Hi, I'm doing my master dissertation. After the mediation anlaysis, the mediator failed to have impact, so I did confirmatory factor anlaysis to find out the problems.
The result showed that my model fit is poor(CFI 0.675, RMSEA .104).
And I have 5 independent variables, 1 mediator. In my questionnaire design, for independent variables, each of them have at least 3-17 items. And my sample size is 107.
In this case, what could I do to revise the current situation or how could I find out the potential problems?
Thank you a lot if you could give me some advice.
I think the sample is very small for that process, also the difference between indicators by factor, I suggest eliminating factors that have few indicators, you can also use boottraping to perform your analysis
Well, the most obvious answer Is that you may not have a positive result. The sample size should be OK for ols regression. If you want to try variable selection (DO NOT just arbitrarily remove IVs) see my recommendations in the attached paper. Good luck, David Booth
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size