How do we know who should receive what intervention? Or who will react in a
certain way to a given stimuli? Psychology deals with many matching effects
whereby individuals have different reactions contingent on their
personalities, and these differences are typically operationalized as
interaction effects. However, popular analyses, such as examining whether
an interaction term is significant and doing simple slopes analyses (e.g.,
looking for differences at +/- 1 SD on a continuous variable) tell us
little about when people shift in their reactions. We therefore propose the
direct estimation of Points of Intersections (POIs) between regression
slopes; thresholds along continuous variables (e.g., personality) when
categorical variables (e.g., stimuli) shift in the direction of their
effects. Using Monte Carlo simulation studies, we evaluate competing ways
of calculating POIs and confidence intervals around POIs, ultimately
recommending a bootstrap procedure.
--
*Keven Joyal-Desmarais*
PhD Candidate, Social Psychology
Social Sciences and Humanities Research Council (SSHRC) of Canada Doctoral
Fellow
Administrative Assistant, Center for the Study of the Individual and
Society (CSIS)
University of Minnesota | N369 Elliott Hall | Minneapolis, MN 55455