June 17 (Bloomberg) -- Expanding the SAT, the most widely used U.S. college entrance exam, made little difference in its power to predict grades, a study found, handing ammunition to critics who seek to minimize use of the test.

The New York-based College Board, which owns the test, released the study today showing that the current SAT rated 0.53 on a measure of predictive ability, compared with 0.52 for the previous version. A result of 1 would mean the test perfectly predicts college performance. Revising the SAT ``did not substantially change'' its capacity to foretell first-year college grades, the research found.

The College Board, a nonprofit membership association, made the exam longer in 2005 by adding a writing section, inserting questions on higher-level math, and removing analogies. Testing time grew to 3:45 from 3 hours, and the fee rose 41 percent. A critic of the SAT and other exams said test-taking ability doesn't translate into academic success.

``Meet the new test -- it's the same as the old test, but longer and more expensive,'' said Robert Schaeffer, public education director for the National Center for Fair and Open Testing, in a telephone interview yesterday. The organization, known as FairTest, is based in Cambridge, Massachusetts.

Since the new SAT was introduced, 41 schools have stopped requiring applicants to submit standardized-test scores. Last month, Wake Forest University in Winston-Salem, North Carolina, became the highest-rated national university to join that group, Schaeffer said.

Wake Forest is ranked No. 30 in U.S. News & World Reports magazine's national rankings, led by Princeton University in New Jersey. A total of 760 accredited four-year schools don't require test scores, according to Schaeffer.

Rival Act

The SAT, which now costs $45, was taken by 1.5 million members of the high school class of 2007, with 1.3 million taking another exam, the ACT, owned by the nonprofit Act Inc. in Iowa City, Iowa.

The College Board had said the SAT changes were meant to make the test ``more closely aligned with current high school curricula.'' The alterations were made after the University of California system, based in Oakland, proposed replacing the SAT with an exam more closely tied to the state's high school requirements.

The College Board also had said in a publication that the writing section, by adding another element to the verbal and math scores, ``can be expected to have greater predictive power of a student's probable success.''

While the predictive power changed little, the study affirmed the validity of the test, according to the College Board.

Attention to Writing

The study showed the SAT predicts college success almost as well as four years of high school grades, said Laurence Bunin, the College Board's senior vice president of operations, in a telephone interview yesterday.

The new writing section, which requires students to compose an essay and answer multiple-choice questions, also has led to a greater attention to writing by high school students and teachers, he said.

``This change was not just about predictive validity,'' Bunin said.

Of the three SAT sections, the new writing test was best able to foretell college grades, with a so-called correlation coefficient of 0.51, compared with 0.48 for critical reading and 0.47 for math, according to the study.

An SAT correlation coefficient of 0.50, for example, means that, on average, a quarter of the difference between any two students' college grades can be linked to their respective SAT scores, the College Board said.

Admission Decisions

High school grades had the highest correlation coefficient, 0.54, and when grades were combined with SAT scores, the coefficient rose to 0.62, according to the study.

``The College Board continues to encourage institutions to use both measures when making admissions decisions,'' the report said.

The SAT was better than high school grades at predicting students' performance at more-selective schools, the College Board said. The test had a correlation coefficient of 0.58 at colleges that accept fewer than 50 percent of applicants. High school grades had a coefficient of 0.55 for students at those schools.

The SAT ``is a national benchmark which colleges can depend on in an era when grade inflation at the high school level is a major problem,'' Bunin said in a teleconference today.

Smith, Bowdoin

Wake Forest officials, when dropping their SAT requirement last month, said they didn't consider the test to be an accurate predictor of college grades or graduation rates. High school grades and participation in college-level classes are better criteria, according to the school.

Smith College in Northampton, Massachusetts, also announced last month it would drop its standardized-test requirement. At Bowdoin College in Brunswick, Maine, which has made test scores optional since 1969, about 80 percent of last year's 6,033 applicants submitted scores.

The new report, said Schaeffer of FairTest, is ``sure to further accelerate the test-optional movement.''

To contact the reporter on this story: Brian Kladko in Boston at bkladko@bloomberg.net.

Last Updated: June 17, 2008 14:44 EDT