Aptitude-by-Treatment Interactions in Research on Educational Interventions

A common theme uniting articles in this special issue is a focus on aptitude-by-treatment interactions (ATIs). This timely and welcome focus allows the field to synthesize current substantive findings on ATIs in educational intervention research in both reading and math domains. In this methodological commentary, we begin by reviewing traditional approaches for detecting and reporting interactions in single-level and multilevel models. Next, we discuss some limitations of traditional approaches for theorizing about and modeling ATIs, and we suggest some solutions. These solutions include interpreting level-specific (unconflated) ATIs, understanding and ameliorating threats to adequate power for detecting ATIs, expanding focus beyond linear ATIs, and increasing the number of measurement occasions beyond two to allow use of a growth modeling framework for investigating ATIs. Incorporating some of these advances into future research can motivate new research questions about educational interventions and lead to new discoveries in the search for ATIs.