I would like to use an exploratory factor analysis to see if there are underlying factors within a measure. This measure was used at a baseline time point and again at six months. It is anticipated the scores will change, and some factors may be more resistant to change than others.
What is the best way to find these factors? Initially, I thought to perform an EFA just on the baseline data, however this may not be applicable to the 6 month data and it would not take into account how these scores change over time.
Thank you!
What creates a set of factors is the relationship (literally, the correlations) among the items. So, you could have the same factors at both times and still have the scores on the underlying variables change, as long as those scores move in a consistent direction.
But whether the factors are the same at both times is an empirical question that you would need to address through Confirmatory (rather than Exploratory) Factor Analysis.
You bring up an excellent point. The phenomena you are wondering about is called response shift (changes in how participants conceptualize or answer a measure). I agree with David Morgan for a common way to assess for that and add that people often use baseline scores to perform an EFA so if you go with examining baseline scores only, you will be fine with reviewers. However, if you are interested in testing for response shift or learning more, I would check out the literature on response shift .
https://www.sciencedirect.com/science/article/abs/pii/S0277953699000477 Article Item response theory and factor analysis as a mean to charac...
Article Methods to detect response shift in quality of life data: A ...
Focus EFA at pre/before assessment. Can conduct CFA between pre-post structures.
Hope this helps
Matt
I agree with David Morgan and Mark Relyea.
Nice literature, Mark.
I really appreciate.
- Badges
- Science topic
- Similar topics
- Correction