Buros-Nebraska Series on Measurement and Testing
Date of this Version
Spring 1996
Document Type
Article
Citation
Published in Multicultural Assessment in Counseling and Clinical Psychology, edited by Gargi Roysircar Sadowsky and James C. Impara (Lincoln, NE: Buros Institute of Mental Measurements, University of Nebraska–Lincoln, 1996).
Abstract
Stanley Sue in "Measurement, Testing, and Ethnic Bias: Can Solutions Be Found" addresses multicultural assessment and research with experienced wisdom and scientific inquiry. His tone is amicable, communicating a problem-solving attitude. Owing to its applicability, Sue's paper will appeal to a wide readership, with each reader finding a particular part especially meaningful. We find journalistic information on negligent diagnosis; a review of diagnostic studies; suggestions for new measurement methods to control for cultural bias in tests; analyses of a White prediction equation for the academic achievement of various Asians in the U.s.A.; ongoing research on MPPI-2 scores of diversely acculturated Asian Americans; hypotheses about Asian-American personality variables that influence responses to mainstream measures of psychopathology; and a discussion on institutional policy matters, something practitioners are rarely concerned about, but which is important to the advocacy of racial and ethnic equity.
One is introduced to what is minority group status, culture, ethnicity, and the overlap of the latter two. Sue cites research where substantial misdiagnosis of American ethnic minorities consists of both over- and underpathologizing, and where misdiagnosis may have resulted from the interaction of client-clinician racial! ethnic match and mismatch. The main point is that American ethnics are more likely to be misdiagnosed than White Americans. Sue notes that the two popular ways of identifying test bias in personality instruments are factor analysis and regression analysis (analysis of items within an instrument has been used particularly in achievement and aptitude tests [Sue, 1994, private communication]).
Sue addresses the nature and extent of bias when one group's regression equation is used as the standard. He summarizes a previous study that reports predictors of Asian academic achievement. A White regression equation both overpredicted and underpredicted various Asian groups. Sue and colleagues used Whites as the standard because prediction formulas established by universities are based primarily on the White-American majority group. Sue emphasizes (1994, private communication) that "over and underpredictions of CPA involving a difference of .17 is quite substantial, not only to student perceptions but also to admissions to graduate school. As one example, UCLA will not as a rule admit as graduate students undergraduates who have a cumula tive CPA of under 3.00. You can imagine how many students receive CP As between 2.83 and 3.00 .... Finally, at some universities (such as UC Berkeley), there were attempts to increase the weight of SAT-Verbal over SAT-Math performance in admission. According to our findings, doing so would probably reduce the ability to identify the best Asian American students."
Comments
Copyright © 1996 by Buros Institute of Mental Measurements. Digital edition copyright © 2012 Buros Center for Testing.