Journal of Applied Psychology Copyright 1998 by the American Psychological Association, Inc. 1998, Vol. 83, No. 3, 438-451 0021-9010/98/$3,00 Can Racial Differences in Cognitive Test Performance Be Reduced by Presenting Problems in a Social Context? Richard E DeShon and Matthew R. Smith Michigan State University David Chan Michigan State University and National University of Singapore Neal Schmitt Michigan State University It is well documented that African Americans tend to score lower on cognitive ability tests than White Americans. The cause of this performance difference is still being hotly debated and actively researched. Recently, J. E. Helms (1992) argued that cognitive ability tests fail to adequately assess African American intelligence because they do not account for the emphasis placed on social relations and the effect of social context on reasoning in the African American culture. To evaluate this argument, racial subgroup performance differences on Wason conditional reasoning items presented in the standard abstract form and in a social relationship form were examined. Contrary to Helms's argument, results indicate that presenting Wason conditional reasoning items in a social context did not reduce the difference between African American and White American performance. The empirical evidence that African Americans tend to score lower on cognitive ability tests than White Ameri- cans is well documented and widely accepted (Hartigan & Wigdor, 1989; Neisser et al., 1996; Schmidt, 1988). The causes of this performance difference, however, are still being vigorously debated. Some individuals have argued that the empirical research suggests that the performance differences in cognitive ability test performance between African Americans and White Americans reflect real dif- ferences in intelligence between the two groups (e.g., Jensen, 1985). However, the sources of these "real" dif- ferences are unclear (Neisser et al., 1996). Others have argued that subgroup performance differences on cogni- tive ability tests are due to bias in the assessment of intel- lectual ability (e.g., Helms, 1992) or pervasive societal discrimination (Mercer, 1984). Whatever the cause, the Richard E DeShon, Matthew R. Smith, and Neal Schmitt, Department of Psychology, Michigan State University; David Chan, Department of Psychology, Michigan State University, and Department of Social Work and Psychology, National Uni- versity of Singapore. Matthew R. Smith is now at AON Con- sulting, Grosse Pointe, Michigan. Correspondence concerning this article should be addressed to Richard P. DeShon, Department of Psychology, Psychology Research Building, Michigan State University, East Lansing, Michigan 48824-1117. Electronic mail may be sent to deshon@ pilot.msu.edu. 438 empirical finding of subgroup differences in cognitive ability tests has serious practical implications for human resource practices in organizations. Therefore, it is im- portant to identify assessment methods that reduce the observed subgroup performance differences while main- taining optimal prediction of organizationally and soci- etally relevant variables. A common explanation for the observed differences between racial subgroups in cognitive ability test perfor- mance is that members of different ethnic cultures are exposed to different information and values, which bias the assessment of cognitive ability. TWo distinct research directions have been developed in response to arguments that subgroup performance differences are caused by cul- tural differences in the exposure to the concepts on ability tests. First, a number of researchers attempted to reduce the potential for bias in the measurement of cognitive ability due to differential exposure to specific values and linguistic concepts by developing abstract, knowledge- free measures of ability (e.g., Cattell & Cattell, 1959; Raven, 1965). However, these abstract, knowledge-free measures of ability tend to yield some of the largest per- formance differences between African Americans and White Americans. Second, in contrast to the measurement of ability through abstract reasoning tasks, an increasing number of cognitive and educational psychologists argue that exam- ining reasoning separate from the contexts in which peo-