Research studies have repeatedly found that group simulations are one of the best predictors of job performance. These results have held across a range of positions and industries. There is a number of reasons to expect that simulations should be valid predictors of job performance. First, they are built on the premise of point-to-point correspondence, which suggests that prediction is improved to the extent that predictors mirror the criterion (Asher & Sciarrino, 1974). Second, unlike many predictors (other than tests of cognitive ability) simulations are ‘show me’ and not ‘tell me’ measures. That is, we do not ask the applicant to tell us how good they are at dealing with a problem or scenario; we given them one to solve. This provides two benefits: mitigating inflation and response distortion as applicants cannot fake proficiency and requiring a behavioural response that yields a more accurate measure of actual ability (Gatewood, Field & Barrick, 2008). In this section, we summarize the current state of the literature on simulation validity and discuss subgroup mean differences.