When is it appropriate to redo a CFA? I am using another researchers survey instrument and the original CFA was done using a population different than the one I am going to use it for. (Canada vs U.S., medical practitioners vs. substance use counselors). Should I redo the CFA? My sample is also larger, 1700 to the original which was 200.
Do you publish a CFA in the same scientific paper as those that report your research findings? I am getting mixed answers on this.
CFA is a large-sample technique. "Large enough" depend on how complex your model is (how many parameters). According to Kline (2011) an ideal sample size-to-parameters ratio would be 20:1, i.e., if 10 parameters require statistical estimates, an ideal sample size would be 20 x 10, or N=200. I don't know how extensive your instrument is, but your large sample allows you to re-do the CFA with more trustworthiness.
I suppose you want to see if there is difference among the groups you collected data from, then a multigroup CFA can be an option, and YES, research findings can be published in the same scientific paper to make the paper more interesting:
Article Assessing understandings of substance use disorders among No...
Greetings from Norway
John-Kåre
As ı know, some journals want new cfa. Because some items can not be valid in your study. You do not want to study invalid iteme in your study. Furthermore, suppose that you want to do structural equation model (sem) and you have three or more scales, before do sem, you need measurement models for each scales. Measurement model = cfa.
Thank you so much for your replies. I had two differing points of view from colleagues and that got my thinking all turned around on this. I feel much more confident now moving forward
.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey