EFA vs CFA
Understanding the crucial distinction between Exploratory and Confirmatory Factor Analysis is the cornerstone of robust scale development.
What is EFA?
- Purpose: Discover underlying factor structure without pre-defined hypotheses.
- Use Case: When you do not know how many factors exist or item allocations.
- Approach: Data-driven; allows items to load freely on multiple factors.
- Typical in: Early-stage research or scale development.
What is CFA?
- Purpose: Test whether a hypothesized structure fits the data well.
- Use Case: When you have a theoretical model or previous evidence.
- Approach: Theory-driven; you specify which items load on factors.
- Typical in: Scale validation or theory testing.
Visual Comparison
| Aspect | EFA (Exploratory) | CFA (Confirmatory) |
|---|---|---|
| Goal | Identify factor structure | Confirm hypothesized structure |
| Use When | Unsure of number/structure of factors | Theory or prior model exists to test |
| Approach | Data-driven | Theory-driven |
| Item Loadings | Items can load on multiple factors | Loadings fixed based on prior structure |
| Rotation | Required (e.g., Varimax, Oblimin) | Not used |
| Model Fit Indices | Not typically reported | Required (CFI, RMSEA, SRMR) |
| Software | SPSS, JASP, R | AMOS, LISREL, Mplus, lavaan (R) |
Real World Examples
EFA Scenario
A researcher creates a new well-being survey but doesn't know how many dimensions it measures. EFA is used to explore whether items group into dimensions like physical health, emotional health, and social well-being.
CFA Scenario
A second researcher tests whether this 3-factor model fits data from a different sample using CFA. They confirm that each item loads only on its designated factor.