Stakeholder perceptions of selection: analysis of applicants' and stakeholders' perceptions of methods used to select medical students in undergraduate medical education

Examining personnel selection solely from an organisational perspective has traditionally been the dominant paradigm within Industrial/Organisational (I/O) psychology (Anderson, Lievens, van Dam, & Ryan, 2004). However, in recent years there has been a move to explore selection from the applicants' perspective. This thesis extends the existing literature by examining applicants' as well as other key stakeholders' perceptions of the validity of the methods used in the high-stakes context of medical student selection. Ryan and Ployhart's (2000) comprehensive model of applicant perceptions was applied in three empirical studies. The first and second studies examined how applicants and non-applicant stakeholders respectively, perceive the validity of the methods used to select medical students into an initial degree, and to identify factors that influence perceived validity. The third study aimed to determine whether individuals' perceptions of each selection method's validity would impact on overall confidence in the selection process. The influence of moderating variables on this relationship was also explored. Results showed that interviews were perceived as the most valid selection method by all participant groups, with cognitive ability tests and references rated as the least valid. Perceived validity was predicted to the extent participants thought the selection methods were susceptible to ethnic discrimination, by the type of skills they thought doctors required, familiarity with the methods, and demographic factors. Perceived validity of cognitive ability and measures of academic performance were significant predictors of applicants' confidence that the best people are being selected as future doctors, while non-applicant stakeholders' confidence was affected by perceived validity of the test of cognitive ability and of interviews. Moderation analyses provided support for the hypothesis of a differential effect for applicants and non-applicants regarding how a method's perceived susceptibility to artificially enhance performance would influence the relationship between its perceived validity and confidence. Strengths and limitations of this thesis are presented, together with implications for practice and future research.