Go to

Transition to Higher Education

College Entrance Exams

The transition from elementary/secondary school to higher education is an important step not only to the individuals making it, but also to a nation committed to the well-being of its citizens. Information on persons making this transition provides opportunities for the assessment of their progress through the stages just completed and their readiness for future activities. In this report, the transition points mark an important opportunity for examining the status of underrepresented groups as they progress through the educational system.

College Entrance Examinations

Persons With Disabilities

Two organizations administer national college entrance examinations-the Admissions Testing Program of the College Entrance Examination Board, which administers the Scholastic Aptitude Test (SAT), and the American College Testing Program, which administers the American College Testing (ACT) Assessment. Results of these examinations are of substantial importance to college admissions decisions and hence to opportunities for college students. A close analysis also offers further insight into the precollege preparation of women and underrepresented minorities. Substantial differences remain in standardized test results among the various groups at the critical transition point from secondary school to higher education.


Scholastic Aptitude Test

The Admissions Testing Program of the College Entrance Examination Board collects and tabulates data on the scores of college-bound seniors who have taken the SAT. The College Board uses the term "college-bound senior" to refer to those students from each high school graduating class who take the SAT Program tests anytime during their high school years. [3] The SAT examination consists of two components: the verbal component, which tests reading comprehension and vocabulary skills, and the mathematics component, which assesses the ability to solve problems by using arithmetic reasoning as well as skills in basic algebra and geometry. [4] The score range for each SAT component is from 200 to 800.

In 1994, almost 1.1 million students took the SAT tests; females constituted 53 percent of the total. (See appendix table 2-25.) Continuing a long-time trend, in 1994 females scored below males in both the mathematics and verbal portions of the SAT. This pattern persists despite the fact that females tend to have higher overall grades in high school than males, [5] and they tend to have better grades in college (see the related discussion on undergraduates in chapter 3). Educators and researchers both in the academic community and within the College Board have been concerned about the underlying causes of this disparity. [6]

SAT Scores and High School Classes

Mathematics. On the mathematics component of the SAT, scores for both sexes have risen during the decade since 1984, a period of increased emphasis on mathematics and science education at the K-12 level. Nevertheless, females in 1994 continued to score considerably below males in the mathematics component, the gap narrowing only slightly over the decade. (See figure 2-7.) Since 1984, females' scores increased 11 points to 460 in 1994, whereas males' scores increased 6 points to 501. Thus, there was a 41-point difference in scores in 1994, down from a 46-point difference in 1984. (See appendix table 2-26.)

See appendix table 2-26.

This large difference in mathematics scores between the two sexes occurred despite the similarity in many of their high school characteristics. In 1994, females who took the SAT exam reported completing an average of 3.6 years of mathematics coursework compared with 3.7 years for males. Females received a grade point average of 2.96 in mathematics, compared with a mathematics grade point average of 2.97 for males. [7] (College Entrance Examination Board 1994, p. 10).

Verbal. In 1994, females also continued to score somewhat lower than males on the verbal component of the SAT. (See figure 2-7.) This occurred even though females reported a higher high school grade point average than males in both English and social sciences/history. [8] Females also took a higher average number of years of coursework in English (3.9 years for females versus 3.8 years for males) and social sciences/history (3.4 years for females versus 3.3 years for males) (College Entrance Examination Board 1994, p.10).

SAT Scores and Level of Difficulty of High School Mathematics
and Science Courses

The propensity for taking difficult coursework in high school may account for some of the differences between males and females in mathematics test scores, according to an analysis of the profile data reported by high school seniors who take the SAT. In particular, although males and females had almost the same percentage taking honors courses and had almost identical grade point averages in the mathematics courses they took, a smaller percentage of females took 4 or more years of mathematics, [9] and a much smaller percentage of females took the most advanced coursework.

The discrepancy in course taking between the males and the females taking the SAT occurs in courses that are generally electives, i.e., those following the geometry course. For example, 96 percent of both males and females took algebra, and 93 percent of both genders reported taking a geometry course. There is a gap of 3 percent, however, in male/female participation in both trigonometry (53 percent for females versus 56 percent for males) and precalculus (34 percent for females versus 37 percent for males). The gap widens to a 5 percent difference in the proportion taking calculus (19 percent for females versus 24 percent for males). (See appendix table 2-27.)

This difference in propensity to take the more difficult mathematics courses undoubtedly contributes to the male-female differences in scores. Females were much less likely than males to place in the top range of scores on the mathematics component of the SAT, i.e., in the 600 to 800 range. In 1994, only 14 percent of females scored in this top range versus 24 percent of males. (See appendix table 2-28.)

A similar pattern is evident in enrollment in natural science classes. Females' grade point averages are very similar to males' in the courses they take; both sexes take about the same number of years of coursework; and they participate equally in the percentage taking honors courses. [10]

As is the case with mathematics, however, a smaller percentage of females take the most advanced coursework in the natural science fields. For example, 97 percent of all students who took the SAT, both female and male, had taken biology, and 83 percent of both sexes had taken chemistry. Only 41 percent of females took physics, however, compared with 51 percent of males. (See the related discussion above concerning a study by Neuschatz and Alpert, American Institute of Physics.) In coursework intensiveness, 45 percent of females took 4 or more years of natural science, compared with 50 percent of males.

SAT II: Achievement Tests

The differences in coursework taken may also affect the differences between males and females in scores received on the achievement tests offered by the Admissions Testing Program of the College Board. [11] Although females took 50 percent of the achievement tests in science and mathematics in 1994, [12] female participation was concentrated in the less advanced mathematics I exam in which females took 57 percent of the total, and in biology (55 percent of the total). Males took the majority of all the other mathematics and science achievement test exams. Female participation was lowest in physics, in which they took only 26 percent of the exams.

In the mathematics and science achievement tests they did take, females' mean scores were lower than the mean scores for males in 1994. (See appendix table 2-29.) The discrepancy ranged from 31 points on the biology test to 53 points on the physics exam. The spread between scores on the new math level IIC was 45 points (650 for females and 695 for males).

Intended Undergraduate Major

Differences between females and males in their intended preference for degree major are striking for students planning to enter college. Perhaps in keeping with their lower scores on the mathematics SAT, relatively few females about to enter college in 1994 intended to pursue a major in engineering. (See figure 2-8.) Only 3 percent of females intended to major in this subject, whereas 17 percent of males intended to major in engineering, the highest percentage for any individual major for males. (See appendix table 2-30.)

See appendix table 2-30.

Twenty-four percent of females cited health and allied services as their most probable major. Business and commerce was the next most popular field for women (13 percent), followed by education (11 percent). For males, business and commerce was also the second most popular probable major (15 percent), followed by health and allied services (13 percent). Education was mentioned by just 4 percent of the males.

Combining all natural science fields, 14 percent of the males intended to pursue these majors, and 10 percent of the females chose these fields as probable majors: two percent of males chose agriculture/natural resources as their major, compared with 1 percent of females. One percent of males chose mathematics as a major, and less than 0.05 percent of females did. Double the percentage of males than females also chose the physical sciences (2 percent and 1 percent, respectively) and computer sciences (4 percent and 2 percent). Only in the biological sciences did a larger proportion of females choose the discipline-6 percent of females chose biological sciences, compared with 5 percent of males.


Scholastic Aptitude Test

Mathematics. An analysis of the descriptive information submitted by students taking the SAT reveals a wide divergence in precollege preparation among the racial/ethnic groups. These differing rates of participation in mathematics and science training in elementary and secondary school are reflected in the scores received on the mathematics portion of the SAT.

Compared with whites, the three minority groups underrepresented in science and engineering-blacks, Hispanics, [13] and American Indians-tend to take fewer courses in mathematics and science. Asians, who engage in science and engineering in larger proportions than their percentage of the general population, take more science and mathematics high school courses than whites. An analysis of scores reveals that, overall, Asians perform better than all other racial/ethnic groups on the mathematics component of the SAT and on the science and mathematics achievement tests; whites score second highest. Asians also tend to take more of the difficult mathematics and science courses in high school than do students in other groups. (See appendix table 2-27.)

On the mathematics component of the SAT, the scores of every racial/ethnic group improved over the decade, again undoubtedly reflecting increased emphasis on improving mathematics and science education at the K-12 level. (See figure 2-9.) The relative standing of the racial/ethnic groups did not change over the 10-year period, however; the groups scored in the same rank order as in 1984.

See appendix table 2-26.

In 1994 Asians continued to have the highest average mathematics SAT scores, followed in order by whites and American Indians, Latin Americans, Mexican Americans, Puerto Ricans, and blacks. (See appendix table 2-26.) Asian students also achieved the highest increase in mathematics scores of any racial/ethnic group, with scores rising 16 points over the decade. Black students achieved the second highest increase in scores since 1984 (15 points), and American Indian students achieved a 14-point increase.

Verbal. On the verbal component of the SAT, whites had the highest mean scores in 1994, followed by Asians and American Indians. (See figure 2-10.) The relative ranking of these groups remained about the same between 1984 and 1994, but several significant changes occurred in the level of the verbal scores. Asians achieved the highest increase in scores of any racial or ethnic group; their verbal scores rose every year for a total increase of 18 points over the decade.

See appendix table 2-26.

Blacks had the second highest increase in mean verbal scores (10 points), whereas American Indians increased their verbal scores by 6 points. Scores by whites fluctuated slightly over the decade but decreased by 2 points overall between 1984 and 1994. Trend data on Hispanics are more difficult to compare because of the data subdivision in 1987. Of the three Hispanic groups, however, only the Puerto Ricans had verbal scores higher in 1994 than in 1987: they rose a total of 7 points by 1994.

SAT Scores and Level of Difficulty of High School Mathematics
and Science Courses

The amount and type of coursework taken in high school are related to the scores achieved on the SAT. In particular, Asians and whites, the two groups with the consistently highest mathematics scores on the SAT, were also the two groups who had taken the most courses in mathematics and natural science in high school.

Science.In 1994, 89 percent of college-bound Asians, 85 percent of whites, and 80 percent of Latin Americans took chemistry in high school; roughly three-quarters of each of the other groups took chemistry. The biggest difference in participation rates among racial/ethnic groups in science coursework was in physics. Sixty-five percent of Asians took physics, compared with 47 percent of whites, 44 percent of Latin Americans, and 40 percent of Puerto Ricans. For all the other racial/ethnic groups, less than 35 percent of the college-bound students took physics. (See appendix table 2-27.)

Mathematics.As with females, high percentages of college-bound students from all racial/ethnic groups took algebra and geometry, but the percentage of participation starts to diverge after these two basic high school mathematics courses. Asians were again the most prepared in terms of coursework taken. Sixty-nine percent of Asians took trigonometry, whereas the next highest proportions were 55 percent for whites and 51 percent for Latin Americans. No other racial/ethnic group had a majority of their college-bound seniors taking trigonometry in high school.

The gap widens even further in precalculus: 53 percent of the Asians took that course in high school. The whites' proportion was 17 percentage points behind; 36 percent took precalculus. All other racial/ethnic groups had fewer than one-third of their students taking precalculus in 1994

Only a minority of all racial/ethnic groups took calculus in high school, yet even here Asians participated at the highest level. Forty percent of Asians took calculus, as did 22 percent of whites. In all other groups, fewer than 20 percent of their student college-bound population took calculus.

Parental Income and SAT Scores

The SAT data show that for every racial/ethnic group, higher reported levels of parental income are generally associated with higher scores on both the verbal and mathematics sections of the SAT. Family income does not uniformly relate to level of achievement, however. The mean SAT mathematics score of 482 for those Asian students at the lowest family income level (under $10,000) exceeded the scores at the highest family levels for several of the underrepresented minorities groups. (See appendix table 2-32.)

Parental Education and SAT Scores

Within every racial/ethnic group, higher levels of parental education were associated with higher students' scores on both the mathematics and verbal portions of the SAT. For example, the difference in mean SAT mathematics scores between the group whose parents did not receive a high school diploma and those whose parents held a graduate degree ranged from 120 points for whites to 85 points for blacks. (See appendix table 2-33.)

A majority of college-bound students in four racial/ethnic groups reported that the highest level of education attained by their parents was a high school diploma or less (Mexican Americans, 70 percent; blacks, 57 percent; Puerto Ricans, 55 percent; and Latin Americans, 54 percent). Although these four groups tended to score lowest on the SAT, within each of these groups the parental education pattern held: average SAT scores increased with the increase in the level of the parents' education.

Citizenship Status and SAT Scores

More than 90 percent of college-bound students taking the SAT in 1994 were U.S. natives or naturalized citizens in all but two of the racial/ethnic groups studied, [14] but only 59 percent of the Asian students taking the SAT and 68 percent of the Latin American students taking the SAT were U.S. natives or naturalized citizens. An additional 27 percent of Asians were permanent residents or refugees, and 15 percent were citizens of another country. For Latin Americans, an additional 23 percent were permanent residents or refugees, and 9 percent were citizens of another country. (See appendix table 2-34.)

Verbal Scores. For all but one racial/ethnic group, verbal SAT scores of U.S. native or naturalized citizens were higher than the verbal scores for either permanent residents/refugees, or for citizens of another country-undoubtedly reflecting the higher proportion of students for whom English is the first language learned. Blacks are the one exception to this pattern of scores. The mean verbal score for black citizens from another country was 29 points above the mean verbal score of black U.S. citizens (381 versus 352). Citizens from another country constituted only 2 percent of blacks taking the SAT, however.

Mathematics Scores. The pattern of higher U.S. citizen scores changes for the mathematics component of the SAT. In all but two racial/ethnic groups-Mexican Americans and Puerto Ricans- the citizens from other countries achieved higher mathematics SAT scores than did the U.S. citizens. The number of foreign citizens in these two ethnic groups was very small, however, constituting one percent or less of each group.

SAT II: Achievement Tests

Approximately 19 percent of all students who took the SAT in 1994 also took at least one achievement test. The proportion of students taking at least one achievement test varies dramatically by racial/ethnic group. Although whites (17 percent), Mexican Americans (19 percent), and Latin Americans (20 percent) all took achievement tests at a rate similar to the national average of 19 percent, the proportion was lower for Puerto Ricans (12 percent), American Indians (11 percent), and blacks (9 percent). On the other hand, the proportion of Asian SAT takers who also took at least one achievement test (42 percent) was far above the national average. (See figure 2-12.)

Intended Undergraduate Major

Racial/ethnic differences in choice of undergraduate major are less dramatic than the differences by sex. Particularly when the social sciences are separated from the natural sciences and engineering, the differences in preference by sex become striking: the proportion of males intending to major in natural sciences and engineering was significantly higher in all racial/ethnic groups than the proportion of females intending to major in these subjects. (See appendix table 2-30.)

For instance, the proportion of males intending to major in natural science/engineering ranged from 28 percent for American Indian and Puerto Rican males to 37 percent for Asian males. For females, however, the proportion intending to study natural science/engineering was much lower, ranging from 12 percent for Mexican Americans to 16 percent for Asians.

At the time they took the SAT in 1994, only 3 percent of all females intended to study engineering, and females in every racial/ethnic group exhibited the same low priority for engineering study. Black and Asian females intended to major in engineering more often than females of other racial/ethnic groups, but their 5 percent participation was still far below the percentage of males intending to major in engineering (19 percent for blacks and 22 percent for Asians). White and American Indian females were the least likely to choose engineering majors (3 percent each).

Persons With Disabilities

Scholastic Aptitude Test

Four percent of college-bound high school students taking the SAT in 1994 reported a disabling condition; they tended to have lower mean scores on the SAT than did seniors who reported having no disabilities. (See figure 2-13 and appendix table 2-35.) In mathematics, the average score for students with disabilities was 436, compared with 483 for other students. On the verbal exam, the average score for students with disabilities was 391, compared with 427 for students who reported having no disabling condition.

See appendix table 2-35.

[3] Students are counted only once regardless of the number of times they take the same test(s). The College Board reports that these test takers represent approximately 42 percent of all students who enter college each year, and approximately 64 percent of all entering first-year, full-time students. (College Bound Seniors, 1994 Profile of SAT and Achievement Test Takers. Princeton, NJ: Educational Testing Service.)
[4] In 1987 the College Board initiated a review of the Admissions Testing Program, and the SAT Program made significant changes in 1993-1994. Through the January 1994 test administration, SAT Program tests included the SAT, the Test of Standard Written English (TSWE), and the Achievement Tests. Beginning in March 1994, the SAT program was revised into two formats: the SAT I: Reasoning Test (the mathematical and verbal sections, with revisions beginning in March 1994) and the SAT II: Subject Tests (formerly known as the Achievement Tests, with the revisions beginning in May 1994).
The College Board reports that the SAT I: Reasoning Test is comparable to the SAT, and therefore scores from this test are included in trend data in this report, and continue to be labeled "SAT." Changes to the Achievement Tests data are noted in the SAT II: Achievement Tests portion of this report. (Data for the TSWE, which is no longer being administered by the College Board, have never been included in the Women and Minorities series.)

[5] Based on data reported by the test takers themselves, 21 percent of the females had overall grades of A or A+, whereas 16 percent of the males scored that well. (See appendix table 2-25.)
[6] See, for example, "How Does the SAT Score for Women?" National Coalition for Women and Girls in Education. Washington, DC, July 1990, or "Sex Differences in SAT Predictions of College Grades," Lawrence Stricker, Donald Rock, and Nancy Button. The College Board Report. No. 91-2. New York, NY, 1991.
[7] Based on the grading of A = 4 points, B = 3 points, C = 2 points, and D = 1 point.
[8] Females earned a grade point average of 3.26 in English, compared with 3.01 for males; they earned a grade point average of 3.24 in social sciences/history, compared with 3.19 for males.
[9] Seventy-one percent of the males took 4 or more years of mathematics in high school, and 68 percent of the females took that much mathematics. (See appendix table 2-27.)
[10] In 1994, female college-bound seniors reported that they had studied natural science for an average of 3.2 years versus 3.3 years for males. Females earned an average grade point average of 3.09 in the natural science courses they took, versus a slightly lower grade point average of 3.05 for males. The percentage who reported taking an honors course in natural science was identical for both sexes (26 percent).
[11] Through January 1994, the achievement test series included multiple choice exams in 14 academic areas. Beginning in March 1994, the Achievement Tests were expanded and renamed. They are now called the SAT II: Subject Tests to reflect the addition of new test offerings in various subjects. (Results for the science and mathematics tests, as well as for the new mathematics test, math IIC, are presented in appendix table 2-29.) The College Board reports that students who take achievement tests tend to apply to selective colleges and universities.
[12] Biology, chemistry, physics, math I, math II, and math IIC (first introduced in 1994).
[13] Data for Hispanic groups are available separately and are presented in this report at the most detailed level possible. SAT data for Hispanics were subdivided in 1987 from two ethnic groups into three ethnic groups, so that the 10-year trend of specific Hispanic subgroups is not comparable. (The subgroup "Latin American" was available as an option beginning in 1987, in addition to the previously available subgroups "Mexican American" and "Puerto Rican.") Since 1987, scores for those who listed themselves as Latin American tended to be higher than the scores for Mexican Americans or Puerto Ricans.
[14] The SAT's descriptive questionnaire also contains a question on citizenship status.

Top of
Table of