Exploratory Factor Analysis (EFA)
Exploratory factor analysis is a latent variable method that identifies a smaller number of underlying factors explaining the correlations among a set of observed items, without specifying the structure in advance.
탐색적 요인분석은 구조를 미리 정하지 않고 관측 문항들 간의 상관을 설명하는 더 적은 수의 잠재요인을 찾아내는 잠재변수 분석 방법입니다.
How do I find out which survey items group together into underlying dimensions?
설문 문항들이 어떤 하위 차원으로 묶이는지 어떻게 알아내나요?
Exploratory factor analysis examines the correlations among items to discover how many latent factors underlie them and which items load on each factor. It fits scale development and early validation when the factor structure is not yet established, with the number of factors guided by parallel analysis or scree plots.
When to use Exploratory Factor Analysis (EFA)
- Developing a new questionnaire and exploring its dimensions
- Multiple Likert-type items thought to reflect a few latent constructs
- Checking the structure of an existing scale in a new population
When to use something else
- Factor structure is hypothesized in advance: use confirmatory factor analysis
- Goal is data reduction without latent constructs: use principal component analysis
Assumptions and requirements
- Adequate sample size relative to the number of items
- Sufficient correlations among items, checked with KMO and Bartlett's test
- Approximately linear relationships among items
- Polychoric correlations considered for ordinal items with few categories
What the analysis reports
- KMO measure and Bartlett's test of sphericity
- Eigenvalues, scree plot, and parallel analysis
- Rotated factor loadings and communalities
- Variance explained by each factor
- Factor correlations for oblique rotations
How to report it
Exploratory factor analysis using [extraction method] with [rotation] rotation yielded [k] factors explaining [value]% of the variance, KMO = [value], Bartlett's χ²([df]) = [value], p = [value].
Related methods
Run Exploratory Factor Analysis (EFA) on Quorant
- Data you need
- a table (CSV or Excel) with item columns (e.g. questionnaire items)
- Price
- 600 credits for one analysis up to 10,000 rows (complexity level 4); larger or wider data costs more. Optional methodologist review: 2400 credits.
- Checking
- Python results are cross-checked against R reference outputs in every deployment
- Before you pay
- A free pre-analysis diagnosis checks your data and gives an exact quote; nothing is deducted until you approve it.