Possible Duplicate:
Should I run separate regressions for every community, or can community simply be a controlling variable in an aggregated model?
I have a SPSS model output with fixed effects. The dependent variable is pay, and I'm testing for the difference between males and females including several variables and interaction terms, e.g. male*years_employed.
How would running a second analysis as two separate regressions, one for male and one for female, differ from my interaction analysis, in terms of tests for expected effects for males and females, and tests for differences between male and female effects? When is it better/more appropriate to use one or the other?
