Interaction Effects in Multiple Regression (Quantitative Applications in the Social Sciences)

Rating:
Summary: Excellent overview of interactions in regression
Review: This is a very nice updated version of the original edition. The book is short at 89 pages, information packed, and solidly grounded. The book is meant for people who *use* interaction effects in thier research. (It is not meant for people who study regression methods.)



STRONG POINTS: It references the key developments in the methodological literature, the most significant of which appeared in Psychological Bulletin. It even discusses in layman terms Chronbach's (of the Chronbach Alpha fame) recent paper in Psyc Bulletin on errors in interactions.



SCOPE: The book covers two-way and three-way interactions. It offers a digestable discussion of variable transformations. The authors clarify two of the biggest misperceptions about testing interactions: (1) It is incorrect to interpret the main effects in the presence of interaction terms and (2) Multicollinearity is rarely a problem with interaction terms if you appropriately transform the variables.



WHAT THE BOOK DOES NOT COVER: The book does not discuss (1) mediation testing and (2) is missing a key reference on that topic. If you're reading this, it will probably be helpful! [1] R. Baron and D. Kenny, "The Moderator-Mediator Variable Distinction in Social Psychological Research: Conceptual, Strategic, and Statistical Considerations," Journal of Personality and Social Psychology, 51 1986, 1173-1182.

Overall, the book is very well written, readable, and usable in doing interaction analyses. It is meant for the consumer not the producer of statistical methods. SPSS exemplars and detailed interpretations really help clarify the points. For fifteen bucks, this is worth owning, dog-earing, and highlighting. Very highly recommended.