Which method yields factor scores that are uncorrelated and standardized with a mean of 0 and a standard deviation of 1?

Which method yields factor scores that are uncorrelated and standardized with a mean of 0 and a standard deviation of 1?

• A. Bartlett method
• B. Anderson-Rubin method
• C. Principal Component Analysis
• D. Regression scoring

Answer: (B)

When estimating factor scores, you can choose a method that shapes the relationship among the scores as well as their scale. The Anderson-Rubin approach constructs factor scores so that the estimated scores for different factors are uncorrelated (orthogonal) and have unit variance, which means each score has a mean of zero and a standard deviation of one across individuals. This combination—zero correlations between factors and standardized scaling—makes the scores easy to interpret as independent latent traits and convenient for subsequent analyses. Other methods don’t guarantee both properties at once. Bartlett scoring aims to minimize bias in reproducing the correlation structure but doesn’t ensure unit variance, while regression scoring provides the best linear predictors of the factors given the observed data but can produce correlated factor scores when the underlying factors are related. Principal Component Analysis yields orthogonal components stemming from variance maximization, but it isn’t the same as the factor-score approach used in exploratory factor analysis.

When estimating factor scores, you can choose a method that shapes the relationship among the scores as well as their scale. The Anderson-Rubin approach constructs factor scores so that the estimated scores for different factors are uncorrelated (orthogonal) and have unit variance, which means each score has a mean of zero and a standard deviation of one across individuals. This combination—zero correlations between factors and standardized scaling—makes the scores easy to interpret as independent latent traits and convenient for subsequent analyses.

Other methods don’t guarantee both properties at once. Bartlett scoring aims to minimize bias in reproducing the correlation structure but doesn’t ensure unit variance, while regression scoring provides the best linear predictors of the factors given the observed data but can produce correlated factor scores when the underlying factors are related. Principal Component Analysis yields orthogonal components stemming from variance maximization, but it isn’t the same as the factor-score approach used in exploratory factor analysis.