Hello everyone! I am performing a personality study to unravel the personality dimensions in an African primate species.
The group of subjects accounts for five individuals what makes my sample size very small. I have read that REFA is a good reduction method for small sample sizes. I have tried to perform this analysis with Matlab, but it is too heavy and it does not work.
Does anybody know another way to perform REFA computationally?
Thanks.
Hi Iván,
I recommend Bayesian Exploratory Factor Analysis using Mplus 7.2 or OpenBUGS. Bayesian EFA is preferred with small samples because of the Markov chain Monte Carlo. However, you cannot find an EFA with only five individuals, once the number of subjects is smaller than the number of parameters...
Thank you, Danilo. I know my sample is very small but I only have five subjects for my pilot study. In any case, I am considering each rater rating each individual as a single case which increases the number of cases vs the number of variables (5 subjects x 7 raters = 35 cases). I hope I'm on the right way... I'll try your suggestion. Thanks a lot!
I suspect you are too far along to consider it now, but Q-method (Brown etc.) may be a way to approach this type of sampling.
http://qmethod.org/faqs
http://qmethod.org/articles/vanExel.pdf
http://qmethod.org/tutoriallinks.php
Ok, Dale. Thanks. I've tried another way to perform my analysis but let's see what the editors say. I'll save your suggestion just in case. Cheers!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size