Discuss Your EFA Results

Exploratory Factor Analysis (EFA) Report

I conducted an Exploratory Factor Analysis (EFA) to examine the underlying factor structure of my questionnaire items related to students’ perceptions of using ChatGPT in academic writing.

1. Extraction and Rotation Methods

For the extraction method, I used Principal Component Analysis (PCA) because it is commonly applied to identify the main dimensions underlying a set of observed variables.

For the rotation method, I used Varimax rotation with Kaiser normalization. This orthogonal rotation method helps produce a clearer factor structure by maximizing the loading of each item on one factor while minimizing cross-loadings on others.

2. KMO and Bartlett’s Test

The Kaiser-Meyer-Olkin (KMO) value was 0.812, which indicates that the sampling adequacy is good and suitable for factor analysis.

The Bartlett’s Test of Sphericity was statistically significant:

• χ² = 356.427
• df = 45
• p < .001

This result suggests that the correlation matrix is not an identity matrix, meaning that the variables are sufficiently correlated to proceed with EFA.

3. Items and Factors Retained

A total of 10 items were included in the analysis.

Based on the eigenvalue-greater-than-1 rule and the scree plot, 2 factors were retained.

The two factors accounted for 67.4% of the total variance, which is considered acceptable.

The retained factors were:

• Factor 1: Perceived Usefulness (6 items)
• Factor 2: Ease of Use and Challenges (4 items)

Items with factor loadings below 0.50 or strong cross-loadings were removed to improve the clarity and reliability of the scale.

These factors were retained because they were theoretically meaningful and statistically supported by the output.

Reply to a classmate 

I found your decision to retain three factors interesting. Could you explain whether you used the scree plot or only the eigenvalue > 1 criterion? Sometimes the scree plot provides a clearer justification for the number of factors retained.