Confirmatory Factor Analysis (CFA)
Statistics & analysis
A technique that tests whether measured items load onto the latent constructs a theory predicts, assessing construct validity.
CFA tests a measurement model you specify in advance: you tell the software which items load on which latent construct, and it reports how well that structure reproduces the observed covariance matrix. It is the confirmatory counterpart to EFA — you arrive with a hypothesis rather than searching for one.
Judge the model on a combination of fit indices rather than any single number: CFI and TLI at or above 0.90 (0.95 preferred), RMSEA at or below 0.08 (0.06 preferred) and SRMR at or below 0.08. Chi-square is reported by convention but is sensitive to sample size and is almost always significant in large samples.
CFA is also where convergent and discriminant validity are established, since standardised loadings, AVE and composite reliability all come out of the same model. Run it before any structural modelling: a path estimated on top of a poor measurement model cannot be meaningfully interpreted.
Where it's used
- Validating the measurement model before structural analysis in CB-SEM
- Confirming an adapted scale works in a new context
Software used
Related concepts
Guides that use this concept
All guidesNot sure where to start?
Book a free 15-minute consult. We'll map your next three steps — no obligation.