Different disciplines have different standards/customs for reporting. I would suggest check some journals in your field to see what is customary.
I would suggest describing the whole reasoning process you are using the regression to model. You should interpret the coefficients at each stage of the reasoning process. You should at a minimum report the zero order correlations among the DV and IVs and the coefficients at the final step. If you have enough cases, you might want to first hypothesize about interactions and then test them. You may want to go further and explore effects that you did not hypothesize about but if you do be sure to specify to your readers that those portions of the analysis are exploratory. How is the construct of parent involvement distinct from the construct of family support? . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
