Effects of academic and psychosocial predictors of performance in medical school on coefficients of determination.

Abstract

This study was designed to answer the following questions: (1) Does a set of selected noncognitive variables predict medical school performance measures? (2) Is there a significant increase in the coefficients of determination when noncognitive measures are added to the conventional cognitive predictors in regression models? Complete data on all measures were available for 88 sophomore medical students. Cognitive (academic) predictors were undergraduate GPA in science and nonscience courses, and scores on science problems, reading and quantitative scales of the Medical College Admission Test. Noncognitive (psychological) predictors were scores on scales of stressful life events, general anxiety, test anxiety, emotionality, external locus of control, intensity and chronicity of loneliness, sociability, self-esteem, perception of early relationships with mother, father and peers, and indices of over- and underconfidence. Criterion measures were freshman and sophomore GPAs in medical school and scores on Part I of the National Board examinations. Results indicate that (1) noncognitive predictors could significantly predict criterion measures and (2) inclusion of noncognitive measures in a model of cognitive predictors could substantially increase the magnitude of the relationships.