Description
Testing whether a hypothesised factor structure fits survey data using indices such as CFI, TLI, RMSEA and SRMR before trusting scale scores or SEM paths.
Confirmatory factor analysis specifies in advance which items load on which factors and tests how well that model reproduces observed covariances. Because the chi-square test rejects almost any model at large N, approximate fit indices are used: the Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) compare the model with a null model; RMSEA estimates misfit per degree of freedom with a confidence interval; SRMR summarises residual correlations. Common guidelines (Hu and Bentler, 1999) are CFI and TLI of 0.95 or above, RMSEA of 0.06 or below and SRMR of 0.08 or below, but these are heuristics, not laws. Rigour comes from reporting all indices with the estimator used, inspecting local misfit through modification indices and residuals, and resisting atheoretical re-specification.
Attributes
- Discipline
- Survey & Psychometrics
- Loss type
- Siloed
- Stage
- Evaluation
- Data
- Quantitative
- Effort
- Days
Steps
- Specify the measurement model from theory or a prior EFA, with at least 3 items per factor.
- Fit with an estimator suited to the data: ML or MLR for approximately continuous items, WLSMV for ordinal items with 5 or fewer categories.
- Report chi-square (df, p), CFI, TLI, RMSEA with 90% CI and SRMR.
- Judge fit against conventional guides (CFI/TLI 0.95 or above, RMSEA 0.06 or below, SRMR 0.08 or below) and treat values near the boundaries cautiously.
- Inspect standardised residuals and modification indices, and only free parameters with a substantive justification.
- Test measurement invariance across key segments (configural, metric, scalar) before comparing segment means, accepting a change in CFI of 0.01 or less.
Sources
- factor_analyzer
- semopy
- pyreadstat (SPSS import)
- scikit-learn (MICE)
- pingouin (reliability)
- pandas
- Kline, R. B. (2023) Principles and Practice of Structural Equation Modeling, 5th ed.
- Hu, L. & Bentler, P. M. (1999) Cutoff criteria for fit indexes in covariance structure analysis. Structural Equation Modeling
- Brown, T. A. (2015) Confirmatory Factor Analysis for Applied Research, 2nd ed.
- The Inquiry Series, C3 — Survey Pipeline in Python — From Raw Responses to Structural Equation Model
- Tenenbaum, G. & Eklund, R. (eds), Handbook of Sport Psychology, 3rd ed. (2007)
Also known as
- CFA model fit
- Goodness-of-fit evaluation
Categories
- Analytics & Behaviour
- Survey & Psychometrics