This report presents findings of the Center for Equal Opportunity’s investigation of undergraduate admissions at the University of California, San Diego (UCSD). It describes the racial and ethnic composition of UCSD applicants, admittees, rejectees, and enrollees, and the racial and ethnic differences in SAT scores and high school grade point averages (GPAs) for these groups. The report also provides a detailed summary of the procedures used to generate these results.

Procedures for Analysis

The report summarizes findings from the UCSD applicant data file for 1995. This data set contains the following information for applicants:

• Racial and ethnic group: non-Hispanic white, Asian American, African American, Hispanic, or American Indian
• High school grade-point average (GPA)
• SAT math score
• SAT verbal score
• Admission status: rejected, admitted but not enrolled, or enrolled

Information on gender, U.S. citizenship status, family income, and financial aid status was not provided.

The statistical analysis and report focuses on the following four racial and ethnic groups: African Americans, Hispanics, Asian Americans, and whites (here meaning non-Hispanic whites). Statistical comparisons were not made with American Indian/Native Alaskan applicants due to the small number of cases reported. Applicants claiming ethnicity as "other" or "decline to state" were also dropped from the statistical analysis, although we believe that these individuals are mostly whites and are treated by UCSD as such. Applicants with missing SAT and GPA data were also dropped from the analyses.

Results

Racial and Ethnic Composition of the Applicant Pool

In 1995, 21,579 persons applied for undergraduate admission to UCSD. Omitting American Indians, those who responded "other," and missing cases from the computations yields 19,880 applicants. Of these, 39 percent of the applicants were Asian American (n=7,852), 44 percent were white (n=8,808), 13 percent were Hispanic (n=2,539), and 3 percent were African American (n=681). (See Figure 1. Note: charts are available by mail upon request).

Racial and Ethnic Composition of Admittees

While the composition of the applicant pool is largely beyond the immediate control of the school, admissions are the responsibility of the UCSD admissions committee. UCSD offered admission to 12,553 persons in 1995, 58.2 percent of the number of applicants. The racial and ethnic composition of acceptances is similar to that of the applicants: 41 percent of those accepted were Asian American (n=4,602), 44 percent were white (n=4,973), 12 percent were Hispanics (n=1,330), and 3 percent were African American (n=280). (See Figure 2.)

Racial and Ethnic Composition of Rejectees

The ethnic composition of the rejectee pool is also similar to the composition of the applicant pool: 38 percent of the rejected were Asian American (n=3,178), 44 percent were white (n=3,696), 14 percent were Hispanic (n=1,147), and 5 percent were African American (n=386). (See Figure 3.)

Racial and Ethnic Composition of Enrollees

The racial and ethnic composition of the enrollees is also similar to that of applicants, admittees, and rejectees: 43 percent of the 1995 entering class were Asian American (n=1,200), 46 percent were white (n=1,295), 10 percent were Hispanic (n=278), and 1 percent were African American (n=42). (See Figure 4.)

Comparisons of Acceptance, Rejection, and Enrollment Rates by Racial and Ethnic Group

In 1995, 53 percent of all applicants were accepted at UCSD. Proportionately fewer African-American applicants were accepted than Hispanic, Asian-American, and white applicants: 43.3 percent of African-American applicants were accepted, 56.7 were rejected; 54.8 percent of Hispanic applicants were accepted, 45.2 percent were rejected; 59.5 percent of Asian-American applicants were accepted, 40.5 percent were rejected; and 58.0 percent of white applicants were accepted, 42.0 percent were rejected. (See Figure 5.)

Overall Yield

The yield, or the proportion of accepted students who actually enroll at UCSD, is 24.4 percent. Among racial and ethnic groups, Asian Americans accepted at UCSD have the highest yield of all ethnic groups, while African Americans have the lowest yield: 25.7 percent of Asian Americans accepted for admission to UCSD enrolled in 1995 as did 25.3 percent of white admittees, 20.0 percent of Hispanic admittees, and 14.2 percent of black admittees. Many of those admitted but not attending UCSD enrolled subsequently in other colleges or universities.

Racial and Ethnic Differences in Academic Qualifications

Method of Statistical Analysis

We do not report group means for grades or SAT scores. Using group means places greater weight on extreme values than is warranted. A few unusually high or low scores can have a substantial effect on the value of the mean. Standard deviations, which are based on squared deviations from the mean, are even more useless for describing the spread of cases for asymmetrical, badly skewed distributions. This is because standard deviations reflect the mathematical square of these extreme values.

The median, however, is far less affected by the values of extreme cases. The median represents the middle of the distribution so that 50 percent of all students have greater scores, and 50 percent have lesser scores. We also report scores at the 25th and 75th percentile, again to deal with the problem of extreme cases. While the median represents the middle of the distribution, the 25th and 75th percentile scores taken together represent the actual spread of scores. For example, a score of 650 at the 25th percentile means that 25 percent of scores were below 650, while 75 percent of scores were above it. A score of 750 at the 75th percentile means that 75 percent of scores were below 750, while 25 percent were above 750.

Math SAT Scores

There are two distinct populations of enrollees at UCSD as measured by SAT math scores: African Americans and Hispanics, on the one hand, and whites and Asian Americans, on the other. There are few overlapping scores between these groups. African-American and Hispanic enrollees have significantly lower math SAT scores than do white and Asian enrollees. (See Figure 6.)

We focus here on the range of scores from the 25th to the 75th percentile for each racial and ethnic group. The interquartile range for each group is as follows: African-American scores range from 480 (25th percentile) to 600 (75th percentile), the median is 540; Hispanic scores range from 480 to 610, the median is 550; Asian-American scores range from 580 to 690, the median is 640; and whites range from 590 to 680, the median is 640.

African-American and Hispanic scores overlap only slightly with those of Asian Americans and whites. The African-American and Hispanic 75th percentile scores (600 and 610, respectively) are roughly the same score as the white and Asian 35th percentile scores.

The meaning of these scores can perhaps be put in more concrete terms by considering the numbers of individuals with scores over 700. Of those enrollees reporting SAT scores (a handful of them do not), there were 3 African Americans (or 8.1 percent of all African Americans reporting scores), 11 Hispanics (4.1 percent), 267 whites (21.2 percent), and 286 Asian Americans (24.3 percent) with these mathematics scores.

Verbal SAT Scores

The racial and ethnic group differences are substantial, but the verbal SAT scores show greater overlap than do math SAT scores. African-American scores between the 25th and 75th percentile range from 435 to 530, the median was 490; Hispanic scores range from 410 to 520, the median was 470; Asian-American scores range from 420 to 550, the median was 490; and white scores range from 490 to 600, the median was 550. (See Figure 7.)

There are only small differences between African-American, Hispanic, and Asian-American scores. White enrollees have significantly higher verbal scores than do members of other groups. The 25th percentile verbal SAT score among whites was equivalent to the African-American and Asian American median scores and slightly higher than the Hispanic median score. Thus about half of all Asian-American, Hispanic, and African-American enrollees have verbal SAT scores equal or less than the bottom quartile of white enrollees. Of those enrollees reporting SAT scores, we find that 3 African Americans (8.1 percent of all African Americans), 22 Hispanics (8.2 percent), 151 Asians (12.8 percent), but 349 whites (27.8 percent) have verbal SAT scores greater than or equal to 600.

High School GPAs

The different racial/ethnic groups are most alike on high school grade point averages (GPAs): African Americans have a median GPA of 3.88, and have 25th and 75th percentile GPAs of 3.50 and 4.12; Hispanics have a median GPA of 3.76, and have 25th and 75th percentile GPAs of 3.52 and 3.97; Asian Americans have a median GPA of 3.86, and have 25th and 75th percentile GPAs of 3.63 and 4.09; whites have a median GPA of 3.92, and have 25th and 75th percentile GPAs of 3.70 and 4.13. (See Figure 8.)

Whites have a higher median GPA than African Americans, who have a higher median GPA than Hispanics and Asian Americans. There were 18 African Americans (or 46.2 percent of all African-American enrollees), 65 Hispanics (24.7 percent), 434 Asian Americans (36.7 percent), and 554 whites (43.8 percent) with GPAs higher than 4.00.

White and Asian-American Rejectees vs. African-American Enrollees

Large numbers of Asian-American and white applicants were rejected by UCSD despite their superior academic credentials:

• 1,909 Asian Americans (or 64.9 percent of all Asian-American rejectees) and 2,041 whites (62.0 percent) who were rejected by UCSD had math SAT scores higher than the African-American enrollee median
• 722 Asian Americans (24.5 percent) and 1,549 whites (47.0 percent) who were rejected had verbal SAT scores higher than the African-American enrollee median
• 583 Asian Americans (19.8 percent) and 1,147 whites who were rejected (34.8 percent) had both higher verbal SAT scores and math SAT scores than did the median African-American enrollee
• 222 Asian Americans (7.2 percent) and 325 whites (9.0 percent) who were rejected had GPAs higher than the median GPA for African-American enrollees
• 49 Asian Americans (1.7 percent) and 138 whites (4.3 percent) who were rejected had higher verbal SAT scores, math SAT scores, and GPAs than did the African-American enrollee medians.

Admissions Policies

In view of the fact that black enrollees differ least from their white and Asian counterparts on high school GPAs and most from them on math SATs, it is of interest to examine the correlations between admissions status, verbal SAT, math SAT, and GPAs. When this is done, we find that GPAs are the strongest predictor of admissions status (r=-.65), followed by the verbal SAT score (r=-.44), and the math SAT score (r=-.43).

Graduation Rates

While we cannot know how these 1995 UCSD freshmen will fare, data available from earlier classes provide a basis for ascertaining their relative chances. UCSD provided data on the percentage that graduated five years after enrollment for the classes of 1988 and 1989:

• 41 percent of the African Americans entering in 1988 graduated in five years, as did 41 percent of those entering in 1989
• 48 percent of Hispanics entering in 1988 graduated in five years, as did 50 percent of those entering in 1989
• 67 percent of Asian Americans entering in 1988 graduated in five years, as did 72 percent of those entering in 1989
• 71 percent of whites entering in 1988 graduated in five years, as did 76 percent of those entering in 1989

How are these differences to be explained? We believe that retention is in considerable part the result of initial preparation. Research indicates that these SATs and high-school grades taken in combination are the best predictors of undergraduate success. This includes both future college grades and the likelihood of eventual graduation. The higher the high-school grade point average and the higher the SAT scores, the better the future enrollees’ grades are likely to be and the better are their chances of eventual graduation. The racial and ethnic disparity reported above is likely due to the fact that African Americans and Hispanics enrolled at UCSD have substantially lower SAT scores than do entering white and Asian students.

Although we do not have the evidence necessary for a complete test of this hypothesis, UCSD has provided some information allowing us a partial test. The class entering in 1989 had the following five-year graduation rates for groupings of students with similar combined SAT scores:

• 35 percent of those with combined SAT scores of 800 or less graduated in five years
• 58 percent of those with combined SAT scores of 801-1000 graduated in five years
• 73 percent of those with combined SAT scores of 1001-1200 graduated in five years
• 74 percent of those with combined SAT scores of 1201-1400 graduated in five years
• 79 percent of those with combined SAT scores of 1401 or more graduated in five years

The entering class of 1989 had the following five-year graduation rates for groupings of students with similar GPAs:

• 53 percent of those with GPAs of 3.29 or less graduated in five years
• 64 percent of those with GPAs of 3.30-3.49 graduated in five years
• 71 percent of those with GPAs of 3.50-3.69 graduated in five years
• 72 percent of those with GPAs of 3.70-3.89 graduated in five years
• 78 percent of those with GPAs of 3.90 or higher graduated in five years

The percentage difference between the top and bottom grouping of SAT scores is 44 percent. The percentage difference between the top and bottom grouping of GPAs is 25 percent. The larger size of the former difference relative to the latter means that it is likely that combined SAT scores are a better predictor of completion rates than are grades.

Earlier we showed that UCSD’s admissions policies weigh grades more heavily than SAT scores, as evidenced by the higher correlation between admission status and GPA than between admission status and SAT scores. We have also shown that this policy cannot be for the purpose of maximizing success rates, because SATs are a better predictor of academic success than are grades. These results, then, support our view that these findings are the consequence of UCSD’s policies of Affirmative Action in admissions.

Conclusion

These findings provide strong, clear evidence of Affirmative Action in undergraduate admissions at UCSD. Racial and ethnic background has not been simply one of many factors used by admissions officers. In all too many cases, it has been the deciding factor, as revealed in racial and ethnic differences on SAT scores among applicants admitted to UCSD. Apart from the unfairness of using non-merit-based criteria in the selection process, these preferences actually hurt the very people they are intended to help. UCSD’s own data on SAT scores, GPAs, and graduation rates show that students with weaker qualifications entering UCSD are less likely to have earned a degree five years later. They may receive preferences going in, but they benefit from none going out. Is it really a good idea to put these students into circumstances for which they are not academically prepared? Or would they be better off attending other schools where they are more likely to earn degrees within five years?