RE: st: Presenting Interaction in LR

[Cameron McIntosh <cnm100@hotmail.com>]

I would also suggest looking at the following for additional guidance on how to examine and discuss your interaction effects:

Hayes, A.F., & Matthes, J. (2009). Computational procedures for probing interactions in OLS and logistic regression: SPSS and SAS implementations. Behavior Research Methods, 41, 924-936. http://www.personal.psu.edu/jxb14/M554/articles/Hayes&Matthes2009.pdf

Aguinis, H, & Gottfredson, R. K. (2010). Best-practice recommendations for estimating interaction effects using moderated multiple regression. Journal of Organizational Behavior, 31, 776-786.http://mypage.iu.edu/~haguinis/JOB2010.pdf

Shieh, G. (2011). Clarifying the role of mean centering in multicollinearity of interaction effects.  British Journal of Mathematical and Statistical Psychology, 64(3), 462–477.

Francoeur, R.B. (2011). Interpreting interactions of ordinal or continuous variables in moderated regression using the zero slope comparison: tutorial, new extensions, and cancer symptom applications.International Journal of Society Systems Science, 3(1/2), 137-158.

Cam
----------------------------------------
> Date: Thu, 8 Dec 2011 18:57:29 +0100
> Subject: Re: st: Presenting Interaction in LR
> From: maartenlbuis@gmail.com
> To: statalist@hsphsun2.harvard.edu
>
> On Thu, Dec 8, 2011 at 6:25 PM, Campo, Marc wrote:
> > We are running a Linear Regression model where we are interested in the effect of a dichotomous predictor (GROUP) on a ratio level rehabilitation outcome scale. It is a pre/post situation where we have adjusted for baseline values (BASELINE) of the scale. There is an interaction between GROUP and BASELINE. People at the low end of the scale do better in one group and at the higher end do better in the other (but not much of a difference in clinical terms).
> >
> > I was going to present the coefficients for GROUP at varying values of baseline. BASELINE however, is somewhat skewed and so when I calculate the coefficient for group at the mean and 1SD above and below the mean BASELINE value we end up with a lower coefficient that is fairly close to the minimum value. Are there other choices? 25, 50, 75Th percentiles? Graphs instead? I don't think transforming BASELINE adds enough to be worth the complexity and the relationship is fairly linear as is..
>
> Quartiles sound fine to me, but the most important thing is that the
> values make some substantive sense. There often are some values on a
> scale that are especially meaningful, e.g. 12 and 16 years of
> education in the American educational system (high school and college
> degree). In that case I would use those values.
>
> Hope this helps,
> Maarten
>
> --------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
>
> http://www.maartenbuis.nl
> --------------------------
>
> *
> * 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/
 		 	   		  
*
*   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/