The Survey of Achievement Responsibility (SOAR): reliability and validity data on an academic attribution scale.

The Survey of Achievement Responsibility (SOAR) was designed to assess students' causal attributions of success and failure in school-related situations. Unlike other measures of locus of control and attributions, the SOAR simultaneously distinguishes among three broad subject areas--math/science, language arts/social studies, and physical education--and between two possible outcomes--success and failure. Data on the psychometric characteristics of the SOAR came from two samples from each of two large school districts, one suburban and one metropolitan. A five-stage study of the psychometric qualities of this new instrument revealed: (a) Content validity is high--judges independently classified items into the attribution categories intended by the authors; (b) internal consistency reliabilities are in a range adequate for group research purposes and are comparable to the reliabilities of other measures of attributions; (c) test-retest reliabilities showed attribution profiles to be stable over a 2-month interval; (d) correlations between the SOAR scales and students' self-reports of ability in various subject areas showed differential relationships in theoretically predicted directions; and (e) correlations between the SOAR scales and achievement tests were statistically significant, though low in magnitude, consistent with reports on other instruments. Finally, use of the SOAR in a spectrum of studies has demonstrated its research utility and indicated directions for further instrument development and research.

Use of objective examinations in medicine clerkships. Ten-year experience.

Student performance during the internal medicine clerkship at the University of Washington School of Medicine has been evaluated by clinical ratings and a written examination containing multiple-choice questions and patient management problems for the past 10 years. Measures of the correlation among the evaluation methods were determined by analyzing data from 1,544 students. The correlations of clinical ratings with the total examination score (r = 0.27), multiple-choice questions (r = 0.23), and patient management problems (r = 0.19) suggest that clinical ratings alone are not adequate for measuring student progress. The relationships of evaluation methods used in the medicine clerkship to other measures of performance such as selection to Alpha Omega Alpha and National Board examinations were also determined. The ability to predict student performance was enhanced considerably when the results of the clerkship examination were considered in addition to clinical ratings. These data suggest that a comprehensive assessment of student performance in medicine clerkships should include written examinations in addition to clinical ratings.

Predicting first-quarter test scores from the new Medical College Admission Test.

The predictive validity of the new Medical College Admission Test as it relates to end-of-quarter examinations in anatomy, histology, physiology, biochemistry, and "ages of man" is presented. It is recognized that the maximum predictive power is attenuated by the reliability of the criterion variables. To determine the value of attempting to increase the reliability in the criteria, the authors corrected the validity coefficients for attenutation. Regression analyses were also undertaken to examine the extent to which the new MCAT subtests can predict scores on end-of-quarter examinations. Results indicate that the Science Knowledge assessment areas of chemistry and physics and the Science Problems subtest were the most useful in predicting student performance, followed by the Skills Analysis: Quantitative and Skills Analysis: Reading subtests and the biology area of the Science Knowledge subtest.