BUS 481 Chapter Notes - Chapter 11: W. M. Keck Observatory, Sony Max, Railways Act 1921

• School: Simon Fraser University
• Department: Business Administration
• Course: BUS 481
• Professor: Peter Charlton

Document Summary

To be considered useful, must add value to prediction of job success. To add value, must go beyond forecasting powers of current predictors. Predictor more useful the smaller its correlation w/ other predictors & the higher its correlation w/ criterion. If correlations b/t new & existing predictor higher than correlations b/t new predictor. & criterion, then new is not adding much value. Multiple regression take correlation among predictors into account. Predictors likely to be correlated when content is similar. Predictor discriminates b/t ppl in terms of likelihood of success on job. May also discriminate disproportionate # of designated members. To the extent this happens, predictor has adverse impact & may result in legal problems. When validity of predictors is same, but 1 predictor has less adverse impact than the other, use predictor w/ less adverse impact. Balancing trade-offs difficult if 1 predictor has high validity/high adverse impact while another has low validity/low adverse impact.