Partial Factorial Invariance

In social and behavioral sciences, when multiple items are used to measure one construct, the condition of measurement invariance implies that the relations between the items and the constructs are the same across different groups and settings. In other words, there is no measurement bias due to group membership. Under the common factor model, which hypothesizes that the shared variances of the items can be attributed to one or more latent factors that coincide with the constructs of interest, factorial invariance refers to the condition that all items relate to the latent factors in the same way across subpopulations, such as demographic subgroups, time points, and measurement conditions. Partial factorial invariance (PFI) happens when only a subset of items function similarly across subpopulations. Because quantitative ...