I have a dataset of n=90 participants with ~20 variables for two conditions, with each variable having 4 data points (not statistically independent) for each condition. I am thinking about conducting multilevel exploratory factor analysis, but I am unsure whether my sample size is insufficient in terms of power. What would be a reasonable sample size for this kind of analysis?
2
-
$\begingroup$ I am a little confused as to what your multilevel structure is. When you say, "...with each variable having 4 data points (not statistically independent) for each condition," does this mean you have longitudinal data such that each of the 90 participants has provided data on the 20 variables at 4 different time points? Or do you mean something different? $\endgroup$Erik Ruzek– Erik Ruzek2025-08-28 20:05:28 +00:00Commented Aug 28 at 20:05
-
$\begingroup$ Sorry for not being clear! The data is not longitudinal. It is a memory study, where the participants have provided 4 memories, then rated each of these according to 20 different variables. $\endgroup$LR1996– LR19962025-08-29 09:25:00 +00:00Commented Aug 29 at 9:25
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
There are a variety of rules of thumb for sample sizes needed for exploratory data analysis. None of them say that 90 is OK (especially with 20 variables).
You have multilevel exploratory factor analysis, which makes it even more challenging.
The only possible caveat is that as the magnitude of the loadings increases, the sample size required to correctly identify the structure decreases - so if your loadings are extremely high, you might be OK. But if your loadings are extremely high, you can tell what's happening by eyeballing the correlation matrix, and you don't need EFA.
