Although controversy over the SAT persists on a number of fronts, the verbal portion of the test renamed “critical reading” in 2005) is a valuable measure of college-bound seniors’ linguistic skills because the test has been administered for several decades and thus permits comparison over an extended period of time. The SAT data reveal a steep decline from 1967 through the early 1980s in the mean verbal score for all college-bound seniors,1 followed by a leveling off, with the mean score through 2011 varying within a relatively narrow band between the upper 490s and 510 (Figure I-5a; scores have been adjusted to take into account the 1995 recentering of the scoring system).

Figure I-5a, Full Size Supporting Data

As was widely reported at the time, the mean SAT verbal score fell sharply in 2006, experiencing its largest one-year drop since 1975. The College Board, the body that administers the SAT, attributed the drop to a decline in the number of students who retook the exam (examinees’ scores tend to increase substantially the second time they take the test). Some observers offered a different explanation, asserting that the diminished average score was due to the increased length of the exam, which included a new writing section. College Board officials responded that an analysis of approximately 700,000 tests produced no evidence to support this theory.2 The average verbal score in 2011 was 497, down six points from 2006.

The average SAT math score also declined over the course of the 1970s. But unlike the SAT verbal average, which has remained fairly constant since the early 1980s, the mean score on the quantitative portion of the math exam rose steadily from 1980–1981 (its lowest point) until 2005. Like the verbal average, the math average dipped in the mid-2000s, with scores remaining at this somewhat reduced level—514–515—through 2011.

From 1967 through the late 1980s, the SAT math average was consistently lower than that for the verbal exam. But with the steady improvement of the math average and the stagnation of the verbal average, American students were demonstrating somewhat stronger math than verbal skills by 1990. The gap has grown since then, so that by 2011 the mean math score was 17 points higher than the mean verbal score. This is a profound reversal from the state of affairs in 1967, when the average verbal score exceeded the math average by 27 points.

Male students’ average verbal SAT score has been consistently higher than that of female students since the early 1970s. Initially the gap was small, but the disparity grew, and during the 1980s the gender gap in verbal scores ranged from 10 to 13 points. The gap has narrowed since then, with the average score of female examinees coming within five points of the male average in 2011.

The difference in mean scores between whites and racial/ethnic minority groups is more pronounced (Figure I-5b). Mean verbal scores for most minority groups increased from 1991 through 2001. However, ten years later, for most groups, average scores were the same or down somewhat. In contrast, the cohort of students who self-identified as Asian, Asian American, or Pacific Islander continued to make substantial gains. In 2011, these students scored 32 points higher, on average, than the cohort of the early 1990s had. However, white students still performed best over the two decades. The gap between African American and white scores was particularly large throughout the time period, with white examinees scoring 100 points higher, on average, in 2011.

Figure I-5b, Full Size Supporting Data

High school seniors graduating in 2006 were the first to take the SAT’s new writing test, which includes both multiple-choice questions and a student-written essay. The mean score on the writing test for all college-bound seniors was 497 in that year (Figure I-5c), with female students scoring higher than their male counterparts, on average (502 versus 491).3 Five years later, the average score for all college-bound seniors was lower (by eight points), as was the average for every racial/ethnic group, with the exception of Asians, whose average score was 16 points higher in 2011 than in 2006, bringing it above that for whites. Every other minority group’s average was lower than that for white students. The magnitude of the performance gap between each of these groups and whites was similar to that for the verbal exam. From 2006 to 2011 the gap between average female and male performance on the writing examination increased, with the female advantage growing from 11 to 14 points.4

Figure I-5c, Full Size Supporting Data

---

Notes