No abstract

This paper describes the application of profile similarity metrics to score Situational Judgment Tests (SJTs) that utilize rating scales to register examinee responses. The paper presents and discusses mathematical analyses that decompose distance-based measures into component indices based on correlation, dispersion and elevation metrics. The mathematical analyses demonstrate that distance measures represent a mixture of variance that can be associated with these separate components. Comparing the validities of distance and component indices using Leader Knowledge Test (LKT) data supports conclusions that the use of component indices (i.e., correlation, dispersion and elevation scores) improves the validity of SJTs that utilize rating scales.

No abstract

No abstract

No abstract