Regions of Significant Criterion Differences in Aptitude-Treatment-Interaction Research

Three techniques (Johnson-Meyman, Potthoff, and Erlander and Gustavsson) for obtaining confidence regions for the difference between regression surfaces are compared. The techniques are potentially useful for analyzing data from aptitude-treatment-interaction research (ATI). The three techniques are found to yield substantially different sized confidence regions when applied to the same data, therefore possibly leading to different decisions in an aptitude-treatment-interaction study.