Chapter 9: Factor Analysis

Question 1

Which are the three ways of obtaining factor scores?

Answer 1

Unweighted least squares, weighted least squares and regresison.

Question 2

How to obtain factor scores using the WLS method?

Answer 2

Minimize sum of squares of errors, weighted by the reciprocal of their variances. From there, normal equations solution.

Question 3

In simple term, what do the loadings represent?

Answer 3

They are the correlations with the unobserved factors.

Question 4

What should be the first thing to look at before deciding on performing factor analysis?

Answer 4

The correlation matrix; if only a few or no instances of high correlation, factor analysis won’t be fruitful.

Question 5

What is the purpose of rotating the loading matrix (L)?

Answer 5

To (hopefully) produce factors with a mix of high and low loadings and few moderate-sized loadings, and to achieve easier interpretation.

Question 6

For standardized data, do we use the covariance or correlation matrix?

Answer 6

Correlation.