An evaluation of student selection in a physician assistant program



Journal Title

Journal ISSN

Volume Title



The research on the selection of physician assistant students has mainly been designed to predict student success and little is known about the predictive validity of this research in the selection process. This study sought to isolate characteristics of applicants to a physician assistant program over a five year period and thereby evaluate the selection process. The rationale for this study was based upon certain factors which affect the admissions process in health professional programs in general and physician assistant programs in particular. The selection of students for health professional programs is complicated by a large number of applicants which necessitates effective and efficient screening in order to select students with the highest probabilities of successfully completing the programs. The cost of educating health care professionals is great, and the educational programs cannot afford to lose students to academic and nonacademic attrition. Validation of selection criteria in educational programs is required by the Civil Rights legislation of the 1960s and 1970s. The traditional cognitive criteria used by most health professional programs have only had moderate success in predicting student performance; therefore, many researchers have sought substantial predictors of success in the noncognitive realm. The physician assistant program used in the study was the Baylor College of Medicine program in Houston, Texas. The first class was admitted in 1971. The selection criteria have changed over the seven years of applicant selection. Health care experience was heavily emphasized for the first several years. Cognitive criteria were first used in 1972 and have had greater emphasis in recent years. The subjects of the study were selected from the applicant pools for each year from 1973 through 1977. Twenty applicants and twenty students were randomly selected from each of the five years included in the study. The total sample included 200 subjects. In order to evaluate the physician assistant student selection process, two null hypotheses were generated to guide the study. Three independent variables and eighteen dependent variables were chosen to represent characteristics of the applicant pool. The independent variables included (1) matriculation status, accepted or not accepted into the program; (2) sex, male or female; (3) year of application to the program, 1973 through 1977. The dependent measures represented psychological, cognitive, and biographical characteristics of the subjects. Seventeen of these were scales chosen from the Survey of Study Habits and Attitudes, the Personal Orientation Inventory, and the Nelson-Denny Reading Test. The other dependent variable was age. The dependent variables were factor analyzed using alpha and principal components models to reduce the number of variables. The factors were the dependent variables in a multivariate analysis of variance. Post-hoc analyses were planned for any significant effects. Three effects in the multivariate analyses of variance were significant. These included the main effects of sex and year of application and the interaction effect of matriculation status by year of application. The only substantial contributor to the difference in mean vectors in both the year of application main effect and the sex main effect was reading ability. Educational orientation and reading ability were the only two major contributors to the significant difference in mean vectors in the matriculation status by year of application effect. Reading ability, the only cognitive measure in the study, consistently differentiated among groups in the three significant effects. The educational orientation factor differentiated among applicants and students by year of application. Age and the two factors representing the Personal Orientation Inventory scales did not contribute to the differences in any of the effects. Therefore, none of the psychological characteristics in this study differentiated among the subjects.