Re: st: factor analysis - calculating Crohnbach's alpha

["JVerkuilen (Gmail)" <jvverkuilen@gmail.com>]

Being one of the resident psychometricians, I can give a more complete
answer later but Cronbach's alpha for an unweighted total score in
Stata is given by -alpha-. It has quite a number of options and is
probably what you're looking for.

Principal components of the correlation matrix gives a different
reliability coefficient, Armor's theta, which is defined to be

       theta = [Q/(Q-1)] (1 - 1/maxeigenvalue)

This is the reliability coefficient corresponding to the first
principal component.

If you use factor analysis to determine weights, you can compute
McDonald's omega:

      omega = (sum of factor loadings)^2 / (sum of uniquenesses + (sum
of factor loadings)^2)

This works for both correlation and covariance matrices.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/