Nonlinear Unidimensionality in Communication Science: Tests, Examples, and Implications
The ability to evaluate predictions and theories hinges on accurate, error-free measurement. As a result, communication scientists seek both highly valid and highly reliable measures. Regarding validity claims, specifically, a hypothesis of content validity is typically tested via factor analysis. Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) assume a linear relationship between indicators and the underlying construct. Thus, EFA and CFA miss instances of nonlinear unidimensionality, such as those encountered when a scale forms a Guttman Simplex. Given that such a measurement model is not uncommon and is difficult to detect, this article presents a method for testing the hypothesis of content validity when this kind of nonlinear relationship exists between scale items and the construct.