Chapter 2:

Precollege Education

Racial/ethnic and disability status differences in high school completion rates contribute to differences in college enrollment. (Women are as likely to graduate from high school as men—among people age 25 or older, 82 percent of both men and women graduated from high school.) Among all people age 25 or older in 1995, 86 percent of non-Hispanic whites, 74 percent of non-Hispanic blacks, and 53 percent of Hispanics were high school graduates. (See appendix table 2-13.)

Gains in high school completion by blacks in recent years have narrowed the educational gap. In 1975, 87 percent of whites and 71 percent of blacks in the 25 to 29 age group had completed high school. By 1995, 92 percent of whites and 87 percent of blacks in that age range had completed high school.

Hispanics (of any race) have the lowest high school completion rates and have experienced the least gains over time. In 1995, 57 percent of those in the 25 to 29 age group were high school graduates, a modest increase from 53 percent in 1975. The low high school completion rates are partly explained by the large number of foreign-born Hispanics who entered the United States without a high school education. The lower high school completion rates for blacks and Hispanics may also be related to family income. Youths between the ages of 16 and 24 who lived in families with low income levels were eight times more likely to drop out than those from families with higher incomes (McMillan and Kaufman, 1997).

Students with disabilities have an annual dropout rate of 5 percent. Students with disabilities who drop out of school are less likely than those without disabilities to eventually receive high school diplomas or certificates. Drop out and graduation rates vary by type of disability, with those with visual and hearing impairments most likely to have graduated with a diploma. Those with serious emotional disturbances are most likely to have dropped out. (See appendix table 2-14.)

Transition to Higher Education

The United States has one of the highest rates in the world of secondary students who go into higher education and earn college degrees (National Science Board, 1998). The transition from secondary school to college is an important step, not only to the person making it, but also to a nation committed to the education of its citizens in a technological world. This section analyzes data primarily on high school seniors who graduated in 1996, many of whom will earn a college degree in the year 2000.

Usually, many people are involved in the student’s decision to attend a college—students and their parents, along with guidance counselors and teachers. For students from low-income families, however, guidance counselors, teachers, friends, and youth leaders are almost as important as parents in helping to make decisions about post-high-school plans (Gallup International Institute, 1996.) These individuals help students to assess their strengths and weaknesses and to clarify their goals; the earlier in high school they are discussed, the more successful students are in attaining their goals (Rodriguez, 1993). Taking assessment tests during high school also helps students ascertain their strengths and weaknesses and choose suitable colleges.

Several organizations are also involved in assessing student aptitude or achievement. In 1900, the College Board was founded as a national membership association of schools and colleges. The College Board currently administers the Advanced Placement, the Preliminary SAT (PSAT), and the SAT tests through the Educational Testing Service. In 1961, American College Testing (ACT) was founded to measure students’ educational development through the ACT Assessment. None of these tests individually provides data that can be considered nationally representative of all college bound seniors, because students in different states tend to take either the SAT or the ACT (see sidebar on State-by-State breakout of SAT or ACT on page 29), but they do provide information on the transition from high school to college.

Almost three-fourths of high school seniors take either the SAT or the ACT in preparation for applying to college.[11]  Results of these exams are important to students for planning purposes and to colleges for admissions purposes. It is important to note, however, that many students who go to 2-year colleges do not take either test, and that approximately 10 percent of 4-year colleges indicate that SAT and ACT scores are optional for admission (calculation by NSF/SRS based on data from National Center for Fair & Open Testing 1997 and from National Center for Education Statistics 1997 IPEDS surveys, unpublished tabulations).

It is important to note that although test scores can help evaluate a student’s academic preparedness in terms of strengths and weaknesses, "directors of admission say that high school grades are still the most important factor in the selection of a freshman class" (College Board, 1997b, p. 4). The percentage of college admissions directors who indicated that school achievement was very important was 87 percent, compared with 46 percent who indicated that test scores were very important; however, highly selective colleges may base admissions on formulas in which standardized test scores account for as much as two-thirds of the calculation (Hernandez, 1997). Although they do not measure many characteristics necessary for success in college, such as motivation, creativity, and persistence, admissions tests are designed to provide a consistent measure across the variety of curricula and opportunities offered in U.S. high schools.

Women

Women accounted for the majority of test-takers of the Advanced Placement (see page 28) tests (55 percent), the SAT (53 percent), and the ACT (55 percent) in 1996. The number of women taking these tests has increased considerably in the 1990s, and women are also increasing their performance on the tests. In fact, they provide most of the increases seen in total test scores for the ACT and over half in the SAT since 1991, although men have increased their scores also. (See text table 2-2.)

The ACT score changes show greater progress for women, as the overall gap in composite scores between men and women narrowed from 0.5 points in 1991 to 0.2 points in 1996, not a significant difference. The SAT scores indicate that women have made greater progress than men from 1991 to 1996 in both verbal and mathematics mean scores. In this period, the mean scores changed in the following ways:

• verbal increased 8 points for women and 4 points for men;
• mathematics increased 10 points for women and 7 points for men.

Even with this progress for women, women’s scores on average are lower than men’s in 1996 on the SAT. Two ways in which men and women taking the college placement tests differ are in their socioeconomic characteristics and type of coursework taken in high school.

Socioeconomic Differences

From the SAT and ACT student data, it is clear that a larger number of women than men from lower income families choose to take college entrance tests. Although the proportions of test takers from the higher family income groups were about evenly split between males and females, among the lowest income groups, women accounted for over 60 percent of the test takers. (See appendix table 2-15.) Given that parental income is related to average scores (College Board, unpublished tabulations), then this fact would mean that the higher proportion of women test takers who are from low-income families would likely reduce the overall averages for women test takers.

Course-Taking Differences

Do course-taking differences account for test score differences among groups? This was believed earlier, as a 1987 National Academy of Sciences report stated that "The general consensus is that these gender differences in college admission mathematics test scores can be largely accounted for by differences in the amount of mathematics, physical science, and computer programming courses that high school and college-bound women take compared to their male peers" (LeBold, 1987, p. 67). More recent studies have shown that although this is a small part of the explanation, there is "the need for more comprehensive research on gender, race, and SES [socio-economic status] differences in science achievement growth" (Madigan, 1997, p. 12).

Differences between quantitative course taking by female and male high school students have lessened in the 1990s, as shown in earlier sections of this report. Similar proportions of women and men took honors mathematics and science classes (29 percent) according to SAT data for 1996 college bound seniors. (See appendix table 2-16.) In terms of level of classes taken, the percentages of women and men taking higher level mathematics were nearly the same. Only small gaps remained between the percentages of men and women who had taken trigonometry (2 percent), precalculus (3 percent), and calculus (4 percent). In science classes, similar percentages of both men and women had taken biology and chemistry; however, a larger gap existed in the percentage who had taken physics (9 percentage points higher among men than women) (College Board, 1996, SAT unpublished tabulations).

The reduction in differences in mathematics course taking leads to the question, Are the average mathematics scores of men and women who took the same level mathematics class more similar? At the lower mathematics level classes (geometry and trigonometry), differences in the mean SAT scores of men and women persist but are smaller than differences at the higher level classes, such as calculus, according to both SAT and ACT data.

An examination of SAT mathematics test scores for only the students who reported taking the highest level of mathematics (calculus) and science (physics) showed that women scored lower on average than men. Among those who took calculus, women averaged 594 and men 631 on the SAT mathematics test; this difference—37 points—is similar to that for men and women test takers in general (35 points difference). Among those who took physics, women averaged 542 and men 577—a 35-point gap (College Board, 1996, SAT unpublished tabulations).

ACT data also show that women who took calculus and physics reported higher grade point averages than men in their high school mathematics and science classes. In the ACT mathematics and science test sections, however, the average scores of women were also lower than those of men. (See text table 2-3.) Educators and researchers in both the academic community and within the testing organizations have been concerned about the underlying reasons for this disparity. In 1997, ETS released a study (Cole, 1997) that had three interesting results that pertain to this issue.

• In nationally representative samples of 12th graders, "the spread of scores of males tends to be larger than the spread for females. This means that there are more males among the very highest scorers and also more males among the very lowest scorers. Below the 10th percentile and above the 90th percentile, there are about 4 females for every 5 males. We see this low-end result perhaps in the presence of more males in some special education classes…. The high-end result is especially important for self-selected groups, such as those taking high-stakes tests. These groups [students taking the SAT] come from the high end of the distribution and, all other things being equal, we can expect more males than females among such groups and higher average scores for males than for females among such groups" (pp. 16–18).
  
  
• Tests measure particular skills on a single day, whereas grades measure a much wider array of skills. "In fact, we view grades as likely measuring a constellation of desirable characteristics that we call ‘studenting skills’—skills that are especially valuable in school or in work. These skills may include characteristics such as persistence, follow-through, doing required work, participating, and performing in different contexts…. Years of results in predicting college grades have, for example, shown that grades are most often the single best predictor…. Also, tests have consistently been shown to add to the prediction of college performance beyond that accomplished by grades alone." (See sidebar on SAT, page 25.)
  
  For example, were the SAT used alone, it would slightly underpredict the overall grade-point average of first-year female college students…. One subject, calculus, has yielded larger differences than…most other subjects examined. Earlier results had indicated…underprediction of college calculus grades when the SAT was used alone[12] …. We found that…adding high school grades corrected the underprediction. In fact, using grades alone would have resulted in underprediction of calculus grades for males in those cases" (pp. 19–20). However, both grade point averages and test scores are the key factors in institutional selection, so there may be problems of severe selection bias associated with these findings.
  
  
• "We found that the differences cut both ways and that 12th grade girls have substantially closed the familiar mathematics and science gap over the past 30 years, but there continues to be a fairly large gap in writing skills that boys have not closed." (See figure 2-8.)

There is a gap between men and women in writing, but is it relevant for scientists and engineers? As National Education Goals for the year 2000, Goal 3, Objective 2 recommends that "the proportion of college graduates who demonstrate an advanced ability to think critically, communicate effectively, and solve problems will increase substantially" (National Education Goals Panel, 1997). "Good communication skills, both verbal and written, rank high among the top priorities of those in business and industry" (Barabas, 1990). Employers of engineering graduates rated speaking/writing as one of the most important areas of competence, and yet these same areas were identified as the most deficient in these graduates (Kimel and Monsees, 1979).

College Credit From Advanced Placement

Half of the high schools in the United States offer Advanced Placement (AP) college-level classes, and the number of students taking AP tests for college credit is increasing each year, reaching over 500,000 students in 1996. Of these, 200,000 qualified for college credit by earning a certain score on AP exams taken in high school (College Board, 1996c).

Of the subjects offered for AP exams, 12 are science or mathematics subjects. Women constituted over half of the test takers in psychology and biology. These subjects were followed by calculus AB[13]  (47 percent female) and chemistry (42 percent female). The science subject in which women were least likely to take an AP exam was computer science, where women accounted for only 20 percent of the computer science A and 12 percent of the computer science AB[14]  test takers. (See appendix table 2-17.)

Underrepresented Minorities

Introduction

Students who decide to go to college usually take a college admissions test if they are planning to apply to a 4-year college or university, but may not take the test if they are planning only to apply to a 2-year college. Two-year colleges play a somewhat larger role in the higher education of minority students than they do of white students. (See text table 2-4.)

College admission test data for "1996 college bound students," then, may not be representative of all students who went on to college, especially for Hispanic and American Indian students.

For those who decide to take a college admission test, registration includes a student questionnaire. Both the ACT and the SAT student questionnaires ask for race/ethnicity.

• On the form for the SAT, students are asked "How do you describe yourself?" and one of the categories offered is an "other" category. Eleven percent of 1996 SAT test takers chose "other" or did not respond to this question.
• On the form for the ACT, students are asked "Which best describes your racial/ethnic background?" and "other," "multiracial," and "prefer not to respond" categories are offered. Eleven percent of 1996 ACT test takers chose one of these three categories.

Therefore, the data provided on ACT and SAT by race/ethnicity groups cover 89 percent of both ACT and SAT test takers.

There are some differences in the proportions of each racial/ethnic group who take the SAT versus the ACT. (See text table 2-5.) Some of these differences reflect the distinct differences found by State (see sidebar on page 29). For example, American Indian/Alaskan Natives are much more likely to take the ACT and Asian/Pacific Islanders are much more likely to take the SAT.

Note that the data presented in this section do not include students in Puerto Rico. The ACT is not given in Puerto Rico and although an SAT-equivalent test is given in Puerto Rico, those test scores are not included in the SAT data for Hispanics. Since Puerto Ricans ages 16 to 19 years living in Puerto Rico account for 14 percent of the Hispanic population in the United States (Puerto Ricans living on the continent account for another 10 percent of the Hispanic population) (U.S. Department of Commerce, 1993a, b), a significant part of the Hispanic population overall is not included in this section. The number of high school graduates in Puerto Rico is increasing (in 1996, a total of 36,600 students graduated from both public and private high schools in Puerto Rico), and many of these graduates consider going to college and taking the SAT. The SAT-equivalent test administered in Puerto Rico by the College Board is given in Spanish; for academic year 1995–1996, 32,490 persons took the test, of whom 58 percent were women. Only about one-fourth of persons who took the test had a parent who had a college degree (College Board, 1995), so most of these students were the first generation in their families to go to college.

Increased Participation in College Admissions Test Taking

One of the major pieces of information derived from data on the ACT and SAT is that the number of high school seniors from underrepresented minority groups taking college admissions tests has increased significantly in the 1990s. Although the number of white students taking the SAT declined slightly (–1 percent) between 1991 and 1996, the number of minority students taking the test has increased 13 percent (the largest percentage increases were among Mexican American and Latin American students). (See appendix table 2-18.) During this same period, the number of ACT test takers increased 11 percent for whites and 29 percent for minority students. (See appendix table 2-19.)

The College Board data indicate that a higher proportion of underrepresented minorities make the decision to take a college admission test late in high school. Students who have plans for college early in high school often begin by taking a commercial "diagnostic SAT" in 10th grade, the PSAT in 11th grade, sometimes an SAT preparation class, and then the SAT in 11th grade and/or 12th grade. Parents play a major role in educating their children to this schedule and process, particularly if the parents have gone to college themselves. But college experience of parents varies greatly by racial/ethnic group, with blacks, Hispanics, and American Indians having the lowest percentages of parents with college degrees. (See figure 2-9.) Among these underrepresented groups, the majority of high school seniors taking the SAT are trying to be the first generation of their family to go to college. (See appendix table 2-20.)

Although guidance counselors and teachers in the schools may help some "first-generation" students understand this process of planning for college, data on the percentage of SAT takers who had taken the PSAT earlier indicate that this process was not taking place for a higher percentage of students in these underrepresented groups. The students with the lowest percents who had taken the PSAT were Hispanic, American Indian, and black men; the next lowest percentages were among the women in these groups. (See figure 2-10.)

Increased Preparation for College

The American College Testing service has been tracking the percentage of their test takers who have taken core courses in high school: 4 years of English and 3 years each of mathematics, science, and social studies. The number of students who have taken at least a core curriculum in high school has been increasing greatly, from 51 percent in 1991 to 61 percent in 1996. Among underrepresented minorities, the increases have been significant also: for blacks—an increase from 45 to 55 percent of test takers; for American Indians—an increase from 40 to 49 percent; and for Hispanics—an increase from 49 to 56 percent. The increase for Asians was 64 to 71 percent of test takers (American College Testing, 1996).

Data on test takers by race are also available by family income. Average ACT scores for each racial/ethnic group are higher for those with core curriculum preparation, as would be expected; average scores also increase with family income. (See appendix table 2-21.)

College Credit From Advanced Placement

The number of high school students enrolling in AP courses and then taking AP tests for college credit has increased considerably in the last few years. Over half of the high schools in the United States offer AP college level classes; in 1996, over 500,000 students enrolled in AP classes. Of these, 40 percent qualified for college credit from their scores on AP exams taken in high school (College Board, 1996c).

Of the subjects offered for AP exams, 12 are in science or mathematics. The AP courses with the highest number of underrepresented minorities enrolled were calculus AB and biology. The proportion of underrepresented minorities among the AP candidates in these two subjects was 9 percent for both; for chemistry and physics B, the representation was 8 and 7 percent, respectively. (See appendix table 2-22.)

According to the American Association of Collegiate Registrars and Admissions Officers (AACRAO), admissions officers are impressed by Advanced Placement course taking—even if the test is not taken. This importance makes it necessary to understand some of the differential opportunities for minority students to take the AP classes. The Advanced Placement at the College Board indicated that the schools that don’t offer AP classes fall into four categories:

• small religious schools,
• small rural schools,
• schools in Wyoming, the Dakotas, and a few other States, and
• large urban schools, particularly if they feed into academic magnet schools.

Although the first three groups may affect large numbers of American Indians, these and the last category may affect large numbers of black and Hispanic students. Wade Curry, head of the AP at the College Board, explained that "magnet high schools draw off the most academically inclined students and produce fairly large numbers of AP scholars. But the schools they draw from then tend not to offer the courses, leaving those students behind." He has found that "African American students who do well on the AP tests tend to be either in the urban magnets or in predominantly white, suburban school systems where there are between five and twenty African American students who take the courses" (Chenoweth, 1997, p. 22).

Percentage of High School Graduates[1]  Taking the SAT or ACT by State: 1996 States Percentage taking SAT Percentage taking ACT States Percentage taking SAT Percentage taking ACT National 41 35 Missouri 9 63 Alabama 8 57 Montana 21 54 Alaska 47 35 Nebraska 9 72 Arizona 28 27 Nevada 31 39 Arkansas 6 64 New Hampshire 70 3 California 45 11 New Jersey 69 2 Colorado 30 60 New Mexico 12 59 Connecticut 79 2 New York 73 16 Delaware 66 4 North Carolina 59 11 District of Columbia 50 5 North Dakota 5 77 Florida 48 34 Ohio 24 58 Georgia 63 16 Oklahoma 8 63 Hawaii 54 15 Oregon 50 11 Idaho 15 59 Pennsylvania 71 6 Illinois 14 67 Rhode Island 69 1 Indiana 57 19 South Carolina 57 13 Iowa 5 64 South Dakota 5 65 Kansas 9 70 Tennessee 14 77 Kentucky 12 62 Texas 48 30 Louisiana 9 73 Utah 4 66 Maine 68 2 Vermont 70 3 Maryland 64 9 Virginia 68 5 Massachusetts 80 5 Washington 47 15 Michigan 11 64 West Virginia 17 54 Minnesota 9 59 Wisconsin 8 63 Mississippi 4 68 Wyoming 11 64 [1] Based on number of high school graduates in 1996 as projected by the Western Interstate Commission for Higher Education and number of students in the class of 1996 who took the SAT or ACT. NOTES: Puerto Rico is not included. A very low percentage of students may have taken both tests. SOURCE: College Board (1996c) and American College Testing (1996).

Persons With Disabilities

Data on disability from the SAT and ACT are collected in two ways: from student questionnaires and from requests for special testing accommodations.

Student Questionnaires

When persons first register to take college placement exams, they are asked on the student questionnaires to (1) "Indicate any permanently disabling condition," with six response choices (SAT), or (2) "Please respond to this item only if you have a physical or diagnosed learning disability," with eight categories (ACT). The ACT states on the student questionnaire that the information is used by colleges only to provide financial aid and special services.

About 5 percent (over 40,000) of students taking the SAT in 1996 checked a category indicating a disability; the ACT also had 5 percent (almost 20,000) of students who indicated a disability category in 1996. Among college-bound seniors in 1996 who indicated a disability and took the SAT, 51 percent were male and 49 percent female; underrepresented minorities accounted for 18 percent of those indicating a disability (compared with 21 percent for all college-bound seniors) (ACT and SAT, unpublished tabulations).

The average SAT scores in 1996 of those who indicated a disability were lower than those who did not. (See text table 2-6.)

Data from the ACT are reported according to the category of disability and show variability among the groups of students in terms of their average scores in 1996. (See text table 2-7.)

Special Testing Format/Conditions

Among SAT test takers, almost 20,000 students took the test under nonstandard conditions. These test takers had average scores (463 verbal and 452 mathematics) that were below the average of all test takers who indicated on the student questionnaire that they had a permanent disability yet did not request special testing formats (472 verbal and 468 mathematics) (College Board, unpublished data).

The ACT also has data available on the almost 20,000 students who requested special testing formats. (See text table 2-8.) Many of these students do not fill out student questionnaires and, therefore, may not be included in the data presented above. Data from ACT on special testing formats are not nationally representative of any of the disability categories and do not include all students with disabilities (many of whom took the regular administration of the test). Still, it is interesting to note the variety of conditions of students who were taking the college placement tests under special testing administration (ACT, 1996, unpublished tabulations).

Students who take the ACT under special administration to accommodate a disability are often flagged as such when their test scores are provided to college admissions offices. The College Board funded a study on how the admissions process may be affected by these flags, as well as the validity of admissions tests scores for disabled applicants. The main finding of that study was that students with disabilities were admitted on much the same basis as the other applicants, though in some instances particular groups of applicants were somewhat less likely to be admitted than would be expected (learning disabled) or somewhat more likely to be admitted (for example, hearing impaired) to special programs. The report also states that "Admissions decisions predicted on the basis of either SAT or HSG (high school grades) correlated slightly less with actual decisions in the case of handicapped applicants than others. This result suggests that factors other than SAT and HSG play a slightly larger role in the case of handicapped applicants, though not necessarily to their advantage" (Willingham, 1988, p. 81).

Advanced Placement Exams do not collect data from the students on any possible disability. If requested and approved, however, AP exams are offered at the high school site in a variety of testing alternative procedures, which include special arrangements, extended time, use of a reader or sign language interpreter, or special test editions (for example, Braille, cassette, photo-enlarged). Students must have official certification of their disability on file at the school to verify the need for special testing arrangements. The number of AP science exams taken in May 1997 under special testing arrangements were

Biology	267
Chemistry	112
Physics B	55
Physics A	50

Note that these numbers would not include the disabled students who took AP exams under regular testing conditions.

Footnotes