Examining the Original Grasmick Scale

Abstract

The original Grasmick, Tittle, Bursik, and Arneklev self-control scale is frequently used in self-control theory studies. Examining self-control requires precision in measurement. Using the Rasch rating scale model analysis, the Grasmick et al. scale is evaluated for construct validity in a college student sample. The results show that the 4-point Likert-type format response categories are not suitable for the scale. The distribution of the measures does not align with the distribution of the item measures in the original Grasmick et al. scale. The items do not fit the Rasch model, they do not form a unidimensional trait, and they function differently for males and females. Thus, the original Grasmick et al. scale does not have construct validity. A revised scale is presented that meets the standards of the Rasch model.