The result of the exploratory factor analysis for the measurement tool I used is not compatible with the original scale. For example, in the original study, item 3 is in factor 1, but in the study I conducted, it is in factor 3. In general, all items in the scale show factor loadings in different sub-dimensions than the original ones. How should I proceed?
When your EFA results don't match the original scale, first re-examine your data and methodology. Consider differences in sample characteristics, cultural or contextual factors, and any potential methodological issues. You may choose to report the new factor structure if it is theoretically justifiable and supported by your data. Additionally, assess the reliability and validity of the new structure before proceeding with further analysis.
The measurement tool I used was adapted to the sample of Turkish adolescents. I also use the measurement tool on Turkish adolescents. Therefore, I cannot explain the situation with a cultural or age-related factor. In this case, do you think it would be appropriate to report my own results and continue the analysis based on my own results? Or is it healthier to consider the data of the original study? However, when I consider the factor structure in the original article, the reliability values also seem lower.
Hi Didem,
as you re-applied a formely established model strucure, I would recommend to conduct a CFA instead of an EFA. As this model will likely fail, one option would be to frame it as a unsuccessfull replication. I would also bet that the model was problematic in the first place. So perhaps you can get the original study of the first attempt and also test the model there and come up with a more appropriate solution. EFA often result in dirty structures as any item can load one more then one factor. This is not problematic perse but I would always try to identify a small number (2-3) indicators that are crystal clear. It is a strange cultural habit that science seem to few a dirty dozen (items) as superior to a small number of clear items. Bad things don't get better if you add more bad things.
And of course, sampling variations should be taken into account. How large was your sample? Was the context of the study (which can shape understanding of the items) different from the one in the former study (the context can be: Different cover story of the study or different other scales / measures in the questionnaire).
Best,
Holger
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication