A version of factor analysis in which specific hypotheses about structure and relations between the latent variables that underlie the data are tested.

A version of factor analysis in which specific hypotheses about structure and relations between the latent variables that underlie the data are tested.

• A. Confirmatory factor analysis (CFA)
• B. Principal component analysis
• C. Component matrix
• D. Compound symmetry

Answer: (A)

The idea being tested is testing hypotheses about how latent constructs are structured and relate to each other. Confirmatory Factor Analysis is this approach: you specify a measurement model that states which observed variables reflect which latent factors, which factors may be correlated, and which loadings or paths are constrained. Then you check how well that hypothesized structure fits the data using fit statistics. This lets you confirm whether your theory about which items measure which factors and how those factors connect holds up. This differs from exploratory methods like principal component analysis, which aim to uncover structure in the data without predefined hypotheses about latent factors or their relationships, and from other terms like a component matrix or compound symmetry, which are not about testing latent-structure hypotheses.

The idea being tested is testing hypotheses about how latent constructs are structured and relate to each other. Confirmatory Factor Analysis is this approach: you specify a measurement model that states which observed variables reflect which latent factors, which factors may be correlated, and which loadings or paths are constrained. Then you check how well that hypothesized structure fits the data using fit statistics. This lets you confirm whether your theory about which items measure which factors and how those factors connect holds up.

This differs from exploratory methods like principal component analysis, which aim to uncover structure in the data without predefined hypotheses about latent factors or their relationships, and from other terms like a component matrix or compound symmetry, which are not about testing latent-structure hypotheses.