Presenter(s)

• Aidan Campagnolio, PhD Student, Temple University School Psychology Program
• Jessica Harris

Abstract

The simple correlation coefficient, directly or indirectly, is at the heart of most statistical procedures ranging from t-tests to multiple regression to factor analysis to structural equation modeling. The number of possible relationships between any two variables is infinite, but the correlation coefficient quantifies only one of them: the linear relationship. With a fair bit of ingenuity, however, statisticians have invented procedures that allow simple correlations to measure a wide variety of non-linear relations. In the school psychology literature concerning the assessment of academic abilities of individuals, almost all analyses remain strictly linear. When non-linear relations are modeled as if they were linear, predictions and interpretations become distorted, and if severe enough, the distortion can interfere with accurate case conceptualization.