In factor analysis, what is the main purpose of rotation?

In factor analysis, what is the main purpose of rotation?

• A. Change the number of factors to fit the data.
• B. Eliminate sampling error.
• C. Transform the factors to maximize large loadings and minimize small loadings, improving interpretability.
• D. Estimate the correlations among observed variables.

Answer: (C)

Rotation in factor analysis is done after extracting the factors to reorient the factor axes so the pattern of loadings becomes easier to interpret. The goal is to produce a simple structure where each observed variable loads highly on a small number of factors and has smaller loadings on the others. This makes it clearer what each factor represents because the variables cohere into distinct, interpretable groups. Rotation does not change how many factors you have or the total amount of explained variance; it just redistributes the loadings to improve clarity. It also doesn’t estimate correlations among observed variables—that part comes from the initial extraction and the correlation structure of the data. Rotation can be done so factors are kept uncorrelated (orthogonal) or allowed to correlate (oblique), depending on whether you expect the underlying constructs to be related.

Rotation in factor analysis is done after extracting the factors to reorient the factor axes so the pattern of loadings becomes easier to interpret. The goal is to produce a simple structure where each observed variable loads highly on a small number of factors and has smaller loadings on the others. This makes it clearer what each factor represents because the variables cohere into distinct, interpretable groups. Rotation does not change how many factors you have or the total amount of explained variance; it just redistributes the loadings to improve clarity. It also doesn’t estimate correlations among observed variables—that part comes from the initial extraction and the correlation structure of the data. Rotation can be done so factors are kept uncorrelated (orthogonal) or allowed to correlate (oblique), depending on whether you expect the underlying constructs to be related.