Multinomial Logistic Regression with Repeated Measures

General help and assistance with jamovi. Bug reports can be made at our issues page: https://github.com/jamovi/jamovi/issues . (If you're unsure feel free to discuss it here)
Post Reply
Whirly123
Posts: 31
Joined: Mon May 06, 2019 3:07 pm

Multinomial Logistic Regression with Repeated Measures

Post by Whirly123 »

I am going to be running a Multinomial Logistic Regression.

I am using a couple of questionnaires as predictors but my outcome variable is the result of repeated measures. Participants make a choice and do so across multiple rounds to either continue or do one of two things before continuing (3 options).

I am checking to see if Jamovi's Multinomial Logistic Regression is appropriate or whether I need to use a mixed-effects model here (currently unavailable in Jamovi for multinomial outcomes). I am concerned that without using a mixed-effects model this would be inappropriate since each participant has only one questionnaire result but the outcome variable is repeated over multiple rounds.
User avatar
jonathon
Posts: 2613
Joined: Fri Jan 27, 2017 10:04 am

Re: Multinomial Logistic Regression with Repeated Measures

Post by jonathon »

hi,

i don't think the multinomial logistic regression in jamovi can handle repeated measures. a mixed effects model would seem to be the right choice to me.

cheers

jonathon
User avatar
mcfanda@gmail.com
Posts: 457
Joined: Thu Mar 23, 2017 9:24 pm

Re: Multinomial Logistic Regression with Repeated Measures

Post by mcfanda@gmail.com »

You can decompose the multinomial model into a set of K-1 logistic models, where K is the number of levels of the DV. You define K-1 dummy variables and predict them with K-1 logistic models. The coefficients and tests you get are practically the same you would get with a multinomial analysis. What you'd miss is the omnibus test of the effect. It's like running an ANOVA with a regression module.

The other few differences between the two approaches are rather technical and appear only for very weird cases. Please see here for theory https://en.wikipedia.org/wiki/Multinomi ... egressions
Post Reply