Abstract
Net of controls for educational credentials, recommendations, age, high school quality, employment sector, firm size and region, white personnel officers tend to assign black male high school graduates to lower paying positions than those assigned to white male high school graduates in the USA. Similar patterns are observed for while female college graduates. The effect of job candidates' race on employers' job placement decisions is examined, using data gathered by the randomised vignette technique. These patterns of apparent bias in job placement are found to be offset to some degree in firms with affirmative action policies. The findings are discussed in the context of Thurow's 1975 theory of statistical discrimination. Further research is needed to investigate potential discrimination in job selection and to examine characteristics of firms and personnel officers with the greatest propensity to discriminate.