When to use it
- •You are developing or adapting a scale and do not yet know its structure.
- •You have a set of items you believe measure a smaller number of underlying constructs.
- •You want evidence for how many factors to retain before confirming them.
Assumptions to check first
- •Adequate sample size — a common rule of thumb is 5–10 respondents per item, but sampling adequacy statistics matter more than any ratio.
- •KMO ≥ .60 and a significant Bartlett's test of sphericity, indicating the correlation matrix is factorable.
- •Items correlate enough to share variance, but not so highly that they are redundant.
What to report
- •KMO and Bartlett's test, to justify that factor analysis was appropriate at all.
- •The extraction and rotation methods used, and why (oblique rotation if you expect factors to correlate — in social science, they usually do).
- •The number of factors retained and the criterion used (parallel analysis is stronger than the old eigenvalue-over-1 rule).
- •The full pattern matrix with loadings, and the variance explained.
APA 7 example
The KMO measure verified sampling adequacy (KMO = .87) and Bartlett's test of sphericity was significant, χ²(66) = 1,204.3, p < .001. Principal axis factoring with oblimin rotation retained three factors explaining 61.4% of the variance. All items loaded above .50 on their intended factor with no cross-loadings above .32.
Numbers are illustrative — the Advisor generates this sentence with your own results.
The mistake reviewers catch
Running EFA and CFA on the same respondents is the single most common reason scale-development papers get rejected. Confirming a structure on the data that produced it is circular — split your sample, or collect a second one.
Run it on your own data
Paste or upload your dataset — nothing leaves your browser. The Analysis Advisor checks the assumptions above, runs efa, and drafts the APA Methods and Results text with your numbers in it.
Open the Analysis Advisor →Related analyses
Frequently asked questions
When should I use efa?
How many constructs underlie my items, and which item belongs where? You are developing or adapting a scale and do not yet know its structure.
What do I need to report for efa in APA 7?
KMO and Bartlett's test, to justify that factor analysis was appropriate at all. The extraction and rotation methods used, and why (oblique rotation if you expect factors to correlate — in social science, they usually do). The number of factors retained and the criterion used (parallel analysis is stronger than the old eigenvalue-over-1 rule). The full pattern matrix with loadings, and the variance explained.
What is the most common mistake with efa?
Running EFA and CFA on the same respondents is the single most common reason scale-development papers get rejected. Confirming a structure on the data that produced it is circular — split your sample, or collect a second one.