Hey researchers,
In my research, I tried to employ one EFA with all of the variables (IV & DV) together but the result came out with both my IV & DV mixed together. I would love your thoughts on whether it is theoretically possible to employ two EFA, one for IV and the one for DV.
Background: The research I am doing is about a new system in my country. Since it's a new system, existing research about it is very limited. There is no prior established model yet. So, I made a new survey and while I was making it, I had in my mind what items are going to be the DV and IV, thus I divided it into different sections.
My reason of doing EFA separately: I personally think it makes a lot more sense to do it separately since IV and DV are meant to measure distinct constructs, forcing them into the same factor structure might create unexpected results.
The goal of doing EFA: find the hidden factors in IV and DV then use it to formulate my hypotheses and proceed with Simple Linear Regression to conclude my research.
DV: Efficiency and Security
IV: User Satisfaction
It probably makes sense to factor analyze the items for the IV(s) and DV in a combined EFA. That way, you can see which items/measures are ambiguous (i.e., do not clearly load onto their expected IV or DV factors) and identify those items/measures that do have clear loadings on their intended factors. You can then potentially exclude items/measures that are not in line with your expected factor structure.
Another approach would be to use confirmatory factor analysis (CFA) to begin with since you seem to have a priori hypotheses regarding which items/measures should load onto which factor.
- Badges
- Science topic
- Similar topics
- Psychology
- Cognitive Science
- Mind