ABSTRACT
While the degree to which any selection method can effectively forecast workplace performance is its most important property, the acceptability of the methods to candidates, administrators and other stakeholders is also incredibly important. Self-reported levels of personality characteristics such as conscientiousness are as good as A-Level grades at forecasting future performance. Nevertheless, psychologists and psychometricians get themselves into much controversy about all of this even the great ones. The aim of this talk is to explore prediction, decision making and cost-benefit in selection methods. At its core, this is about the bottom line for an organisation. One of the biggest mistakes people make is to confuse forecasting/prediction errors and decision errors. Cronbach and Gleser showed that cost-benefit is actually linear to the validity coefficient, not its square. Cronbach found that any financial gain at all from an assessment method easily overshadowed the cost of testing, which could be lower than $100 per person tested.
