ABSTRACT
The bifactor structure is a specific design of within-item multidimensionality. In bifactor modeling, all items load on a general or primary factor, and subsets of items each load on their own specific factor (Gibbons & Hedeker, 1992; Gibbons et al., 2007). Take a mathematics test, for example. Suppose that the test has three subdomains, or constructs, where items within each subdomain measure one of three mathematics domains: algebra, geometry, and trigonometry. However, all items across the three subdomains measure the general or primary factor of mathematics. If one wants to model the general or primary factor as a single construct while modeling each subdomains specific factor, the bifactor modeling suits the situation.
