Use of Z-scores to rank applicants to professional degree programs

Criteria for assessing suitability of applicants for professional degree programs such as veterinary medicine are usually treated as distinct components of a composite scoring procedure that determines applicant ranking. Some components are valued more than others, which is reflected in the relative weights assigned to each component. However, the patterns of dispersal of individual components have the potential to alter the assigned relative weights. Components with larger variances can have greater influences on composite scores than intended. Such unintended altered weighting can be avoided through standardization. Yet non-standardized approaches continue to be used for admissions ranking in several programs. In this study, we documented the potential for differential selection of applicants when non-standardized scoring approaches are applied to admissions assessment components. At our medical school, applicants' component scores with differing variances are standardized by determining Z-scores with a mean of 0 and standard deviation of 1 before mathematically combining to calculate composite scores and admissions ranking. We retrospectively and hypothetically ranked one applicant cohort using non-standardized methods and identified differences in ranking between the standardized and non-standardized approaches. Most differences were observed for applicants in the second, third, and fourth quintiles of the admissions rank list, that is, those for whom admissions cut-off decisions make a marked difference. Observations were supported by lower Spearman's rank correlation coefficients in these quintiles. Although standardization of component scores is not a novel topic, we document the implications of using non-standardized scoring approaches for applicant ranking and underscore the importance of standardization of component scores.