Psychological tests measuring general intelligence using progressive matrices are often considered “markers†for Spearman’s g and are frequently used in the world of psychological assessment (Jensen, 1998). This study examines age differences in scores from a matrices test developed for and used in employee selection for medium to high complexity jobs. The sample of (N=24,116) was collected between year 2011 and 2014 in high stake selection processes conducted in Sweden. Main part of the data was collected in the private sector. The results show that age have a curvilinear effect on test scores. This also applies when gender and education are held constant. Differences in age between high scorers in the youngest age group (15-30 years) compared with the low scorers in the oldest age group (61-75 years) is relatively high: d = 0.83, 95% CI [0.70, 0.96]. Implications for employee selection are discussed and suggestions are given how the effect of adverse impact can be reduced by using mixed tests combining non-verbal with verbal test items, or by using age-specific norm groups.