Test-Makers to Revise Nat. Merit Exam to Address Gender Bias - FairTest Complaint Will Lead to Millions More for Girls

University Testing

(See also Test-Makers Try to Dodge Legal Implications)


After two and a half years of investigation and negotiation with the defendants, the U.S. Department of Education's Office for Civil Rights has announced a settlement of FairTest's complaint charging the Educational Testing Service (ETS) and the College Entrance Examination Board (College Board) with illegally assisting gender discrimination by manufacturing, sponsoring, and administering a gender biased exam used as the sole criterion to select National Merit Scholarship semifinalists (see Examiner, Winter 1993-94).


To avoid a finding of discrimination that would jeopardize their federal funding, ETS and the College Board agreed to revise the Preliminary SAT/National Merit Scholarship Qualifying Test (PSAT/NMSQT) in 1997 to "sufficiently address the allegations" in FairTest's complaint. But, rather than move to a system based on genuine academic performance in the classroom, as FairTest proposed, the test-makers simply plan to add a multiple-choice "writing" component to the PSAT/NMSQT. Data from similar exams, such as the SAT II Writing Test, suggest that females will score higher than males on the new section, thus increasing their chances for selection as scholarship semifinalists.


The FairTest complaint argued that the lion's share of National Merit Scholarships went to boys each year because they score higher on the PSAT/NMSQT, although girls earn better grades in both high school and college when matched for the same academic courses. Even the test-makers' own research admits that the test underpredicts the performance of females and over-predicts the performance of males (see Examiner, Winter 1994-95, Summer 1993, and Spring 1992). For years, ETS, the College Board, and the National Merit Scholarship Corporation stonewalled requests from women's and civil rights activists to overhaul the selection procedure.


By agreeing to change the PSAT/NMSQT, ETS and the College Board effectively admitted what FairTest had charged -- their college admissions exams are gender biased and inaccurate. As a result of the alteration, millions of dollars more in scholarships each year should go to the young women who earned them through their superior academic performance.


The settlement has already led to calls for similar changes in the SAT and other exams such as the GRE which also underpredict the performance of females. Unfortunately, by relying on a multiple-choice test that really measures proof-reading skills, the new "writing" section may have a harmful effect on high school curricula.


Title IX of U.S. education law prohibits recipients of federal funds from discriminating on the basis of sex or significantly assisting in such discrimination. The Department of Education's Office for Civil Rights is responsible for enforcing the equal opportunity requirements. Both ETS and the College Board receive U.S. Department of Education funds through a variety of contracts and grants. The complaint did not target the National Merit Scholarship Corporation because that organization does not directly receive federal funds.


The ACLU Women's Rights Project ably represented FairTest in the National Merit gender bias complaint. Excellent pro bono legal counsel was also provided by Mark Friedman and Julie Brandfield from the firm of Debevoise and Plympton as well as by former FairTest Executive Director Cinthia Schuman, who was of counsel.


This is the second legal victory in a case initiated by FairTest against gender bias in the SAT exams. In a 1989 lawsuit spurred by FairTest research and litigated by the ACLU, a federal district court judge declared New York State's system of awarding college scholarships based solely on SAT scores unconstitutional and in violation of Title IX (see Examiner, Spring 1989).


FairTest and its legal team plan to closely monitor the settlement to make sure that young women receive the scholarships they truly merit. The Office for Civil Rights retained jurisdiction to assure the settlement is implemented.


Test-Makers Try to Dodge Settlement's Message

Like tobacco company executives who distort the truth and cover up their own research to preserve their product's image, those who profit by promoting standardized tests are scrambling to deny that the National Merit settlement represented an admission that the exam was flawed and dangerous. For example, College Board President Donald Stewart grossly misstated several significant facts in a letter responding to a New York Times editorial endorsing FairTest's critique of gender bias in the SAT.


Dr. Stewart claimed the Times "misread" a study by two researchers at the Educational Testing Service, the company that manufactures and administers the PSAT and SAT for the College Board. But that report clearly concludes that the SAT underpredicts the math grades of women in college courses and urges that test scores not be used as the sole factor to select National Merit Scholarship Semifinalists (see Examiner, Summer 1993).


Similarly, Dr. Stewart's letter was misleading in stating that this study is inconsistent with "the larger body of research on gender differences in academic abilities." In fact, a large number of reports, several of which were published by the College Board itself, show that girls perform as well or better than boys in both high school and college when matched for identical courses. It is the SAT and similar fast-paced, multiple-choice tests that are out of synch with measures of genuine academic merit.


Dr. Stewart also argued that the study covered only one school even though the authors state that it is based on 47,000 students at 51 colleges and universities. In-depth research at many schools, such as the University of California at Berkeley (see Examiner, Summer 1995) and the Massachusetts Institute of Technology, confirm the same point: the SAT falls far short of its sole advertised claim -- the accurate prediction of first year college grades -- by consistently underestimating the academic performance of females.


Dr. Stewart, a member of the New York Times Corporation Board of Directors, apparently managed to get his letter published without any fact-checking. He owes the paper's editorial board an apology for accusing them of sloppy research. Even more importantly, he should make amends to the more than one million young women each year who take College Board tests by eliminating their exams' gender bias. It's time for exam-makers to tell the truth by acknowledging that test scores do not accurately measure merit.


1998 Competition Remains Gender Biased

The two-thirds of a million girls from the high school class of 1998 who entered the National Merit Scholarship competition by taking the Preliminary SAT (PSAT) this fall still face a "stacked deck" from the exam's gender bias. Of the 1.2 million juniors who sat for the exam, only the 1.5% with the top test-scores in each state will be eligible for $25 million in National Merit awards.


A FairTest state-by-state analysis of the high school class which graduated in 1996 concluded that only 39.5% of the more than 15,000 National Merit Semifinalists -- selected solely on the basis of test scores -- are female, while 55% are male. The gender of 5.5% of the Semifinalists could not be determined from the names. Fifty-six percent of those entering the National Merit competition are female. Data for the class of 1998 are expected to be similar because the selection exam has not yet changed.


A selection system which accurately reflected academic achievement would result in young women earning about 1,000 additional scholarships annually. Unfortunately, a flawed test will, once again this year, cheat talented females out of $2 million in scholarships as well as immeasurable prestige.


* For a copy of the state-by-state gender analysis, send a stamped, self addressed envelope to "National Merit" at FairTest.