Hello all,
I am trying to examine an indirect model (with several moderators between
the indpendent variables and the dependent variable).
All data comes from questionnaires I handed employees, their employers and
from company records (no laboratory experiments etc.).
Going over some articles that used indirect models, I understand that I
should use the linear regression command in SPSS (v. 10), enter in the
first block my control variables (demographics, group size etc.), then
choose the "Enter" method, enter the indpendent variables in the second
step, choose the "ENTER" method again, and, finally, enter the hypothesized
mediating variables (and choose "ENTER").
Am I right? How should I interpert the results (i.e. what will prove my
hypothesized mediating effect)? I'm afraid the control variables have a
major influence over the dependent variable, should it realy be entered in
the first step, or maybe last (I've seen the latter technic somewhere).
Thank you for any help or reference,
Eric Lifschitz (An inexperienced graduate ob student from Israel - sorry
about the Awkward "Statistical" English...).
|