday, March 24, 2012 - 6:27 am
See slides 28 and 31 of the Topic 2 short course handout on our website.
 Anonymous posted on Thursday, March 29, 2012 - 1:39 am
Thanks. I have fitted the grouped data model in Mplus using the freqweight option and am able to get the same results as in other stats programmes.

I note that when I ask for influence and cooks values I get values for both the 'successes' (outcome=1) and the 'failures' (outcome=0) because of the way the data is constructed. I wondered if when examining these it is necessary to look at both successes and failures, or just to look at the values for the successes, as the model fitted is estimating the log odds of 'success'?
 Linda K. Muthen posted on Thursday, March 29, 2012 - 4:25 pm
I would look at both. These are person-related diagnostics not model-related.
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: