Regency EF
Exploring 2- and 3-Way Interactions in Regression Models with R and RFSA (304011)
*Joshua Lambert, University of CincinnatiKeywords: Interactions, regression, effect, modification, rFSA,
A significant statistical interaction could exist when the effect of a variable changes based on the specific presence or value of another variable or variables. Interactions in a regression models are sometimes overlooked as usually the investigator had no knowledge of their potential existence or there were simply too many interactions to check. This poster will present a novel new tool for exploring interactions in large datasets where complete enumeration is impractical. This tool, rFSA, is highly flexible R package that can handle different modeling strategies (linear, logistic, cox, survey weighted regression, etc.), criteria functions, and can give multiple interactions to check for inclusion. Also, this poster will give tips and tricks about interactions. These include: an overview of when interactions might exist, how to plot an interaction effect, and how do I interpret a multiple linear regression model with an interactions?