SAT Increase -- The Real Story, Part II
CER Press Release
Washington, DC
August 22,1996
The new 1996 SAT scores are being released today, amidst great fanfare. Another increase this year, writes the College Board, its sponsor. SATs are now the highest in 25 years screams their press release. The problem is, no one, including the College Board, knows for sure.

The difference between what was given before and now is too great, and the scores have been re-formulated using the new recentering method that was announced last year. Thus this year's average of 505 on verbal and 508 on math sounds really great. Adjusted for grade inflation, it is a 1 and 2 point increase over last year respectively. But throughout the College Board’s multi-page press release, there in only one real reference to the recentering in the narrative that most reporters would read, and no reference to the changes in the test instituted last year. We recap those from last year's CER alert:

• The tests have changed dramatically. As reported in the August 24, 1995 Wall Street Journal, "...the reason (for the SAT increase last year) may have more to do with changes in the test than improvement in schools" overall.

Those changes included a test that has fewer questions, longer reading passages, fewer multiple-choice math questions and no antonym section in the verbal section (WSJ, 8/24/95). Students also have longer to take the test, an additional 30 minutes, and may use a calculator.

• Those who claim that there has been no real decline in American education, such as authors of The Manufactured Crisis David Berliner and Bruce Biddle, have argued that the only declines in SAT scores since 1963 are due to more demographic changes among SAT takers (i.e. more disadvantaged).

In fact, as reported by State University of New York at Binghamton Professor Lawrence Stedman, the largest decline in SAT scores in the last two decades was among white students. In fact, only about 30% of the decline in the ‘70s was due to demographic changes in the scores, Stedman says. Only 40% who take the test report that they rank in the top 20% of their class. Yet in the last twenty years, the number of such college bound students who scored above 600 on the verbal portion of the test has slipped from 112,530 in 1972 to 73,080 in 1993, a 36% drop, despite the fact that the total number of test-takers has risen over 500,000.

• We are delighted that minorities now make up 30% of SAT takers, double the number in 1976. Over that period black students’ SAT scores have increased 20 points in verbal, and 34 points in math, while white students' scores have declined 8 points in verbal and increased 2 points in math. Still, blacks continue to trail their white peers by over 100 points on each of the sections, math and verbal.


• There is good news. Students taking the test report doing more academic course work than in previous years, and taking more AP classes. The only problem is that it is unclear whether the quantity of course work translates into rigor. A glance at reading, writing, history and geography results from the National Assessment of Educational Progress would suggest that rigor is still largely missing from even the best students’ course work.


• Chester Finn and Diane Ravitch, writing in the Educational Excellence Network’s 1995 Report Card (the 1996 issue is just now out) said "The College Board's decision to 'recenter' the SAT scores has considerably reduced the utility of those scores as a national barometer of the educational performance of college bound students. For some two decades, the SAT has served this function, mainly because of its stable scale. The recentering however magically...gives the impression that the nation's educational deficit has been eliminated."

In a way, the apologists have won. They've been able to destroy the SAT as a useful means of student assessment, a goal that the College Board argued last year it never intended.

• Why did they artificially recenter the scores to 500 if not to boost grade inflation? In previous years, students scores were predicated on the number of questions that were correct. A perfect 1600 could only be made by getting every question right. Now up to four questions can be wrong and a student would still get a 1600. In addition, last year's average of 428 on verbal and 482 on math has been recalibrated according to the new scores to 504 and 506 respectively, thus not showing the wide gap between performance on the two disciplines. Thus students from this day forward -- and the media, and colleges -- will have no real means of comparison on which to gauge what someone's 1985 score of 510 means compared to today's score of 510.

According to College Board graphs, both original and recentered scales show that 1972 was the highpoint for SAT scores. Precipitous drops through the early eighties were followed by modest increases in the mid-eighties, another decline from ‘86-’91 has been followed by marginal gains from 1994 through today.

SAT scores, even with the inflationary measures, are still roughly 25 points below the 1972 high in verbal, and a few points lower from their 1972 high in math.

# # #