EFA Procedure and Results

EFA Procedure and Results

Trần Huỳnh Gia HânHUF04 -

For this study, Principal Component Analysis (PCA) was used as the extraction method, and Varimax rotation was applied to achieve a clearer factor structure. The ...

更多...

For this study, Principal Component Analysis (PCA) was used as the extraction method, and Varimax rotation was applied to achieve a clearer factor structure. The Kaiser-Meyer-Olkin (KMO) value is expected to be above .60, indicating that the sample is adequate for factor analysis. Additionally, Bartlett’s Test of Sphericity is expected to be statistically significant (p < .05), confirming that the variables are sufficiently correlated to proceed with EFA. A total of 15 items were included in the analysis. Based on the eigenvalue greater than 1 criterion and the theoretical structure of the questionnaire, three factors were retained. These factors correspond to:

  • Frequency and variety of use
  • Accuracy and confidence
  • Challenges in using complex sentences

Items with factor loadings above .40 were retained, while any items with low loadings or strong cross-loadings would be considered for removal.

EFA Procedure and Results

Vo Dao Trang ThyHUF04 -
Your explanation is clear and well organized. I like that you mentioned both the extraction and rotation methods, and your decision to retain items with loadings above .40 ...

更多...

Your explanation is clear and well organized. I like that you mentioned both the extraction and rotation methods, and your decision to retain items with loadings above .40 also seems reasonable. One thing you might improve is that some parts are still written as expectations rather than actual results, especially the KMO and Bartlett’s test values. It would be even stronger if you included the real statistics from your output.

EFA Procedure and Results

Nguyễn Đăng HảiHUF04 -
I think your approach is well-structured, especially your use of PCA with Varimax rotation and the clear criteria for retaining items. Linking the number of factors to both...

更多...

I think your approach is well-structured, especially your use of PCA with Varimax rotation and the clear criteria for retaining items. Linking the number of factors to both eigenvalues and theory is also a strong decision.

I do have a few questions though. First, since you’re using PCA, did you consider whether Exploratory Factor Analysis (e.g., Principal Axis Factoring) might be more appropriate for identifying underlying constructs rather than just data reduction? It could slightly change how the factors are interpreted.

Also, how will you handle items that load above .40 on more than one factor? Do you already have a rule (e.g., difference threshold) for deciding whether to keep or remove them?

Finally, I’m curious about your third factor—“challenges in using complex sentences.” Do you expect this to behave as a negative construct compared to the others? It might be interesting to check whether it correlates differently with the other factors.