Yea-Saying, Nay-Saying, and Going to Extremes: Black-White Differences in Response Styles

Abstract

This paper reports an important racial difference in patterns of responding to Likert-type items: blacks are more likely than whites to use the extreme response categories, particularly the positive end of agree-disagree scales. This finding appeared consistently in several large-scale nationally representative surveys of youth. Response style indexes (Agreement, Disagreement, Acquiescence, and Extreme Responding) display ranges of individual differences and cross-time stabilities comparable to commonly used personality measures. Response styles show no consistent differences by sex, and no appreciable correlation w;ith indicators of family socioeconomic level or with educational accomplishments and aspirations. For both races, agreement tendencies are stronger among those in the South, especially in nonmetropolitan areas; however, controlling for geography does little to reduce overall black-white differences. The findings reveal potential pitfalls in dealing with racial differences in survey and personality measures, and illustrate the need for great caution in reporting and interpreting such differences. Jerald G. Bachman is Program Director and Patrick M. O'Malley is Senior Study Director at the Survey Research Center of the Institute for Social Research, The University of Michigan. The authors thank Pamela R. E. Kittel for her assistance in the preparation of this report. Public Opinion Quarterly Vol 48:491-509 ? 1984 by the Trustees of Columbia University Published by Elsevier Science Publishing Co., Inc. 0033-362X/84/0048-491/$2.50 This content downloaded from 157.55.39.138 on Mon, 27 Jun 2016 06:05:10 UTC All use subject to http://about.jstor.org/terms 492 JERALD G. BACHMAN AND PATRICK M. O'MALLEY others are more likely to use the middle values (e.g., mostly agree, mostly disagree) (Hamilton, 1968). As Couch and Keniston pointed out, one perspective is to view such response styles as primarily statistical nuisance that must be controlled or suppressed by appropriate mathematical techniques, but another perspective treats response styles as a manifestation of deep-seated personality syndrome (1960:151). Both perspectives are relevant for the present study. For several years we have been aware of some fairly substantial differences between black and white high school seniors in their patterns of responding to Likert-type questionnaire items. Specifically, blacks are more likely than whites to agree in response to agree-disagree items, and to use the extreme ends of response scales. We did not originally set out to search for these differences in response styles; rather, we discovered them serendipitously in the process of making more general comparisons of blacks and whites across wide range of questions tapping attitudes, opinions, and personality dimensions. But once aware of the overall differences in response distributions, we considered it worthwhile to investigate them further. The present paper, based on those further investigations, is intended to document the extent of black-white differences in response styles, to explore some possible explanations, and to illustrate some special problems and pitfalls facing those who analyze and report racial differences in attitudes and personality.