st: Harman One Factor test

[Marco Ercolani <m.g.ercolani@bham.ac.uk>]

Dear Statalisters,

I wish to learn how to apply a simple "Harman One Factor test" to
identify "common methods variance" (and do so in Stata not SPSS).

I know that the use of this test has been discouraged in a previous thread:
www.stata.com/statalist/archive/2012-04/msg00582.html
but let us assume I wish to learn it for didactic reasons.

For argument's sake, let the measured (log) wage be made up of 3 parts:
lnWage = Trait_Variance + Common_Method Variance + Error_Variance
and we wish to check if Common Method Variance is a significant feature
in a wage regression.

To make things more concrete, let us assume that I set up the following
Stata dataset and analysis:
-begin program---------------------------------------------------------
webuse nlsw88.dta, clear
generate lnWage=ln(wage)
regress lnWage age married collgrad south c_city union hours tenure
* Do I then run the “factor” command? How do I interpret the results?
factor lnWage age married collgrad south c_city union hours tenure, pcf
-end program-----------------------------------------------------------

Please tolerate my ignorance on this matter, hence the post.
Simple explanations most welcome.

sincerely,
 Marco

Dr Marco G. Ercolani
Department of Economics
University of Birmingham
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/