Test Scores Are Not "Merit"

General Testing

As the assault on affirmative action intensifies, the debate increasingly has focused on the use of test scores in college admissions and employment hiring. Affirmative action opponents consistently try to equate high test scores with high qualifications and "fair" selection. But if the goal is allocating opportunity on the basis of necessary skills or projected performance on the job or in the classroom in order to allocate opportunity fairly, exam scores do a very poor job. In fact, dozens of academic studies and court rulings have confirmed that standardized tests contribute to the ongoing discrimination many minorities and women still face.


For example, admissions exams like the SAT and ACT claim to predict how well students will do in their first year of college. Yet the test-makers own research shows their scores explain only one-sixth of the differences among students first-year grades and an even smaller portion of four-year performance. All the other factors that contribute to academic success studiousness, creativity, and the like are not measured by a standardized exam.


There is also ample data indicating that these tests underpredict the achievement of entire groups such as women. Based on the overwhelming evidence, in 1989 a Federal Court barred the use of SAT scores as the sole factor to award millions of dollars in New York State scholarships concluding, "SAT scores capture a student's academic achievement no more than a student's yearbook photograph captures the full range of her experiences in high school."


California the First Target

In spite of the data, the admissions practices of the entire California public higher education system are under attack by test-score advocates. One attorney has brought a consumer fraud lawsuit against the University of California (U.C.) Board of Regents for not telling applicants that factors other than test scores and grades are considered in admitting law and medical school candidates.


This July, the Regents voted to eliminate all forms of affirmative action, stating that "race, religion, sex, color, ethnicity or national origin shall not be the criterion for admissions in exception to U.C. eligibility requirements." Sponsors ignored a study at the Berkeley campus concluding that heavy weighting of test scores in the current admissions eligibility formula is itself a source of bias (see p. 5).


A so-called "California Civil Rights Initiative" banning affirmative action throughout the state is scheduled for the 1996 ballot. It has the strong backing of Governor Pete Wilson, a candidate for the Republican presidential nomination. Ironically, one of Wilson s chief backers in that bid and a fellow conservative, Massachusetts Governor William Weld, offered an effective rebuttal, "Who among us would want to have an S.A.T. score emblazoned on our foreheads, and have that alone stand for all that we are, all of our potential?"


Employment Testing

Similarly, a number of court cases have shown that employment exams such as those used to select teachers, doctors, police officers and members of hundreds of other occupations have little relationship to work performance. That is hardly surprising since no pencil-and-paper test can possibly measure dedication, innovation or the ability to work with others the skills most employers say they really want.


Still, supporters of civil service tests and the like argue that selection decisions should be made solely by ranking candidates on exam scores. Anti-affirmative action forces refer to a handful of studies justifying such employment tests, but many of these have been severely criticized for exaggerating very weak correlations. The most widely administered federal employment tests, the General Aptitude Test Battery (GATB) and the Administrative Careers with America (ACA) exams, have both been replaced due to concerns about bias and inaccuracy (see Examiner, Fall 1990 and Winter 1994-95).


The evidence does not deter those who want to make political points or contort reality to confirm their prejudices. Several years ago, Congressional conservatives made a major public relations campaign out of a drive to repeal what they labeled "race-norming" on the GATB, a policy of reporting candidates' scores based on their test score rank within their own racial groups (Examiner, Summer 1991).


Yet, a report by the National Research Council of the National Academy of Sciences had specifically endorsed the policy of "within-group scoring." The Academy pointed to the higher percentage of false, negative predictions for African Americans and Latinos concluding that "the disproportionate impact of selection error provides scientific grounds for the adjustment of minority scores so that able minority workers have approximately the same chances of referral as able majority workers."


Exaggerated Claims

In addition to ideological and racial motivations, test score advocates are goaded by exam-makers who make highly exaggerated promotional statements about their products. The College Board, for instance, has long labeled its SAT "A Common Yardstick" which provides equal measure for all students. The Educational Testing Service has spent millions on public relations trying to establish the SAT, GRE and similar exams as symbols of what one author called "the meritocratic system of education."


Manufacturers of the Wonderlic Personnel Test allege that it predicts accurately for thousands of jobs. For those who never read the test-makers' technical reports where such claims are carefully qualified, this rhetoric frames perception.


Another intentional source of confusion is the "National Merit" schoalrship competition, which relies exclusively on Preliminary SAT scores to narrow the pool of applicants from over one million to fifteen thousand. As a FairTest complaint now under investigation by the U.S. Education Dept. Office for Civil Rights demonstrates, bias in that test causes girls to win an unfairly small portion of the awards (Examiner, Winter 1993-94).


Using any more realistic standard of performance, including grades in identical high school or college courses, class rank and graduation rates, females would have earned a majority share of scholarships. What the tests reward is not what the academic system otherwise views as merit.


This year, the greatest single reinforcement of the anti-affirmative action, pro-testing movement was the publication of The Bell Curve. Backed by a huge publicity campaign, that book advances the claim that test scores accurately measure the ability of individuals and groups, and hence their "merit." Yet, a number of independent researchers have demonstrated that the data on which authors Charles Murray and Richard Herrnstein based their argument are often fraudulent, exaggerated, misinterpreted, or deliberately distorted by researchers and funding organizations which believe in racial superiority (see Examiner, Winter 1994-95).


Opportunity to Perform

In truth, there are much better ways to make high-stakes selection decisions. Nationally, at least 235 colleges do not require admissions applicants to submit test scores. Instead, they rely on high school grades or class rank (both of which test-makers admit are more accurate predictors than exam scores), extracurricular activities, community service, and, increasingly, portfolios of school work.


A growing number of employers also use performance-based measures, particularly work history. They recognize that merit cannot be reduced to a test score.


Ending the test score fixation and looking at broader, less biased, and more meaningful measures of "merit" is the best way to promote excellence and fairness while combating the racist and sexist attacks that are at the core of the affirmative action debate.