Study: Relying on Above-Average Scores on Entrance Exams Hinders DiversityJanuary 5, 2015 |
While above-average scores on college entrance exams may predict better academic performance and graduation rates in college, relying on the scores in admission decisions will “severely reduce” the number of minority and low-income students on campus.
That is the key finding from a new study released Monday, titled “Efficacy vs. Equity: What Happens When States Tinker with College Admissions in a Race-Blind Era?”
The study is meant to illuminate the challenges associated ensuring diversity on campus by means of supposedly “objective” criteria such as SAT or ACT scores.
Those challenges come into play in the era ushered in by the 2014 Fisher v. University of Texas case, which holds that public universities must show there are no “workable solutions” not based on race before they resort to race-based admissions strategies to bring about diversity.
To demonstrate the difficulty of using criteria that is supposedly race-blind, the researchers — UT Austin professors Sandra Black and Jane Arnold Lincove and Texas A&M University professor Kalena Cortes — show what the impact would be if college entrance exam scores and other criteria, such as the number of college prep courses taken and high school exit exam scores, were applied in Texas, where the top 10 percent of students in all high schools are automatically admitted to the public university of their choice.
They found that SAT or ACT scores are “significant predictors” for the top 10 percent of high school graduates and can boost the average GPA by 6.3 percent and increase the four-year graduation rates by 5.9 points.
“However, adding these new admission criteria also severely reduces both minority and low-income representation in our analysis,” the study states.
Specifically, it says, using the college entrance exam scores would eliminate automatic admissions eligibility for 69 percent of Hispanics, 73 percent of Blacks and 62 percent of low-income student who were admitted based on class rank alone.
“Thus the design of automatic admissions policies reflects a classic tension between equity and efficiency,” the study states of the Texas 10 percent plan, which was meant to ensure diversity on campus by guaranteeing admission from the top 10 percent of students in all public high schools, which tend to be heavily segregated. “While average student performance might be improved with additional criteria, each additional criterion might also limit college access among underrepresented students.”
College entrance exam representatives took exception with the study’s finding.
ACT Vice President of Research Wayne Camara conceded that “from a policy perspective, yes, eliminating tests and just using grades will improve diversity.”
“But it will result in a lower academic success and lower graduation rates and lower persistence rates,” Camara said, noting that universities may not be willing to sacrifice higher GPA and higher graduation rates to admit more minority or low-income students.
Camara also said the study wrongly downplays the difference in GPAs predicted by college entrance exams.
The difference between a 2.9 and a 3.1 is “huge,” he said, because while a GPA is based on a scale of 0 to 4 points, college students cannot survive with anything below a 2.0.
“So that six percent in the study is not reflecting the true magnitude of the difference because, yes, you can get a GPA of 0.1 or 1.1 but you’re not going to be a college student next semester,” Camara said. So the 6.3 percent difference in GPA mentioned in the study is “really a difference of 12 or 15 percent in reality,” he said.
Zach Goldberg, director of external communications at The College Board, the company that administers the SAT, did not directly address the Texas study but noted how the SAT and other assessments have “cast a light on the inequities in our education system.”
He cited a different study — “The Test-Optional Movement at America’s Selective Liberal Arts Colleges: A Boon for Equity or Something Else?” — that found test-optional colleges were not more diverse.
For the most recent study out of Texas, the researchers looked at all students who graduated from Texas public high schools in 2008 and 2009.
The analytic sample included students who graduated in the top 10 percent of their high school classes during these two years, enrolled in a Texas selective four-year public university directly after high school and attempted a full-time course load in their first fall semester, the study states.
This includes approximately 22,000 students from a total of roughly 500,000 graduates, the study states.
Although the top 10 percent tended to be higher-income students and had less racial and ethnic diversity, the sample was still more diverse than other groups that were the focus of college readiness studies, the researchers claim.
Specifically, they said the sample was 26 percent Hispanic, 6 percent Black, 11 percent Asian and 27 percent from families with income below $40,000.
The average SAT/ACT score for the sample group is was 1170 versus 1029 for all high school graduates.
While requiring above average SAT/ACT would eliminate only 8 percent of White students, 10 percent of Asian students and 7 percent of high-income students from eligibility, doing so would eliminate 40 percent of Hispanics, 49 percent of Blacks and 36 percent of low-income students, the study states.
The researchers also found that college entrance exams “are not equally predictive of college outcomes for Black and White students.”
“The potential racial bias of SAT/ACTs is well documented and suggests that average scores for minorities are lower due to factors that are not associated with college success,” the study states. “It is likely that white students in Texas and beyond have better access to test preparation, which, by design, weakens the association between ability and performance by teaching students how to improve their scores with no meaningful gains in actual college readiness.
“Thus, the use of these criteria in race-blind admissions might inadvertently introduce inequity,” the study states.
Camara, of ACT, disagreed that the ACT is “biased.”
“I think that’s unfair. That’s wrong. That’s incorrect,” Camara said. “What they really mean is the scores on the ACT or any cognitive test for that matter have shown lower performance for minorities, or low-income students doing well on them.”
The Texas researchers maintain that “applying the same SAT/ACT score criteria to a Black and a White student may be inappropriate given these differential effects.”
But since applying differential criteria for different racial and ethnic groups is now illegal for state universities, the researchers says percent plans such as the 10 percent plan in Texas are a “more attractive solution.”
“Our results suggest that the fewer ‘objective’ criteria that are used in admissions, the less inequity will be introduced,” the study states. “In the case of Texas, efficiency gains from adding criteria come at a very high cost of dramatic reductions in equity.”
The study concludes with this advisory: “Admissions officers should use caution in applying minimum standards across the board when diversity continues to be a goal of admissions.”