# Talk:Interaction between categorical and continuous variables

### From PsychWiki - A Collaborative Psychology Wiki

Revision as of 19:18, 18 March 2011 by Flojo (Talk | contribs)

Centering the categorical variable in this case would also be necessary, unless you want the regression to return coefficients that apply to the population of people with a score of 0 on the categorical variable (and since I'm assuming one would code the example categorical variable 1,2, this would be meaningless). See this reference for a good discussion: Kraemer, H. C., & Blasey, C. M. (2004). Centring in regression analyses: a strategy to prevent errors in statistical inference. International Journal of Methods in Psychiatric Research, 13(3), 141-151. doi:10.1002/mpr.170