Executive Summary

All of the eight Michigan public colleges and universities that we studied show a qualifications gap between white and black students admitted for enrollment. There was no school at which the black median SAT score, ACT score, or GPA was equal to or higher than the white median for students admitted in 1995 -- showing that most or all of them use Affirmative Action to increase black enrollment. At the University of Michigan in Ann Arbor, for example, the average white admittee scored 230 points higher on the combined SAT (out of a possible 1600), 6 points higher on the ACT (out of a possible 36), and nearly half a point higher on grades (on a 4-point scale) than the average black admittee.

Most of the schools studied also exhibit a qualifications gap between white and Hispanic students. There were only a few instances in which the Hispanic median SAT score, ACT score, or GPA was higher than the white median for students admitted in 1995 -- suggesting quite strongly that most of these schools use Affirmative Action to increase Hispanic enrollment. At Michigan State University, for instance, the average white admittee scored 100 points higher on the combined SAT, 3 points higher on the ACT, and two-tenths of a point higher on grades than the average Hispanic admittee.

There is no evidence proving that Asian Americans benefit from Affirmative Action at the schools studied, and there is strong evidence suggesting that they in fact do not receive special considerations at all.

Schools routinely reject white and Asian students with higher test scores than black and Hispanic students who are admitted. These rejected students, however, usually have lower GPAs than black and Hispanic students who are admitted. Despite this, the University of Michigan in Ann Arbor refused admission in 1995 to hundreds of white and Asian students who had both higher test scores and GPAs than the black admittee median.

The University of Michigan at Ann Arbor is by far the greatest offender among that state’s public colleges and universities when it comes to using Affirmative Action.

The six-year graduation rates of white and Asian students are higher than those of blacks at every school studied and higher than those of Hispanics at all but one. Although we lacked the data to demonstrate conclusively that Affirmative Action depress the graduation rates of blacks and Hispanics -- which we have done elsewhere when the data were available -- this finding is consistent with the hypothesis that preferences have a negative impact on graduation rates.

If Michigan schools were to admit students on a colorblind basis, black admissions would not drop sharply across the board. Only the University of Michigan at Ann Arbor probably would experience significant declines in black enrollment. The effect on other schools probably would be much less apparent, and perhaps negligible. Black students would still have many opportunities in Michigan to earn a higher education.

Introduction

For more than 20 years, Affirmative Action have played a key role in how admissions officers at public colleges and universities in the United States have chosen their schools’ undergraduate classes. A system of Affirmative Action in the admissions process operates by establishing different standards of admission for individuals based on their racial or ethnic background, with some students held to a higher standard and others admitted on the basis of a lower standard. Earlier this century, some colleges and universities denied admission to Jews, blacks, women, and the members of other groups even when their grades, test scores, and other measures of academic achievement surpassed those of white males who were offered an opportunity to enroll. The passage of new civil rights laws in the 1960s made this kind of blatant discrimination illegal.

Since then, however, many colleges and universities created affirmative action programs meant to boost the enrollment of students whose background previously had excluded them for pursuing a higher education--especially blacks and, to a lesser extent, Hispanics--by granting preferences to them during the admissions process. These policies were immediately controversial, and they remain so today. Defenders of Affirmative Action claim that their policies are not discriminatory and help administrators choose between equally or almost equally qualified students, giving a slight edge to people who likely have faced discrimination or come from disadvantaged backgrounds. Critics of Affirmative Action say that these policies are no better than the discriminatory ones they replaced and that the advantages they confer upon certain applicants are much greater than supporters are willing to admit.

Public colleges and universities have seen their ability to use Affirmative Action increasingly restricted in the last two years. The enactment of California’s Proposition 209 (also known as the California Civil Rights Initiative) forbids discriminating against or granting special treatment to any applicant on the basis of race, ethnicity, and sex in the country’s largest state. Grassroots activists elsewhere are bound to consider placing similar proposals on their own state ballots, and lawmakers in state capitals may draft legislation modeled on the new California law. The University of Michigan is currently facing multiple lawsuits charging it with a preferential admissions policy. (In 1992, about 84 percent of all Michigan residents were non-Hispanic white, 2 percent were Hispanic, 14 percent were black, 1 percent were Asian, and less than 1 percent were American Indian, according to the U.S. Bureau of the Census.)

This study examines the extent to which Affirmative Action are used in a cross-section of Michigan’s public universities and attempts to determine how the elimination of racial preferences would affect the enrollment patterns of these schools. It submits actual admissions data supplied by the schools themselves to a rigorous statistical analysis, and its conclusions are the result of one of the most extensive investigations of how Affirmative Action operate in the undergraduate admissions process of Michigan’s institutions of higher education.

The report is the latest in series published by the Center for Equal Opportunity, a Washington, D.C.-based public policy research organization. Earlier CEO studies have focused on the public colleges and universities of Colorado as well as the University of California at Berkeley and the University of California at San Diego. The previous reports have shown that blacks and Hispanics receive enormous preferences in the undergraduate admissions process at the schools studied.

Methods of Study

The eight universities examined in this study represent a cross-section of Michigan’s public institutions of higher education, from the most selective to the least selective. Each of the schools studied here provided the raw data for our statistical analysis of undergraduate admissions for the fall 1995 class of incoming freshmen. The universities are: Central Michigan University (CMU), Ferris State University (FSU), Michigan State University (MSU), Michigan Technical University (MTU), Northern Michigan University (NMU), the University of Michigan at Ann Arbor (UM), the University of Michigan at Dearborn (UMD), and Saginaw Valley State University (SVSU). Figure 1 provides a list of these schools, followed by their ranking in the 1997 edition of Barron’s Profiles of American Colleges. Several other schools in the state refused to comply with our requests for information or supplied inadequate data. They include Eastern Michigan University, Oakland University, University of Michigan at Flint, Wayne State University, and Western Michigan University.

We received data from four schools--MSU, MTU, UM, and UMD--on individual applicants, admittees, and rejectees, regarding ethnic group membership, verbal and math SAT scores, ACT scores, and grade-point averages (GPAs), gender, and high-school rank. We received data from Ferris State on individual applicants, admittees, and rejectees containing ACT scores and GPAs, while CMU, NMU, and SVSU provided data on enrollees only.

We omit from our data analyses those cases for which the ethnicity is listed as either other, missing, or unknown. We also omit Native Americans because of their small numbers. Lastly, we omit cases with missing academic data.

We do not report group means for grades or test 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 less useful 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 and related statistics, however, are less affected by the values of extreme cases. The median represents the middle of the distribution. Fifty percent of all students have greater scores than the median, and 50 percent score lower.

We also report scores at the 25th and 75th percentiles, 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 combined SAT score of 800 at the 25th percentile means that 25 percent of combined scores were below 800 while 75 percent of scores were above it. A score of 1400 at the 75th percentile means that 75 percent of scores were below 1400 while 25 percent of student scores were above 1400.

We look at both SAT and ACT scores. The SAT and ACT are not quite comparable tests. SAT scores range from 200 to 800 on the verbal and math tests; combined scores range from 400 to 1600. In contrast, the ACT is actually four tests on English, reading, mathematics, and science reasoning. A student’s raw scores for all four tests are combined and converted into scaled scores. The scaled scores are converted into one composite ACT score. The possible range of composite ACT scores is from 1 to 36. Grade-point averages are calculated on a four-point scale.

Applicants, Admissions, and Enrollments

Figure 2 shows the relatively high admissions rates at the Michigan schools studied. Most who apply are admitted. On average, 85 percent of black applicants were admitted, along with 91 percent of Hispanic and Asian American applicants, and 89 percent of white applicants. The admissions rate of Asian American applicants is higher than that of white applicants at every school.

In all schools but one, the white admissions rate is greater than the black admissions rate. The exception is the University of Michigan at Ann Arbor, where the black admissions rate exceeds the white admissions rate by nine percent.

The pattern is mixed when we compare admissions rates between whites and Hispanics. Hispanic and white applicants were admitted at similar rates at FSU, MSU, and MTU. Whites were admitted at a 5 percent higher rate at UMD. At UM, the Hispanic admissions rate exceeds the white admissions rate by 18 percent.

Differences in Qualifications

Figure 3
White-Black Admittee Differences
in Median Verbal SAT Scores
We examine three sets of differences in qualifications: white-black, white-Hispanic, and white-Asian. Treating each pair of comparisons separately makes it easier to see whether substantial differences in racial and ethnic preferences exist and for which groups they are greatest.

Differences Between Whites and Blacks

Figure 4
White-Black Admittee Differences
in Median Math SAT Scores
Whites on average have better credentials compared to blacks, regardless of whether we examine SATs, ACTs, or grade-point averages. Figures 3,4, and 5 show the white-black gap in verbal and math SAT scores, the white-black gap in median ACT scores, and the gap in median GPAs.

Four schools provided SAT data on blacks and whites. For verbal SAT scores, the white-black gaps range from a high of 100 points at UM to a low of 60 points at MSU. The white-black gaps are even more substantial for math SAT scores, ranging from a high 140 at UMD to 110 at MTU.

Seven schools provided ACT data on blacks and whites. Gaps range from a high of 6.0 at UM, down to 2.0 at SVSU.

Figure 5
White-Black Admittee Differences
in Median ACT Scores
The gaps are smallest regarding grades. The largest gap between whites and blacks is less than a half grade-point. The white-black gap in median GPAs range from a low gap of 0.17 points at NMU to a high of 0.45 at SVSU.

The size of the white-black gap varies depending on the school and the requirements. The pattern however is clear. There is no case in which the black median is equal to or greater than the white median, be it grades or test scores. For four different measures of qualifications (verbal SAT scores, math SAT scores, ACT scores, and grades), there are 23 cases (four schools for the verbal SAT scores, four schools for the math SAT scores, seven for the ACTs, and eight for grades) in which the white median is higher than the black median, and none in which the reverse is true. The probability of this happening by chance is exceedingly small. It is the same as the probability of flipping a coin and getting 23 heads in a row.

Figure 6
White-Black Admittee Differences
in Median GPAs

This is incontrovertible evidence of the use of Affirmative Action to increase black enrollment at Michigan’s public universities.

Differences Between Whites and Hispanics

Figure 7
White-Hispanic Admittee Differences
in Median Verbal SAT Scores
White-Hispanic differences are smaller than the ones between whites and blacks. In two cases, the average Hispanic admittee performed as well as or better than the average white.

At UM and MSU, the average verbal SAT score for whites is higher than that of Hispanics. The average is higher for Hispanics at Michigan Tech. At all three schools, the average white has a larger math SAT score compared to the average Hispanic.

Figure 8
White-Hispanic Admittee Differences
in Median Math SAT Scores
Seven schools provided ACT data on Hispanics and whites. At all seven schools, the average white scored higher compared to the average Hispanic. The white-Hispanic ACT gap ranges from a high of 4.0 at UM to a low of 1.0 at FSU and SVSU. As for GPAs, gaps range from 0.30 at UM, down to no gap in average grades at UMD. All white-Hispanic gaps in median GPAs are less than a half grade-point.

For four different measures of qualifications (verbal SAT scores, math SAT scores, ACT scores, and grades), there are 20 cases of white-Hispanic comparisons (three schools for the verbal SAT scores, three schools for the math SAT scores, seven for the ACTs, and seven for grades). In 18 cases, the white median is higher than the Hispanic median.
Figure 9
White-Hispanic Admittee Differences
in Median ACT Scores
In two cases, the Hispanic median is equal to or greater than the white median. The probability of these gaps happening by chance is about 1 in 2500.

This is striking evidence of Affirmative Action being used to increase Hispanic enrollment in Michigan’s public universities.

Figure 10
White-Hispanic Admittee Differences
in Median GPAs

Differences Between Whites and Asians

Figure 11
White-Asian Admittee Differences
in Median Verbal SAT Scores
There is no discernible pattern when comparing whites and Asians. The white-Asian gap in verbal SAT scores ranges between 50 points at UMD, to -10 at UM (where the median Asian verbal SAT is higher than that of whites). For median math SAT scores, the average Asian admittee has a score equal to or greater than that of the average white admittee at all four schools. The white-Asian gap in ACT scores ranges from 2.0 to -1.0. At three schools, the average white outscores the average Asian, but at two others, the median scores are the same, and at MTU, the average Asian outscores the average white. The range of differences is even smaller regarding grades. The largest gap favoring whites is at UMD, but the gap is only one-tenth of a grade-point, while the largest gap favoring Asians is roughly the same.

Figure 12
White-Asian Admittee Differences
in Median Math SAT Scores
Overall, there are 20 cases of white-Asian comparisons. In four cases, there is no gap. Asian qualifications are higher in seven cases, while in nine cases, white qualifications are superior.

These data suggest strongly that Asian Americans do not receive Affirmative Action from Michigan’s public universities.

Figure 13
White-Asian Admittee Differences
in Median ACT Scores

Figure 14
White-Asian Admittee Differences
in Median GPAs

Biggest Gaps at the Best Schools

Admitting students based on Affirmative Action results in schools accepting minorities with lower test scores and grades than white students at the same school. This creates a situation in which a group of enrollees who were admitted under a scheme of Affirmative Action actually have the same qualifications as whites of a lesser school. Admissions officers essentially reach down into the applicant pool and pull up certain students. This practice generally results in at least some whites and Asians with better credentials than black and Hispanic enrollees being rejected from the same schools, despite their superior qualifications.

Critics of race preferences have argued that as better schools reach down into the applicant pool to accept minorities, creating a gap in qualifications between white and minority enrollees. If this is the case, then there should be a positive relationship between the quality of the school and the white minority gap in qualifications.

One way to show this gap is to combine the data reported above. A useful way to do this is to develop models which predict the probability of admissions at each school for members of the different racial and ethnic groups, holding constant their qualifications. This can be best done by computing prediction equations for the admissions decision by race and ethnicity and including test scores and high school grades as statistical control variables.

The technique we employ for this purpose is logistic regression. We report the admissions odds ratios derived from these equations as a means of measuring the extent of racial preferences at any particular school. Odds ratios are somewhat like correlation coefficients, except that unlike correlation coefficients, which vary between -1 and 1, odds ratios vary between zero and infinity. An odds ratio of 1 is equivalent to a correlation coefficient of zero. In both cases there is no relationship between the two variables. An odds ratio of greater than 1 means that the odds of members of one group being admitted is larger than those of members of the other group being admitted. This is something like a positive correlation: an odds ratio of less than 1 is equivalent to a negative correlation. For a more complete discussion, please see Applied Logistic Regression by David W. Hosmer and Stanley Lemeshow (John Wiley and Sons, 1989).

We report the odds ratios for the admission decision for blacks relative to whites (See Figure 15). Because of the way that the data for MSU and MTU were provided, we were not able to adjust for both grades and ACTs at the same time. This probably understates the degree of preferences exhibited at these schools. However, we report two separate results for each school, one controlling for GPAs and the other controlling for ACTs.

We find a correspondence between the quality of the school and the magnitude of the odds ratio. The odds of a black applicant with the same qualifications as a white applicant being offered admission to the University of Michigan at Ann Arbor is 173.7 to 1. In other words, if two students with equal test scores and grades apply to UM, and UM has only one spot to offer them, it is almost 174 times more likely to choose a black applicant instead of a white one.

At UMD, the odds ratio is not nearly as high, though still quite substantial: 36.5 to 1. The ratios at other schools are smaller but measurable. At MSU it is 3.26 to 1 (controlling for GPA) and 1.80 (controlling for ACTs) to 1, at MTU it is 1.46 to 1 (controlling for GPA) and 2.67 to 1 (controlling for ACTs), and at Ferris State it is 1.79 to 1. (This last result--as well as the result for MTU’s grades--is not statistically significant. All the other results are statistically significant at the p<.05 level or better.)

These results reveal several things. First, the degree of Affirmative Action exhibited at both UM and UMD are astonishingly large. Second, blacks are admitted at a higher rate than whites controlling for their qualifications. Third, the degree of preference exhibited roughly declines with the selectivity of the school in question. Ferris State, which is a virtually open-admissions university, exhibits little or no Affirmative Action in admissions. The University of Michigan at Ann Arbor, which is by far the most selective school, exhibits by far the largest degree of racial preference. MSU and MTU also exhibit evidence of preferences, but less than at either UM or UMD. (Using SATs instead of ACTs as a control variable makes no difference in the degree of preference accorded to black admittees.)

How Preferences Affect Graduation Rates

If students gain admission to college for any reason other than their academic preparation, it is likely that they will face more hurdles in school compared to their peers who have been admitted under a higher standard. They may, in fact, fail to earn their degrees. So it makes sense that Affirmative Action will have a negative effect on the graduation rates of students who are said to benefit from them. Indeed, the six-year graduation rates for whites and Asians, who do not receive preferences in the admissions process, are much higher than those of blacks and Hispanics, who do receive preferences, at almost every school examined.

Graduation rates at the six schools for which we have data are shown in Figure 16. White graduation rates are higher than both black and Hispanic rates at all schools. The gap between white and black graduation rates range from 12 to 29 percent. The white-Hispanic gap in graduation rates ranges from five to 20 percent. Of course, students can drop out of college for many reasons. The lower rate of graduation for blacks and Hispanics could be due to economic factors. Because we do not have data on all of the students applying to all of the public colleges and universities in Michigan, we are not able to determine conclusively that Affirmative Action depress the graduation rates of minorities--as we were able to do in Colorado, for example, because we had the necessary data. Nevertheless, our findings on the differences in graduation rates among racial or ethnic groups are consistent with the hypothesis that preferences have a negative impact on graduation rates.

Black Admissions vs. White Rejectees

The use of Affirmative Action invariably leads to less qualified black students taking the place of more qualified white (and sometimes Asian-American) students. Only four schools submitted data on rejectees. Figure 17 compares median test scores and grades of white rejectees with those of black admittees.

At two of the four schools, the median verbal SAT of white rejectees is higher than the median of black admittees. At three of four schools, white rejectees have higher median math SATs compared to black admittees. As for ACTs, white rejectees and black admittees have the same median ACT at three schools. At MTU, the average white rejectee has a higher score, while black admittees have higher ACTs compared to white rejectees at UM.

Grades are another matter. The median grade-point averages of white rejectees at all four schools are lower than those of the median GPAs of black admittees. The smallest gap is at UM, where the median GPA of black admittees is 0.40 points higher than that of white rejectees. This gap also exists when comparing the 75th percentile scores of white rejectees to black admittees, although it is smaller.

Ascertaining the actual numbers of individuals enables us to determine the extent to which a substantial number of white rejectees are as or better qualified than the average black admittee The gap is by far the largest at UM. We find that there were 564 whites rejected at UM with both grades and ACT scores higher than those of the average black admitted to UM. There were 224 white rejectees with SAT scores and grades higher than those of the average black admittee.

At UMD, we find that 78 white rejectees had higher ACTs than the median black admittee and three white rejectees had higher GPAs than the black median admittee. There was one white admittee with both ACT scores and GPAs higher than the black admittee median. At MTU there were 2,284 white rejectees with higher ACTs than the black median and nine with higher grades than the black acceptee median. At MSU there were 871 whites with higher ACTs than the black acceptee median and seven whites with higher GPAs than the black acceptee median. At Ferris State, there were three white rejectees with higher GPAs than the black admittee median.

What Colorblind Admissions Would Mean for Michigan

If Affirmative Action ended in Michigan public universities, would minorities be locked out of opportunities to pursue a higher education? To address this question, we compared the median ACT scores and GPAs of black admittees at the eight schools in this report and compared them to the 25th percentile to whites admitted at the schools (See Figures 19 and 20). If the median score equaled or surpassed the white 25th percentile score in any of the categories, we assumed that the applicant could gain admission to that school.

Figure 19
Probable Access to Schools Based on Median ACTs of Black Admittees

Figure 20
Probable Access to Schools Based on Median GPAs of Black Admittees

If based solely on ACT scores, the average black admittee at the University of Michigan at Ann Arbor would qualify for all schools in our study except UM. Grade-point averages, however, show that the average black admittee at MSU, MTU, UM, and UMD would qualify for all schools but UM. Those at the other four schools would be more restricted in choices, but there is no place where the median GPA for black admittees falls below the white score at the 25th percentile. To claim that minorities would be cut off from higher education if admissions were colorblind is not sustained by this analysis.

There are even fewer restrictions on the top quartile of black admittees at these schools, especially for those at the more elite schools. As Figure 21 shows, the top quartile of black admittees based on ACTs currently admitted to MTU and UM would be eligible for all schools but one. Those similarly situated at MSU would have access to all schools but MTU and UM, while those at UMD would have access to four schools.

Based on GPAs, the top quartile of black admittees has more choices (see Figure 22). Those at MTU, UM, and UMD would qualify for all schools, while those at CMU, MSU, and NMU would qualify for all but one. The most restricted would be those at Ferris State. However, there are no cases where the top quartile of black admittees would fail to meet the minimum cut-off.

Figure 21
Probable Access to Schools Based on Top-Quartile ACTs of Black Admittees

In short, black admittees at all these schools would have access to higher education, if not always at the most selective ones. Colorblind admissions probably would reduce the number of black students attending the University of Michigan. Other schools would be significantly less affected. Clearly, talk of colorblind admissions ending access to higher education for minorities is highly exaggerated.

Figure 22
Probable Access to Schools Based on Top-Quartile GPAs of Black Admittees

Central Michigan University

There were 2,730 enrollees at Central Michigan University in 1995.

Central Michigan Enrollees:

• 3.8 percent black
• 1.5 percent Hispanic
• 1.1 percent Asian
• 93.6 percent white

Differences in Enrollee ACTs

Figure 23
Enrollee ACT Scores, Central Michigan University
Figure 23 shows ethnic group differences in ACTs. The gaps are largest between whites and blacks. White scores are the highest at the 25th, 50th, and 75th percentiles, while black scores are the lowest at the same percentiles. The white-black gap in median ACT scores is four points. The black median is lower than the white score at the 25th percentile, while the black score at the 75th percentile is two points lower than the white median.

Hispanic scores are slightly lower that white scores at the 50th and 25th percentiles, although they are the same at the 75th percentiles. Asian scores are lower than Hispanic scores at all three percentiles.

Differences in Enrollee GPAs

Figure 24
Enrollee GPAs, Central Michigan University
Figure 24 shows ethnic group differences in GPAs. Gaps are significantly smaller when comparing GPAs, although blacks have the lowest GPAs at all three percentiles. The white-black gap in median GPAs is 0.44, which is less than half of a grade-point. The median GPA for black enrollees however is lower than that of whites at the 25th percentile (and is only slightly higher than Hispanic and Asian GPAs at the same percentile). This means that the average black at CMU was admitted with lower average grade-points compared to Hispanics, Asians, and whites.

Six-Year Graduation Rates

• 43 percent of blacks
• 41 percent of Hispanics
• 55 percent of whites
• 51 percent of Asians

Whites graduate at the highest rate, followed by Asians, blacks, and Hispanics. The white-black gap in graduation rates is 12 percent, while the white-Hispanic gap is 14 percent. The white-Asian gap is four percent.

Ferris State University

Of the 4,400 persons who applied to Ferris State in 1995, 96.7 percent (n=4,254) were admitted and 39.4 percent (n=1,732) enrolled.

Ferris State Applicants:

• 21.0 percent black
• 1.7 percent Hispanic
• 1.1 percent Asian
• 76.0 percent white

Ferris State Admittees:

• 20.0 percent black
• 1.8 percent Hispanic
• 1.1 percent Asian
• 77.1 percent white

Ferris State Rates of Admission:

• 91.9 percent of blacks
• 96.6 percent of Hispanics
• 100 percent of Asians
• 98.0 percent of whites

Overall admissions rate: 96.7 percent

Ferris State Enrollees:

• 16.5 percent black
• 1.5 percent Hispanic
• 1.1 percent Asian
• 80.0 percent white

Differences in Admittee ACTs

Figure 25
Admittee ACT Scores, Ferris State University
Figure 25 shows ethnic group differences in ACTs. White scores at the 25th, 50th, and 75th percentiles are higher than Asian, Hispanic, and black scores at the same levels. The median white score is the same as the Asian and Hispanic ACT score at the 75th percentile. It is higher than the black score at the 75th percentile. The black score at the 75th percentile is the same as the white score at the 25th percentile. This means that 75 percent of blacks admitted to Ferris State score lower than 75 percent of white admittees.

Differences in Admittee GPAs

Figure 26
Admittee GPAs, Ferris State University
Figure 26 shows ethnic group differences in GPAs. Asian GPAs are slightly higher than white GPAs, which in turn are higher than Hispanic and black GPAs. Differences in GPAs are not as large as ACT differences. All are less than half a grade-point. The white-black gap in median GPAs is .25, which is a quarter of a grade-point.

Rejectees vs. Admittees

Composite ACT scores for white rejectees (n=32) are 14, 15, and 18.8 at the 25th, 50th, and 75th percentiles. Scores for black admittees (n=509) are 13, 15 and 17 at the 25th, 50th, and 75th percentiles. Scores for Hispanic admittees (n=52) are 15.3, 18, and 19 at the 25th, 50th, and 75th percentiles. White rejectees’ scores on the whole are lower than those for Hispanic admittees, although scores of white rejectees at the 75th percentile are higher than the Hispanic median. Twenty-five percent of whites rejected by Ferris State had higher ACT scores compared to the average Hispanic admittee.

The median ACT score for white rejectees is the same as that for black admittees, meaning that half the whites rejected by Ferris State had higher ACT scores compared to half the blacks admitted.

The GPAs for white rejectees (n=60) are 1.41, 1.56, and 1.65 at the 25th, 50th, and 75th percentiles. GPAs for black admittees (n=509) are 2.09, 2.45 and 2.82 at the 25th, 50th, and 75th percentiles. GPAs for Hispanic admittees (n=52) are 2.28, 2.51, and 2.98 at the 25th, 50th, and 75th percentiles. GPAs for white rejectees are considerably lower than those for black and Hispanic admittees at comparable percentiles. The median GPA for white rejectees is almost a full grade-point lower than the median GPA for black admittees. The GPA for white rejectees at the 75th percentile is almost a half-point lower compared to black admittee scores at the 25th percentile. Given these extremely low GPAs for white rejectees, it is highly unlikely than any policy of race preference affected admissions decisions at Ferris State, especially if the admissions process gives equal or greater weight to grades compared to test scores.

Six-Year Graduation Rates

• 26 percent of blacks
• 35 percent of Hispanics
• 57 percent of Asians
• 55 percent of whites

There are large differences in graduation rates among ethnic groups. Asians and whites at Ferris State graduate at 20 percent higher rate compared to Hispanics. The gap is greater between Asians and whites versus blacks. The graduation rate for Asians and whites is more than twice the graduation rate of blacks; the latter is roughly one in four.

Michigan State University

In 1995, 24,200 people applied for admissions, 91 percent (n=21,910) were admitted, and 26 percent (n=6,336) subsequently enrolled. The overwhelming proportion of applicants were white.

Michigan State Applicants:

• 11.1 percent black
• 3.1 percent Hispanic
• 5.5 percent Asian
• 80.4 percent white

Michigan State Admittees:

• 10.0 percent black
• 3.2 percent Hispanic
• 5.7 percent Asian
• 81.1 percent white

Michigan State Rates of Admission:

• 82.1 percent of blacks
• 93.5 percent of Hispanics
• 93.7 percent of Asians
• 91.4 percent of whites

Michigan State Enrollees:

• 8.8 percent black
• 2.4 percent Hispanic
• 4.5 percent Asian
• 84.3 percent white

Differences in Admittee Verbal SAT Scores

Figure 27
Admittee Verbal SAT Scores, Michigan State University
Figure 27 shows the range of scores by ethnic groups. White verbal scores at the 25th, 50th, and 75th percentile are higher than comparable scores for blacks, Hispanics, and Asians. The white-black gap in median scores is the largest (60 points). The white score at the 25th percentile is the same as the black median, meaning that 75 percent of white admittees have scores equal to or greater than 50 percent of black admittees.

Asian scores have the greatest range. Asian scores at the 75th percentile are the highest of the four groups, and there is a 30-point gap in median scores between Asian and black admittees, Asian scores are the same as black scores at the 25th percentile, and lower than Hispanic and white scores at the same percentile.

Differences in Admittee Math SAT Scores

Figure 28
Admittee Math SAT Scores, Michigan State University
Figure 28 shows the range of math scores by ethnic groups. Asian scores are highest, followed by whites, Hispanics, and blacks. The Asian-black gap in median scores is 170 points, while the white-black gap is 120 points. Asian and white scores at the 25th percentile are higher than black scores at the 50th percentile, meaning that 75 percent of Asians and whites had higher scores compared to the average black admitted to Michigan State.

Differences in Admittee ACT Scores

Figure 29
Admittee ACT Scores, Michigan State University
Figure 29 shows the range of composite ACT scores. Asian and white admittees have higher scores compared to blacks and Hispanics. The Hispanic median ACT is the same as the white and Asian ACT score at the 25th percentile, while the black median ACT falls below it. This means that the average Hispanic or black was admitted to Michigan State with ACT scores equal to or lower than 75 percent of Asians and whites.

Differences in Admittee GPAs

Figure 30
Admittee GPAs, Michigan State University
Figure 30 shows ethnic group differences in GPAs. Black and Hispanic GPAs are lower than white and Asian GPAs at the 25th, 50th, and 75th percentiles. The differences however are relatively small compared to test scores. The gap in GPAs between Asians and whites versus blacks and Hispanics is less than one-half of a grade-point.

Rejectees vs. Admittees

Verbal SAT scores for Asian rejectees (n=52) ranged from 280 at the 25th percentile to 420 at the 75th percentile, with a median score of 340. Scores for white rejectees (n=426) ranged from 360 at the 25th percentile to 460 at the 75th percentile, with a median score of 410. Scores for Black admittees (n=814) ranged from 370 to 430, with a median of 480. Scores for Hispanic admittees (n=288) ranged from 400 to 510, with a median of 450. Asian rejectee scores at the 25th, 50th, and 75th percentiles do not fall above the black and Hispanic admittee medians, but white scores at the 75th percentile do.

Math SAT scores for Asian rejectees (n=52) range from 440 at the 25th percentile to 600 at the 75th percentile, with a median score of 515. Scores for white rejectees (n=426) range from 397 to 530, with a median score of 460. Scores for black admittees (n=814) range from 390 to 530, with a median score of 450. Scores for Hispanic admittees (n=288) range from 450 to 600, with a median score of 510. Median math scores for white and Asian rejectees are higher than the median score for black admittees, although not compared to Hispanic admittees. This means that the average white and Asian rejectee was denied admission to Michigan State, despite having higher scores compared to the average black admitted. The median score for Asian rejectees is also higher than that for Hispanic admittees.

Composite ACT scores for Asian rejectees (n=54) range from 16 at the 25th percentile to 21 at the 75th percentile with a median of 18. Scores for white rejectees (n=1,490) range from 18 to 22, with a median of 20. Scores for black admittees (n=1,962) range from 17 to 22, with a median score of 20. Scores for Hispanic admittees (n=571) range from 19 to 24, with a median of 21. The median ACT for white rejectees is the same as that for black admittees, meaning that the average white rejected by Michigan State had the same ACT score as the average black admitted. Although the median ACT for white rejectees is lower than the median ACT for Hispanic admittees, the ACT score at the 75th percentile for white rejectees is higher than the Hispanic median. In other words at least 25 percent of whites rejected by Michigan State had higher ACT scores compared to the average Hispanic admittee.

Similarly, the median ACT for Asian rejectees is lower than the medians of black and Hispanic admittees. However, ACT scores at the 75th percentile for Asian rejectees are equal to the Hispanic and higher than the black median score. At least 25 percent of Asian rejectees were denied admission to Michigan State despite having the same or better ACT scores compared to the average Hispanic and black admittee respectively.

GPAs for Asian rejectees (n=62) range from 2.21 at the 25th percentile to 2.55 at the 75th percentile, with a median GPA of 2.34. GPAs for white rejectees (n=1,608) range from 2.37 to 2.71, with a median GPA of 2.55. GPAs for black admittees (n=1,540) range from 2.86 to 3.16, with a median GPA of 3.47. GPAs for Hispanic admittees (n=476) range from 2.91 to 3.22, with a median GPA of 3.52. Comparisons of GPAs, unlike test scores, show black and Hispanic admittees with considerably higher GPAs compared to Asian and white rejectees at the same percentiles. The median for black and Hispanic admittees are more than half-a-grade-point higher (0.61 and 0.67 respectively) compared to the median GPA of white rejectees. Even Asian and white rejectees at the 75th percentile have GPAs lower than the median black and Hispanic admittee GPAs.

Six-Year Graduation Rates

• 45.80 percent of blacks
• 54.57 percent of Hispanics
• 69.47 percent of Asians
• 71.48 percent of whites

Whites graduate at the highest rate, followed by Asians, then Hispanics and blacks. Roughly 26 percent more whites and 24 percent more Asians compared to blacks graduate from Michigan State. Hispanics fall between the black and white rates. (These figures are based on the Fall 1995 averages for seven freshmen classes, from 1983 to 1989.)

Michigan Technical University

In 1995, 2,689 people applied to MTU, 95 percent (n=2,545) were admitted, and 47 percent (n=1,263) enrolled.

Michigan Tech Applicants

• 4.0 percent black
• 1.4 percent Hispanic
• 1.8 percent Asian
• 92.8 percent white

Michigan Tech Admittees:

• 3.9 percent black
• 1.4 percent Hispanic
• 1.8 percent Asian
• 92.9 percent white

Michigan Tech Rates of Admission:

• 91.7 percent of blacks
• 92.1 percent of Hispanics
• 95.8 percent of Asians
• 94.7 percent of whites

Overall admissions rate: 94.6 percent

Michigan Tech Enrollees:

• 29.6 percent black
• 34.2 percent Hispanic
• 29.2 percent Asian
• 48.2 percent white

Differences in Admittee Verbal SAT Scores

Figure 31
Admittee Verbal SAT Scores, Michigan Technical University
Figure 31 shows the range of scores by ethnic groups. Hispanic scores at the 25th and 50th percentile are higher than white and Asian scores, while black scores are lower. The black score at the 75th percentile is lower than the Asian and white medians, and the same as the Hispanic score at the 25th percentile. This means that 75 percent of blacks are admitted to MTU with lower verbal scores than the average Asian and white admittee, and with lower verbal scores than 75 percent of Hispanics admitted to MTU.

Differences in Admittee Math SAT Scores

Figure 32
Admittee Math SAT Scores, Michigan Technical University
Figure 32 shows the range of math scores by ethnic groups. Asians and whites have comparable scores at the 25th, 50th, and 75th percentiles. Hispanic scores are somewhat lower, while black scores are lower still. The white-black gap in median scores is 110 points. The Hispanic-black gap is 70 points.

White and Asian scores at the 25th percentile are all higher than the black score at the 75th percentile. The Hispanic score at the 25th percentile is the same. In other words, 75 percent of blacks were admitted to MTU with math SAT scores equal to or lower than 75 percent of Hispanic, Asian, and white admittees.

Differences in Admittee ACT Scores

Figure 33
Admittee ACT Scores, Michigan Technical University
Figure 33 shows the range of composite ACT scores. Here, too, white, Asian, and Hispanic ACT scores are higher than black scores at the same percentiles. The gap in median ACT scores is six points between whites and blacks; seven points between Asians and blacks; and 4.5 between Hispanics and blacks. Black scores at the 75th percentile are the same as Hispanic and white scores at the 25th percentile (and slightly higher than those of Asians at the 25th percentile). In other words, 75 percent of blacks were admitted to MTU with ACT scores slightly lower than 75 percent of Asian admittees, and equal to or lower than 75 percent of whites and Hispanics.

Hispanic scores fall between scores of whites and blacks at the same percentiles. The Hispanic score at the 75th percentile is the same as the white median, while the Hispanic score at the 25th percentile is the same as the black score at the 75th percentile.

Differences in Admittee GPAs

Figure 34
Admittee GPAs, Michigan Technical University
Figure 34 shows ethnic group differences in GPAs. Gaps in GPAs are smaller than in test scores. The gap in median GPAs for blacks compared to Hispanics, Asians and whites is roughly a quarter of a grade-point. Black GPAs at the 75th percentile fall between the 50th and 75th percentile of the other groups.

Rejectees vs. Admittees

SAT verbal scores for white rejectees (n=19) are 490 at the 25th percentile, 510 at the 50th percentile, and 570 at the 75th percentile. Scores for black admittees (n=43) are 430, 470, and 520 at the 25th, 50th, and 75th percentiles. Scores for Hispanic admittees (n=11) are 520, 570, and 590 at the 25th, 50th, and 75th percentiles. Twenty-five percent of white rejectees had verbal SAT scores equal to the median score for Hispanic admittees, and higher than the verbal SAT scores of black admittees at the 75th percentile.

Math SAT scores for white rejectees (n=19) are 470, 530, and 590 at the 25th, 50th, and 75th percentiles. Scores for black admittees (n=43) are 450, 530, and 560 at the 25th, 50th, and 75th percentiles. Scores for Hispanic admittees (n=11) are 560, 600, and 630 at the 25th, 50th, and 75th percentiles. The median math score for white rejectees is the same as the median for blacks admitted to MTU. In other words, half the whites rejected by MTU had equal or higher math scores compared to the average black admitted to MTU. The median math score for white rejectees is much lower compared to the Hispanic median, but the score for white rejectees at the 75th percentile is close to median for Hispanic admittees.

ACT scores for white rejectees (n=119) are 19, 21, and 23 at the 25th, 50th, and 75th percentiles. Scores for black admittees (n=90) are 18, 20, and 23 at the 25th, 50th, and 75th percentiles. Scores for Hispanic admittees (n=32) are 23, 24.5, and 26 at the 25th, 50th, and 75th percentiles. The median ACT score of white rejectees is higher than the median score for black admittees, although it is lower compared to Hispanic admittees.

GPAs for white rejectees (n=130) are 2.39, 2.69, and 2.88 at the 25th, 50th, and 75th percentiles. GPAs for black admittees (n=32) are 2.92, 3.20, and 3.58 at the 25th, 50th, and 75th percentiles. GPAs for Hispanic admittees (n=13) are 3.03, 3.46, and 3.80 at the 25th, 50th, and 75th percentiles. The median GPA for white rejectees is more than half a grade-point lower compared to black and Hispanic admittees (0.51 and 0.77 respectively).

The GPA for white rejectees at the 75th percentile is lower than that for black and Hispanic admittees at the 25th percentile, meaning that 75 percent of white rejectees had lower grades compared to 75 percent of blacks and Hispanics admitted to MTU.

Six-Year Graduation Rates

• 44 percent of blacks
• 60 percent of Hispanics
• 60 percent of Asians
• 65 percent of whites

Asians and Hispanics at MTU graduate at somewhat lower rates compared to whites, while a significantly lower percentage of blacks graduate at MTU. (These figures are based on Fall 1995 averages of students entering MTU between 1983 and 1989.)

Northern Michigan University

There were 1,192 enrollees at Northern Michigan University in 1995.

Northern Michigan Enrollees:

• 0.9 percent are black
• 0.3 percent are Hispanic
• 0.5 percent are Asian
• 98.2 percent are white

Differences in Enrollee GPAs

Figure 35
Enrollees GPAs, Northern Michigan University
Figure 35 shows ethnic group differences in GPAs. The gaps are not large, although white scores at the 25th, 50th, and 75th percentiles are higher than scores for blacks, Asians, and Hispanics at the same percentiles. White median GPAs are only slightly higher that the median GPAs for Asians and Hispanics. The same is true for white versus Asian and Hispanic scores at the 25th and 75th percentiles.

The white-black gap in median GPA is almost half a grade-point (0.45). Moreover, the black median GPA is lower than the white GPA at the 25th percentile, meaning that the average black at NMU enrolls with a GPA lower than 75 percent of all white enrollees. The black median GPA is only slightly higher than the Asian and Hispanic GPAs at the 25th percentile.

Since the largest gap is less than a half-grade point, and data were not sent regarding ACTs or on rejectees, it appears that NMU does not engage in any systematic form of race preference, although more data might suggest a different conclusion.

Saginaw Valley State University

According to Barron’s 1995 College Guidebook, Saginaw Valley State University accepted 1,548 out of 1,689 applicants for enrollment in 1995, and 785 freshmen enrolled.

Saginaw Valley State Enrollees:

• 4.4 percent black
• 2.7 percent Hispanic
• 92.9 percent white

Differences in Enrollee ACTs

Figure 36
Enrollees ACT Scores, Saginaw Valley State University
Figure 36 shows ethnic group differences in ACTs. Although the ACT differences are not large, there is still a white-black gap in ACT scores, and a smaller one between whites and Hispanics. The white-black gap in median ACT scores is 2 points, with the black median ACT score the same as the white ACT score at the 25th percentile. This means that 75 percent of all whites enrolled at SVSU had equal or higher GPAs compared to the average black. The white median is the same as the black ACT score at the 75th percentile, meaning that 75 percent of blacks enter with the same or a worse ACT score compared to the average white.

Differences in Enrollee GPAs

Figure 37
Enrollees GPAs, Saginaw Valley State University
Figure 37 shows ethnic group differences in GPAs. The pattern is similar regarding grades, although Hispanic scores at the 75th percentile is higher than that of whites and blacks. White scores at the 25th and 50th percentiles are greater than those for Hispanics and blacks. The white-black gap in median GPAs is 0.45, which is almost a half grade-point difference. It is closer to the 25th percentile score for whites and Hispanics. Similar gaps occur between white-black and Hispanic-black scores at the 75th percentile.

University of Michigan at Ann Arbor

In 1995, 16,292 individuals applied for admissions and 74 percent (n=12,124) were admitted. Twenty-nine percent (4,759 admittees) subsequently enrolled. The overwhelming proportion of applicants were white.

University of Michigan Applicants:

• 8.5 percent black
• 3.9 percent Hispanic
• 16.9 percent Asian
• 70.8 percent white

University of Michigan Admittees:

• 9.3 percent black
• 4.7 percent Hispanic
• 16.8 percent Asian
• 69.2 percent white

University of Michigan Rates of Admission:

• 81.8 percent of blacks
• 90.7 percent of Hispanics
• 73.8 percent of Asians
• 72.8 percent of whites

University of Michigan Enrollees:

• 10.4 percent black
• 5.1 percent Hispanic
• 12.3 percent Asian
• 72.1 percent white

Differences in Verbal SAT Scores

Figure 38
Admittee Verbal SAT Scores, University of Michigan at Ann Arbor
Figure 38 shows the range of verbal SAT scores by ethnic group. White and Asian verbal SAT scores are higher at the 25th, 50th and 75th percentiles compared to blacks and Hispanics. The white-black gap in median SAT scores is 100 points. The white-Hispanic gap is 60 points. The Asian-black and Asian-Hispanic gap in median verbal scores is 110 points and 70 points respectively.

White and Asian verbal SAT scores at the 25th percentile are higher than the median verbal SAT scores for blacks and Hispanics. Seventy-five percent of whites and Asians admitted to UM had higher verbal SAT scores compared to the average black and Hispanic admittee.

Differences in Math SAT Scores

Figure 39
Admittee Math SAT Scores, University of Michigan at Ann Arbor
Figure 39 shows the range of math SAT scores by ethnic group. Like their verbal SAT scores, white and Asian math scores are higher at the 25th, 50th and 75th percentiles compared to blacks and Hispanics. The median white math SAT score is 130 points higher than the median black math score, and 70 points higher than the median Hispanic score. The Asian-black gap in median scores is 170 points, while the Asian-Hispanic gap is 110 points. Black scores at the 75th percentile are roughly the same as Asian and white scores at the 25th percentile. This means that approximately 75 percent of blacks admitted to UM have lower math SAT scores compared to roughly 75 percent of Asian and white admittees.

Differences in Composite ACT Scores

Figure 40
Admittee ACT Scores, University of Michigan at Ann Arbor
Figure 40 shows the Composite ACT scores for admittees. White and Asian scores are higher than those for black and Hispanics. The median ACT score for whites and Asians is six points higher than the black median, and four points higher than the Hispanic median. The ACT scores for whites and Asians at the 25th percentile are higher than black scores at the 75th percentile, meaning that 75 percent of blacks were admitted with lower scores compared to 75 percent of white and Asian admittees.

Differences in GPAs

Figure 41
Admittee GPAs, University of Michigan at Ann Arbor
Figure 41 shows the GPAs for admittees. While white and Asian GPAs at the 25th, 50th, and 75th percentile are higher than those for blacks and Hispanics, there is more overlap in GPAs compared to test scores. Nevertheless, the median GPAs for blacks and Hispanics is lower than white and Asian GPAs at the 25th percentile, meaning that at least half the blacks and Hispanics admitted had lower GPAs than 75 percent of Asian and white admittees.

Rejectees vs. Admittees

Verbal SAT scores for Asian rejectees (n=648) range from 420 at the 25th percentile, to 540 at the 75th percentile, with a median verbal score of 490. White rejectees’ verbal SAT scores (n=2,167) range from 460 at the 25th percentile, to 560 at the 75th percentile, with a median score of 510. Verbal scores for black admittees (n=686) range from 430 at the 25th percentile, to 550 at the 75th percentile, with a median score of 480. Hispanic admittee scores (n=455) range from 460 at the 25th percentile, to 590 at the 75th percentile, with a median score of 520. While somewhat lower compared to Hispanic admittees, verbal scores for white rejectees at the 25th, 50th, and 75th percentile are all higher than comparable scores for black admittees. Verbal scores for Asian rejectees are roughly the same as those for black admittees. This percentile may be more meaningful as an absolute number. There were 1,378 whites and 327 Asians rejected despite having higher verbal scores than the median black admittee.

The math SAT scores of Asian rejectees ranged from 580 at the 25th percentile, to 680 at the 75th percentile, with a median math score of 640. White rejectees scored from 540 at the 25th percentile, to 650 at the 75th percentile, with a median math score of 600. Black admittees scored from 470 at the 25th percentile, to 622 at the 75th percentile, with a median math score of 540. Hispanic admittees scored from 530 at the 25th percentile, to 670 at the 75th percentile, with a median math score of 600. Scores for Asian and white rejectees at the 25th, 50th, and 75th percentile are higher than scores for black admittees at the same percentiles. Asian and white rejectees’ scores at the 25th percentile are equal or higher than the median for black admittees, meaning that 75 percent of Asians and whites rejected by UM had higher scores than 50 percent of blacks who were admitted.There were 1,571 rejected whites and 546 Asians who had higher scores than the median black admittee.

Combining the verbal and math section of the SAT reveals that there were 297 Asian and 1,090 white rejectees who had both verbal and math SAT scores higher than the black admittee medians.

Composite ACT scores of Asian rejectees (n=292) ranged from 22 at the 25th percentile, to 26 at the 75th percentile, with a median score of 25. White rejectees (n=1,924) scored from 23 at the 25th percentile, to 27 at the 75th percentile, with a median score of 25. Black admittees (n=849) scored from 20 at the 25th percentile, to 25 at the 75th percentile, with a median score of 23. Hispanic admittees (n=289) scored from 22 at the 25th percentile, to 28 at the 75th percentile, with a median score of 25. Median ACT scores for Asian and white rejectees are identical to the median ACT score for Hispanic admittees and that of black admittees at the 75th percentile. In other words, half the Asian and white rejectees had higher ACT scores than half the Hispanics and 75 percent of blacks admitted to UM. In raw numbers this translates into 180 Asians and 1,334 whites with higher ACT scores than the median black admittee.

GPAs for Asian rejectees (n=718) range from 2.90 at the 25th percentile to 3.40 at the 75th percentile, with a median GPA of 3.10. GPAs for white rejectees (n=3,125) range from 2.90 at the 25th percentile to 3.40 at the 75th percentile, with a median GPA of 3.20. GPAs for black admittees (n=1,094) range from 3.00 at the 25th percentile to 3.60 at the 75th percentile, with a median GPA of 3.30. GPAs for Hispanic admittees (n=560) range from 3.10 at the 25th percentile to 3.78 at the 75th percentile, with a median GPA of 3.40. GPAs for Asian and white rejectees at the 25th, 50th, and 75th percentile are lower than comparable scores for black and Hispanic admittees, although their GPAs at the 75th percentile fall at or above the median GPAs for black and Hispanic admittees.

There are 182 Asians and 1,052 whites, a total of 1,234 individuals, rejected from UM with GPAs higher than that of the average black and Hispanic admittee.

It is of interest to combine grades with either SAT scores or ACT scores to make the strongest case for the existence of Affirmative Action operating to the disadvantage of white and Asian applicants. There were 49 Asians and 564 whites with ACT scores and grades above the black admittee medians. There were 42 Asians and 224 whites with SAT scores and grades above the black admittee medians. In the first case there were 613 individuals and in the second case 266 individuals who were rejected who could be said to be more qualified than the median black admittee.

In various individual instances, these differences in qualifications were astoundingly large. Forty-nine of these individuals had ACTs greater than 29, 77 had combined SATs greater than 1200, and, most amazing of all, 4 had SATs greater than 1400.

In sum, therefore, using ACTs and grades together there were 613 white or Asian individuals who were rejected even though they were above the black admittee median and using the SATs and grades there were 266 white or Asian individuals who were rejected even though they were above the black admittee median. This is a very large number of individuals who in all likelihood suffered from reverse discrimination by not being admitted despite their superior qualifications to the average black enrollee.

Six-Year Graduation Rates

• 66 percent of blacks (n=1,159)
• 76 percent of Hispanics (n=663)
• 86 percent of Asians (n=1,716)
• 87 percent of whites (n=14,599)

Whites and Asians have the highest graduation rate, followed by Hispanics, then blacks. On average, about 20 percent more whites and Asians graduate compared to blacks; 10 percent more whites and Asians graduate compared to Hispanics. (These figures are an average of the freshmen classes entering UM between 1987 and 1990.)

University of Michigan at Dearborn

In 1995, 1,399 persons applied to UMD, 86 percent (n=1,200) were admitted, and 41 percent (n=580) enrolled.

University of Michigan at Dearborn Applicants:

• 11.9 percent black
• 3.1 percent Hispanic
• 8.5 percent Asian
• 76.4 percent white

University of Michigan at Dearborn Admittees:

• 10.6 percent black
• 3.0 percent Hispanic
• 9.1 percent Asian
• 77.3 percent white

University of Michigan at Dearborn Rates of Admission:

• 76.0 of percent black
• 81.8 of percent Hispanic
• 91.6 of percent Asian
• 86.8 of percent white

Overall admissions rate: 85.8 percent

University of Michigan at Dearborn Enrollees:

• 9.0 percent black
• 2.6 percent Hispanic
• 6.6 percent Asian
• 81.9 percent white

Differences in Admittee Verbal SAT Scores

Figure 42
Admittee Verbal SAT Scores, University of Michigan at Dearborn
Figure 42 shows the range of scores by ethnic groups. (Hispanic scores are not reported because there were only five Hispanic admittee verbal and math SAT scores.) Whites have the highest verbal scores at all three reported percentiles. The white-black gap in median verbal SAT scores is 90 points, with the median Asian score falling in between. The black median is lower than white scores at the 25th percentile. That is, the average verbal score of blacks admitted to UMD is lower than scores of 75 percent of whites admitted to UMD. At the 75th percentile for black admittees, black scores are lower than the white median, meaning that 75 percent of blacks admitted to UMD have lower verbal SAT scores compared to the average white admittee.

Asian scores at the 25th, 50th, and 75th percentile are lower than those of whites at the same level. The white-Asian gap in median scores is 50 points. The Asian median is only slightly higher than white scores at the 25th percentile. This shows that the average Asian admittee to UMD has lower scores compared to 75 percent of all white admittees.

Differences in Admittee Math SAT Scores

Figure 43
Admittee Math SAT Scores, University of Michigan at Dearborn
Figure 43 shows the range of math scores by ethnic groups. Unlike verbal scores, math scores of Asian admittees are higher than those of whites at the same percentiles. The Asian-white gap in median scores is 30 points, with white scores being lower.

Black scores are significantly lower compared to Asian and white scores at the same percentiles. The Asian-black gap in median scores is 170 points; the white-black gap is 140 points. The black score at the 75th percentile is 43 points lower than the Asian score at the 25th percentile, and 10 points lower than those of white admittees at the 25th percentile. 75 percent of blacks were admitted to UMD with scores lower than 75 percent of Asian and white admittees.

Differences in Admittee ACT Scores

Figure 44
Admittee ACT Scores, University of Michigan at Dearborn
Figure 44 shows the range of composite ACT scores. Asian and white scores at the 25th, 50th, and 75th percentiles are higher than those of Hispanics and blacks. Hispanic ACT scores at these percentiles are higher than black scores at the same percentiles.

The white-black gap in median ACT scores is five points. The black ACT score at the 75th percentile is lower than the white score at the 25th percentile, meaning that 75 percent of blacks were admitted to UMD with lower ACT scores compared to 75 percent of whites admitted to UMD.

There is greater overlap between white and Hispanic scores compared to white and black scores. The gap in median scores between Hispanics and whites is smaller compared to the white-black gap. The difference in median scores between white and Hispanic admittees is three points, and the Hispanic score at the 75th percentile is the same as the white median.

Differences in Admittee GPAs

Figure 45
Admittee GPAs, University of Michigan at Dearborn
Figure 45 shows ethnic group differences in GPAs. The differences are minimal. The largest difference is 0.20--less than a quarter of a grade point. Hispanic and white median GPAs are the same, while black and Asian medians are slightly lower.

Rejectees vs. Admittees

Verbal SAT scores for white rejectees (n=7) are 340, 360, and 400 at the 25th, 50th, and 75th percentile. Scores for black admittees (n=18) are 365, 400, and 478 at the 25th, 50th, and 75th percentile. Most white rejectee scores fell below the median score for black admittees, but the scores of those at the 75th percentile are the same as the black median. There are two whites with scores equal to the black median. Math SAT scores for white rejectees (n=7) are 340, 370, and 430 at the 25th, 50th, and 75th percentile. Scores for black admittees (n=18) are 398, 430, and 500 at the 25th, 50th, and 75th percentile. The differences are the same as in the verbal scores. Most white rejectee scores fell below the median score for black admittees, but the scores of those at the 75th percentile are the same as the black median. There are two whites with scores equal to the black median. Because the absolute number of cases is low, we should hesitate to draw too many conclusions.

Composite ACT scores for white rejectees (n=128) are 17.25, 19, and 21 at the 25th, 50th, and 75th percentiles. Scores for black admittees (n=118) are 17, 19, and 21 at the 25th, 50th, and 75th percentiles. Scores for Hispanic admittees (n=31) are 19, 21, and 24 at the 25th, 50th, and 75th percentiles. The ACT scores at the 50th and 75th percentiles for white rejectees is the same as those for black admittees. This translates into 64 whites rejected by UMD despite having the same or higher ACT scores compared to the average black admittee.

GPAs for Asian rejectees (n=10) are 2.28, 2.55, and 2.75 at the 25th, 50th, and 75th percentiles. GPAs for white rejectees (n=136) are 2.20, 2.40, and 2.60 at the 25th, 50th, and 75th percentiles. GPAs for black admittees (n=122) are 2.90, 3.20, and 3.60 at the 25th, 50th, and 75th percentiles. GPAs for Hispanic admittees (n=34) are 3.03, 3.40, and 3.60 at the 25th, 50th, and 75th percentiles. GPAs of white and Asian rejectees are significantly lower at all percentiles compared to black and Hispanic admittees. The gap in median GPAs between white rejectees and black admittees is almost a full grade-point. White and Asian rejectees GPAs at the 75th percentile are lower than black and Hispanic GPAs at the 25th percentile.

Six-Year Graduation Rates

• 22 percent of blacks
• 40 percent of Hispanics
• 48 percent of Asians
• 47 percent of whites

Six years after entry, fewer than one of every four black enrollees graduate from UMD. This is less than half the graduation rate for whites and Asians. Four in ten Hispanics at UMD graduate after six years, which is lower than the graduation rates for Asians and whites, but considerably higher than blacks. (These figures are based on the Fall 1995 averages of students entering UMD between 1983 and 1989.)

We would like to thank the Michigan Association of Scholars for helping obtain the data used in this report.