Brian O'Connor   UBCO Psychology   UBCO  

SIMPLE:

All-in-one programs for exploring interactions in moderated multiple regression


Reference:

O'Connor, B. P. (1998). All-in-one programs for exploring interactions in moderated multiple regression. Educational and Psychological Measurement, 58, 833-837.


Upon discovering an interaction in moderated multiple regression, users must conduct time-consuming simple slope analyses that are not performed by current statistical software programs. This paper describes SAS and SPSS programs that provide complete simple slope statistics and plots all in one job run. Programs are available for two and three-way interactions, and for continuous and categorical moderators.

The programs were revised in May, 2005.

Click on a program to view its contents. Copy and paste into a file.


SPSS: SAS:
Simple-2way Simple-2way For interactions between two continuous 
predictors, or for interactions between a 

dichotomous IDV and a continuous moderator
Simple-3way Simple-3way For 3-way interactions with two continuous
moderators and a continuous or dichotomous IDV
Simple-2grps Simple-2grps For interactions between a continuous IDV 
and a 2-category moderator
Simple-3grps Simple-3grps For interactions between a continuous IDV 
and a 3-category moderator
Simple-4grps Simple-4grps For interactions between a continuous IDV 
and a 4-category moderator
Simple-5grps Simple-5grps For interactions between a continuous IDV 
and a 5-category moderator
Simple-6grps Simple-6grps For interactions between a continuous IDV 
and a 6-category moderator

All of the programs contain commands (near the top) that generate artificial data that can be used for trial runs. These commands should of course be removed for analyses using real data.

Thanks to Asier R. Larrinaga, Dept. de Biolox'a Celular e Ecolox'a, A Coru–a, Spain, for help with Simple-6grps.sps program and with the simultaneous regions of significance computations.

Thanks to Maarten Vissers, Dept. of Communication Studies, University of Twente, The Netherlands, for help with the simultaneous regions of significance computations.

For more information on the programs, see the Program Notes.

Brian P. O'Connor
Department of Psychology
University of British Columbia - Okanagan
Kelowna, British Columbia, Canada
brian.oconnor@ubc.ca