Math and Verbal Scores Up on Revamped SAT
A revamped SAT yielded improved scores for college-bound students in 1995, although officials denied that this year's examination was easier than in past years.
This year's version of the Scholastic Assessment Test, called the SAT I: Reasoning Test, sought more emphasis on critical-reading skills, application of concepts, and interpretation of data.
Officials of the College Board, which sponsors the SAT I, said 1995 takers of the college-entrance exam reported having taken record numbers of honors-level courses. Such high-level preparation, not the revised test, yielded the improved test scores, they said.
"This is our best-prepared class in recent memory," said Donald M. Stewart, the president of the College Board. "We expect scores to rise if students are taking more difficult classes."
Officials of the American College Testing Assessment, the nation's other most frequently used college-entrance exam, also announced last month their 1995 results. National composite scores on this year's ACT held steady at 1994 levels, they said.
SAT Scores Up
On the SAT I, the 1995 average verbal score rose five points over last year's, reaching 428 on the test's scale of 200 to 800. The average math score made a three-point jump from last year, to 482.
In the fourth consecutive year of rising scores for women, verbal and math scores reached 426 and 463, marking respective five- and three-point increases. Men's average verbal and math scores reached 429 and 503, up four and two points, respectively.
Both verbal and math scores for minority groups increased overall. The exception was among blacks and Puerto Ricans, whose math scores did not change, and Mexican-Americans, whose math scores fell one point.
Of the more than 1 million students who took the test in 1995, 25 percent came from households with incomes of more than $70,000, and 22 percent were in the top tenth of their high school classes.
More Time, Fewer Questions
College Board administrators said the SAT I: Reasoning Test had been painstakingly tested to ensure the same level of difficulty and reliability as the old form.
"What we're measuring hasn't changed--it's how we're measuring that's changed," said Gretchen Rigol, the executive director of the College Board's admissions and guidance services. "We're still measuring very basic reasoning skills."
The new test now has 78 questions in the verbal portion, down from 85. Students have 75 minutes to answer questions instead of 60. That change is complemented by more reading questions and reading passages designed to be more accessible.
Similarly, the mathematics portion has been adjusted to allow 75, rather than 60 minutes; to recommend the use of calculators; and to offer 15, rather than 20, quantitative comparisons. However, students now have to answer 10 new math questions by producing their own answers and entering them on grids.
The College Board plans to "recenter" the scoring of the test so that averages will be closer to 500, but that will not happen until next year. (See Education Week, Sept. 7, 1994.)
ACT Scores Steady
Meanwhile, after two years of gains, the 1995 national average composite score on the ACT held steady at 20.8, on a scale of 1 to 36.
The major change for this year's act, was that the number of test-takers increased to 945,000, up from 892,000 in 1994.
Act President Richard L. Ferguson said a large increase in students taking the test could lead to instability in the score results. In a statement, he referred to the fact that the national average remained stable as "cause for satisfaction, though not reason for rejoicing."
The average composite ACT score for women remained the same as last year--20.7, while the men's score rose slightly to 21.0 from last year's 20.9.
Test Accuracy Questioned
As the score results were released, one organization that promotes equitable testing pointed to the persistent gap between scores of men and women as evidence that the tests are gender-biased and therefore inaccurate.
Pamela Zappardino, the executive director of FairTest, said that even though women tend to perform better than men in high school courses, they score consistently lower on standardized tests. Research has shown that both the questions and the structure of standardized tests may be biased against women, she said.
The Cambridge, Mass.-based group advocates eliminating standardized tests as a college-entrance requirement--a step 235 four-year colleges and universities have taken. "When you try to reduce a person down to two or three numbers," Ms. Zappardino said, "by necessity you cannot have a fair system."