What have the difference scores not been telling us? A critique of the use of self--ideal discrepancy in the assessment of body image and evaluation of an alternative data-analytic framework.

Difference scores are often used as a means of assessing body image satisfaction using silhouette scales. Unfortunately, difference scores suffer from numerous potential methodological problems, including reduced reliability, ambiguity, confounded effects, untested constraints, and dimensional reduction. In this article, the methodological problems are outlined and an alternative framework is discussed. The alternative consists of a minimum of testing the constraints implied by the difference score model, and at most evaluating more exact body image hypotheses by incorporating nonlinear terms in a regression and testing features of the response surface. Two empirical examples are used to illustrate the utility of these methods. The first example uses both polynomial regression and response surface methods to examine eating disorder outcomes, whereas the second example uses polynomial regression to examine the outcomes related to muscle dysmorphia. Directions for future research related to assessment of body image are discussed.