How and Why Standardized Tests Systematically Underestimate African-Americans' True Verbal Ability and What to Do about It: Towards the Promotion of Two New Theories with Practical Applications

Article excerpt

INTRODUCTION

In this Article, I want to raise a number of issues, both theoretical and practical, concerning the need for a total reassessment of especially the verbal intelligence of minority individuals. The issues to be raised amount to a critical reappraisal of standardized multiple-choice tests of verbal intelligence, such as the Law School Admissions Test ("LSAT"). I want to probe very deeply into why such standardized tests systematically underestimate verbal intelligence.

This leads me first to review the prospects for a new standardized test of verbal intelligence associated with the studies of Joseph Fagan and Cynthia Holland.1 These studies show us that the races are equal; this result leads us to question the construct validity of many current standardized tests of verbal aptitude. Then, I briefly review my own studies of standardized tests that suggest a systematic underestimation of the ability of minorities. My studies question not only the construct validity, but also the reliability of the scores used to assess individual test performance, especially for minority students and even White students from lower socio-economic strata. In order to correct some additional problems associated with standardized testing, I present in some detail a new theoretical model to explain the concept of "guessing' as it occurs on standardized tests. After assessing the empirical adequacy of this new guessing theory, I then apply it to help clarify why it is absurd to describe the test behavior of especially low scoring students as being due to "lucky guessing" especially when it comes to describing minorities' choices in response to hard test items. Properly assessing guessing behaviors provides us with another means to criticize the validity and reliability of scores reported for many current standardized tests. Next, I touch on student mentoring issues, the temporal aspects of testings,2 the structure of law school course work,3 and how these issues can affect the predictive validity of tests such as the LSAT. Finally, I review a central point of contention for this conference: namely, the validity of using primarily incoming students' LSAT scores to rank the quality of law school faculty. I will take up each of these validity and reliability issues in turn.

I. FAGAN & HOLLAND: A TEST DEMONSTRATING RACIAL EQUALITY IN VERBAL INTELLIGENCE

I first wish to discuss a new standardized test of verbal intelligence that successfully embraces a culture-free assessment of ethnic ability. The test is due to Fagan and Holland.4 The Fagan-Holland test is important because it provides us with a standardized test that demonstrates racial equality in verbal intelligence.

Let me repeat the above assertion. The important result in the Fagan-Holland test of verbal intelligence is that no evidence for ethnic differences in intelligence occurs using their procedures."' Some of my earlier work suggested that if one used what I called a revised SAT scoring method, this would lead to a decrease of about one-third in the mean separation of AfricanAmericans and Whites.6 By contrast, the Fagan-Holland test leads to a total erasure of the mean difference between these two groups." To be sure, there are still strong individual differences to be found-that is, some students always score very high on this test, while others score consistently low-but when mean racial groups are compared, there are NO significant differences. This I consider to be a major breakthrough in assessing verbal intelligence in sharp contrast to the very negative things that, for example, Arthur Jensen has said in the past regarding racial differences in intelligence.8

So, here is what I would like to suggest to law school admissions officers. I would like you to consider using the Fagan-Holland test of verbal aptitude either as a replacement for the LSAT or at least in addition to the LSAT. That is, if one really wants to honestly assess an incoming student's verbal aptitude in a manner that is totally free of racial bias, here is the opportunity to do it. …