A power analysis software such as G3 can determine the minimum required sample size for logistic regression, but I can't find a software to determine the sample size for a multinomial logit regression
Hi Nicola, I have no idea, but start with this link and see if it leads anywhere: http://samplesizeshop.org/ Otherwise, the usual way to get at that is through simulations. Bob
http://samplesizeshop.org/
If I remember correctly from Econometrics classes, common practice dictates a sample size that exceeds 385 for multinomial logit regression
I am not aware of any software for power analysis for multinomial outcomes. Of course you could always perform a Monte Carlo simulation study. The software has build in commands for simulation studies and you can probably use the free student version for problem at hand.
Here's a suggestion:
http://www.methodspace.com/group/quantitativeresearch/forum/topics/gpowerbased-sample-size
Hi,
I feel the sample size is not the criterion for MLR.
Rather category and type of data shall determine the same.
Just see to it that the criterion variable is non metric and at least dichotomous.
Joanne M Eaves , G*Power can do *multiple* logistic regression, but I don't think it can do *multinomial* logistic regression.
- Badges
- Science topic