Issues of validity and reliability in the use of real–ideal discrepancy scores to measure self-regard.

Abstract

The validity and reliability of real-ideal discrepancy scores as measures of selfregard are investigated using a large sample of adolescent subjects. Real selfratings are found to be superior to real-ideal discrepancy ratings in predicting several measures of global self-esteem and in predicting several measures of teachers' evaluations of the self-attitudes of the subjects. Further, similar-sized discrepancies do not translate into similar levels of self-regard at different points on scales of real self-evaluation, and the more dependent the discrepancy score is made on the real self-score relative to the ideal self-score, the more effective a predictor it is of global self-esteem. Finally, it is shown that the test-retest correlations are quite low for a real-ideal discrepancy scale. Researchers are cautioned against the use of such measures. Discrepancy scores between self-ratings and ideal self-ratings on lists of descriptive adjectives and self-descriptions have been widely used in research. The discrepancy scores have been seen as indexes of self-acceptance, personal adjustment, self-regard, self-esteem, and cognitive development. The most widely used measures come from different theoretical traditions. The Butler and Haigh Q sort (Butler & Haigh, 1954) stems from the work of Rogers and Dymond (1954), and the Interpersonal Checklist (LaForge & Suczek, 1955) stems from the theoretical work of Leary (1957) and his co-workers. Other widely used measures include the Index of Adjustment and Values (Bills, Vance, & McLean, 1951) and the Self-Activity Inventory (Worchel, Note 1). Many researchers