This document outlines the five steps of exploratory factor analysis: 1) assessing if the data is suitable, 2) determining the factor extraction method, 3) criteria for factor extraction like eigenvalues and variance, 4) choosing an orthogonal or oblique rotation method, and 5) interpreting the results by evaluating the sample size, multicollinearity, number of factors, rotation method, and naming the extracted factors. It provides guidelines for suitable sample sizes, variable ratios, and correlation values when assessing if data is suitable for factor analysis in the first step.

1. Exploratory factor analysis: A five-
step guide for
novices
brett.williams@med.monash.edu.au
Ted Brown
Andrys Onsman

2. Objectives of Exploratory Factor Analysis
• Reduce the number of variables
• Examine the structure or relationship between variables
• Detection and assessment of unidimensionality of a
theoretical construct
• Evaluates the construct validity of a scale, test, or
instrument
• Development of parsimonious (simple) analysis and
interpretation
• Addresses multicollinearity (two or more variables that are
correlated)
• Used to develop theoretical constructs
• Used to prove/disprove proposed theories

3. The Five-Step Exploratory Factor Analysis Protocol
1) Is the data suitable for factor analysis?
2) How will the factors be extracted?
3) What criteria will assist in determining
factor extraction?
4) Selection of Rotational Method
5) Interpretation

4. Is the data suitable for factor
analysis?
• General Rule:
• A number of textbooks cite the work of
Comrey and Lee in their guide to sample sizes:
– 100 as poor,
– 200 as fair,
– 300 as good,
– 500 as very good,
– and 1000 or more as excellent.

5. Is the data suitable for factor
analysis?
• Sample to Variable Ratio (N:p ratio):
• Another set of recommendations also exist
providing researchers with guidance regarding
how many participants are required for each
variable
– 3:1, as poor,
– 6:1, as fair,
– 10:1, as good,
– 15:1, as very good,
– or 20:1 or more as excellent.

6. Is the data suitable for factor
analysis?
• Factorability of the correlation matrix:
• Henson and Roberts pointed out that a
correlation matrix is most popular among
investigators.
– ±0.30=minimal,
– ±0.40=important,
– and ±.50=practically significant

7. Is the data suitable for factor
analysis?
• Kaiser-Meyer-Olkin (KMO) Measure of
Sampling Adequacy/Bartlett's Test of
Sphericity
• 0.50 considered suitable for factor analysis.
• 0.70 considered important for analysis
• 0.90 considered practically significant for
analysis

8. How will the factors be extracted?
• Principal components analysis (PCA)
• Principal axis factoring (PAF)
• Maximum likelihood
• Unweighted least squares
• Generalised least squares
• Alpha factoring
• Image factoring

9. What criteria will assist in
determining factor extraction?
• Cumulative Percentage of Variance and
Eigenvalue > 1 Rule
• Scree Test
• Parallel Analysis

10. Selection of Rotational Method
• Orthogonal rotation OR Oblique rotation.
• Correlation between factors

11. Interpretation
• Is the sample size adequate?
• Would you exclude any items on
questionnaire on the basis of multicollinearity
or singularity?
• How many factors should be retained?
• What method of rotation used and Why?
• Which item load on to which items? Can we
name these factors psychologically?