I have developed a tool and validated it through EFA and CFA process. I split my sample into two for these two processes. Now I want to calculate composite scores for the sub scales and full scale. Which sample data should I use? Is it EFA sample or CFA sample or can I combine these two into the original full sample which of these two samples were derived from?
hi Emmanuel
Thanks for the answer. I didn't mean the composite reliability score which I have done already. I mean the composite score for the scale in order to compare, correlate and do regression with the scale score. Hope this clarify your question.
Thanks
Kanthi Hettigoda, I'm aware that you asked your question some time ago, and I hope you've been able to move ahead successfully by now. However, I'll offer an opinion in case it helps others.
If your CFA and EFA results support each other, which I assume they do by reading between the lines of your question, I think you could go ahead and combine both data sets for subsequent analyses.
If the CFA and EFA results don't support each other, that's a different kettle of fish - but let's not worry about that if it if that situation doesn't apply in your case. If/when it does apply, I think some very judicious decision making, and perhaps further analyses, would be in order.