I have seen many articles in reputed impactor journals which have used only CFA before running multilevel SEM to test multilevel hypothesis on nested data. for example Hirst, G., Van Knippenberg, D., Chen, C. H., & Sacramento, C. A. (2011). How does bureaucracy impact individual creativity? A cross-level investigation of team contextual influences on goal orientation–creativity relationships. Academy of management journal, 54(3), 624-641.
however, many researchers used multilevel CFA before running multilevel SEM to test multilevel hypothesis on nested data. why there is contradiction and is it allowed to use simple CFA for nested data before running multilevel SEM to test multilevel hypothesis. Any authentic reference ?
To me, it seems inconsistent/illogical to use single-level CFA but multilevel SEM when dealing with nested/hierarchical data (unless some other kind of adjustment for nesting/non-independence of observations was made in the CFA). The non-independence that can arise from nested data affects CFA fit statistics (e.g., the chi-square test of model fit), standard errors, and potentially parameter estimates in a similar way as it would affect SEM. I wonder what justification (if any) was provided in the papers that fit single-level CFA and multilevel SEM to the same nested data.
Many papers are published in AMJ and if we look on the article which i have shared in above link. No justification about using single level CFA found any where. AMJ is one of the top IF journal in management. If this is the case , why editor and reviewers neglected this
Many reviewers and editors are not knowledgeable or interested in statistical issues.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Coating