SAT Scores Drop

SAT scores are down this year. And while the College Board played down that news and attributed the falling scores to growth in the test-taking population, the downward shift runs counter to recent patterns. The data also show continuation of a trend that has concerned many educators for years: growing gaps by race and ethnicity in how students perform on average on the test.

The trend in recent years has been a point up in one part of the SAT, offset by a point down in another part -- with minimal movement in total. But this year saw a three-point decline in critical reading, a one-point decline in mathematics, and a two-point decline in writing.

SAT Scores 2011

The writing score has been gradually falling since it was introduced in 2006. Of the reading and mathematics tests, which have been around a lot longer, the combined score of 1011 is the lowest total since 1995. In the years since 1995, those two scores combined have reached as high as 1028.

The ACT, the SAT's rival in the college admissions testing field, saw modest gains in its composite scores this year.

Proponents of both the ACT and SAT note that those who take college preparatory courses earn better scores, on average, than those who don't. And this year's data for both tests reinforce that message. But the data also show both longstanding and growing gaps in scores, when examined by socioeconomic groups.

The College Board breaks down score averages by family income, starting at income levels below $20,000 and going to more than $200,000. With 10 income levels, and three parts of the SAT, the average score rises on each test at each income level, from 460 in mathematics in the below $20,000 range to 586 in mathematics for those with family incomes of more than $200,000.

Gaps are also evident by racial and ethnic group. Asian Americans continue to show gains -- even as this year other groups do not. Over the last three years, the combined average scores of Asian American test takers have gone up by 30 points, while every other group showed declines.

The gaps in the above table point to a major challenge for the College Board and for colleges. Institutions that tend to admit applicants of different ethnic groups with notably different SAT averages get attacked by critics of affirmative action, and sometimes these colleges are sued. Many colleges have cited their discomfort with the gaps as one reason to end SAT requirements, and institutions that have done so have generally reported surges in black and Latino applications.

The gaps also play into the debate over why this year's scores dropped. At a news briefing, James Montoya, vice president of higher education relationship development at the College Board, said that the average scores were "down slightly," and that this was not surprising given that more students and a more diverse pool of students took the exam this year.

He also noted the "devastating" inequities in American society in terms of the quality of high schools attended by students from different socioeconomic groups.

Others questioned the College Board's analysis.

Robert Schaeffer, public education director of the National Center for Fair and Open Testing, noted that the SAT has been growing steadily in recent years (and not experiencing overall score declines). He also argued that the changes within ethnic groups suggest that more is at play than just having more diverse test-takers.

He said that the College Board has endorsed the testing emphasis of No Child Left Behind reforms in elementary and secondary schools, and is not admitting that this year's results add to evidence that all of the testing has not worked. "A key promise of the high-stakes testers was that achievement gaps would narrow" if more schools did more-systematic testing. "But this shows we are going backwards, and that achievement gaps are growing. Whether you like tests or not, that's bad news for the country," he said.

While College Board officials regularly talk about the inequities in American education, the organization is sensitive to allegations made by FairTest and others that the SAT contributes to those inequities.

In his remarks at the briefing, Gaston Caperton, president of the College Board, spoke with pride about the $37 million in revenue that the College Board loses by granting fee waivers to low-income students who want to take the SAT. (The basic fee for the SAT is $49, with some potential additional charges.) Caperton said that there were "no limits" on the fee waivers for low-income students.

Actually, though, there are limits. Students who can afford the fees may take the SAT as many times as they would like and report only their best scores. Currently, about 13 percent of students take the SAT three or more times. But the College Board will only grant fee waivers for two SAT tests, so that 13 percent is most likely made up of students who are not in the lowest income groups. A spokeswoman said that when Caperton said that there were "no limits" on fee waivers, he meant that there was no budget limit on the program.

A New Benchmark

Caperton and other College Board officials on Wednesday spoke about a new way to analyze SAT scores. They announced that they were introducing a benchmark -- currently at 1550 across all three sections of the SAT -- at which they could confidently predict that students have a 65 percent chance of earning at least a B- average during the first year of college.

Board officials said that they were not recommending that colleges set any minimum score requirements based on the benchmark or base admissions decisions on it. But officials said that colleges and high schools could track their success at producing or admitting students who meet the goal.

Wayne Camara, vice president for research at the College Board, said that the benchmark "is designed to give us a holistic index of the kind of preparation students are receiving." College Board officials spent much more time in the news briefing on the benchmark than on the decline in scores, although news coverage -- as is typically the case -- focused on the scores.

Jeff Rickey, vice president and dean of admissions and financial aid at St. Lawrence University, said he didn't understand why the College Board was introducing this benchmark. "I'm not sure why the College Board feels the need to continue to grasp for measures of the validity of the SAT when the number of test-takers goes up each year," he said.

Rickey was a member of a commission of the National Association for College Admission Counseling that in 2008 urged colleges to reconsider their use of standardized testing, but that did not rule out its effective use. St. Lawrence is test-optional, and he said that last year, about 74 percent of applicants who were admitted did submit test scores.

Of the benchmark, Rickey said that "it seems like repackaging SAT results in new paper." While Rickey said he and his colleagues continue to use SAT scores for applicants who submit them, he said he was bothered by the boasts of College Board officials about the value of the benchmark. Rickey noted that repeated research studies have found that "the high school G.P.A. is the single best predictor of first-year college grades ... and they don't even acknowledge that."