📸 CFA in AMOS — IBM SPSS's structural equation modeling module
CFA vs. EFA: The Key Distinction
Exploratory Factor Analysis (EFA) discovers factor structures from data. Confirmatory Factor Analysis (CFA) tests a pre-specified factor structure against the data. CFA is used in scale validation, measurement model testing before SEM, and replications of previous EFA findings.
Model Fit Criteria
- CFI ≥ .90 (good); ≥ .95 (excellent)
- RMSEA ≤ .08 (acceptable); ≤ .05 (good) with narrow 90% CI
- SRMR ≤ .08 (good)
- χ²/df ≤ 3 (good)
Running CFA in AMOS
Step 1: Open AMOS from SPSS: Analyze → IBM SPSS Amos.
Step 2: Draw latent variables (ovals) and observed indicators (rectangles). Draw arrows from each latent to its indicators. Add error terms to all indicators.
Step 3: Connect your SPSS data file. View → Analysis Properties → Output: Standardized estimates, Modification indices.
Step 4: Calculate Estimates. Review fit indices and factor loadings.
📸 CFA model fit indices — all thresholds met
APA Reporting
CFA indicated acceptable model fit for the 3-factor measurement model: χ²/df=2.41, CFI=.947, RMSEA=.061 (90% CI [.048, .074]), SRMR=.054. All factor loadings were statistically significant (λ=.61–.88, all p<.001).