If non-trivial cross-loadings (β = .10; Asparouhov et al., 2015 ) were detected in the ESEMs and substantively important differences between the sets of results were observed, we interpreted this as evidence of practically significant bias due to the strict unidimensionality assumption imposed in CFA and analysis of scores (i.e., in the CFA or independent cluster models, items were only allowed to load on their target factor).
← all excerpts
Are Higher-Order Constructs in Evolutionary Psychology Attributable to Omitted Cross-Loading Bias? An Exploratory Structural Equation Modeling Approach.
1
—
—