Situational judgments tests for selection:traditional vs construct-driven approaches

Context Historically, Situational Judgement Tests (SJTs) have been widely used for personnel selection. Their use in medical selection in Europe is growing with plans for further expansion into North America and Australasia in an attempt to measure and select on ‘non-academic’ personal attributes. However, there is a lack of clarity regarding what such tests actually measure and how they should be designed, scored and implemented within the medical and health education selection process. In particular, the theoretical basis from which such tests are developed will determine the scoring options available, influencing their psychometric properties and, ultimately, their validity. Objective The aim of this article is to create an awareness of the previous theory and practice that has informed SJT development. We describe the emerging interest in the use of the SJT format to measure specific constructs (e.g. ‘resilience’, ‘dependability’ etc.), drawing on the tradition of ‘individual differences’ psychology. We compare and contrast this newer ‘construct-driven’ method with the traditional, pragmatic approach to SJT creation, often employed by organisational psychologists. Making reference to measurement theory, we highlight how the anticipated psychometric properties of traditional versus construct-driven SJTs are likely to differ. Conclusions Compared to traditional SJTs, construct-driven SJTs have a strong theoretical basis, are uni- rather than multidimensional, and may behave more like personality selfreport instruments. Emerging evidence also suggests that construct-driven SJTs have comparable predictive validity for workplace performance, although they may be more prone to ‘faking’ effects. It is possible that construct-driven approaches prove more appropriate at early stages of medical selection, where candidates have little or no healthcare work experience. Conversely, traditional SJTs may be more suitable for 3 specialty recruitment, where a range of hypothetical workplace scenarios can be sampled in assessments.