Optimizing Differences Between Groups is Important

Abstract Firms rely on aggregate responses from engagement survey data in order to differentiate between teams of direct reports under specific managers, divisions, business units, etc. In this chapter, the authors show that items vary with respect to being able to differentiate groups and provide concrete and practical information relevant to selecting items used in engagement surveys. Specifically, the authors show how item-level intraclass correlation coefficient type 1 (ICC[1]) values can provide useful information to optimize group differentiation beyond traditional psychometrics. Using multiple examples from applied settings, they provide insights regarding interpreting ICC(1) values, wording of survey items, stability of ICC(1) values, and the difference between ICC(1) values and traditional psychometric values.