

EFA with repeated or aggregate? 

Message/Author 


I am quite unfamiliar with MPlus, but in desparate need of help with EFA. I have quite complex data and therefore do not know how to do EFA in MPlus. Let's say that I have 100 participants that rated 60 targets on 4 items. I would like to see how these 4 items construct factors. The 100 participants filled in each item 60 times (once for each target). We have several possibilities now: 1. We use all the separate ratings as given by the participants 2. We have to calculate how each target is rated and use these scores 3. We can construct an aggregate score, for example, we sum all 60 times that a participant gave ratings to a target 4. We can construct an aggregate score based on how the target has been rated There are two main differences in these approaches: 1. The participant is leading (thus the ratings per participant are used, independent of whom they give ratings to) versus the target is leading (thus the received ratings per target, which is a combination of all 100 ratings they receive). 2. Using all separate ratings versus using an aggregate score. I truly hope one of you has experience in this. I am happy to provide clarification if needed. Thanks a lot! Maaike 


I have no experience with this, but I wonder if it could be analyzed as a twolevel analysis where participants are level2 and targets are level1, and where your outcomes are the scores on the 4 items. I don't know what the within model would be  one possibility is a 1factor model. You may want to ask this question on SEMNET. 


Thank you so much for this answer! 

Back to top 

