Affirmative action – A question of merit

Once again, the US Supreme Court is reconsidering nearly four decades of legal precedent – essentially, allowing public universities to use race and ethnicity as one among many factors when considering student admissions.

As in previous affirmative actions cases, a Euro-American student filed a lawsuit against a highly selective public university, in this case the University of Texas at Austin (UT-Austin). The plaintiff, Abigail Noel Fisher, claims overt racial discrimination when UT rejected her freshman application in 2008. Her lawyers filed the case that same year, and it wound its way through lower courts.

The Supreme Court decided to review the case and began deliberations last October. This court is decidedly more conservative than in the past, and seemingly more sympathetic to simply ending affirmative action.

The details of the case are not very different from earlier affirmative action cases, but the court’s decision to revisit precedent is driven, in large part, by the conservative majority among the nine justices.

Abigail Fisher’s lawyers claim that she achieved an academic record that should have allowed her to be accepted and then to enrol at the Austin campus, and that others with less achieved merit had been allowed to enter in her place based on racial preferences.

But this raises a number of important questions: what are the proper criteria to determine merit? Is admission to a selective institution (in which the demand far outstrips the supply of enrolment spots) a reward determined, as in many national higher education systems, via a single national test? Or should it be a set of criteria that helps predict academic performance and levels of engagement once the student is enrolled?

Just as importantly, what is the appropriate locale for determining what constitutes merit?

There is a long history of US courts deferring to the academic community when it comes to making decisions not only related to admissions, but also hiring faculty and setting curricula. The Supreme Court’s pending decision will revisit this precedent.

Here I attempt to focus on the issue of merit and its complexity, and why universities in the US and elsewhere need to be the primary location for shaping their student bodies – within broader goals sets by state or national governments.

That issue revolves around what admissions criteria result in a student body that is diverse in talent and fully engaged in meeting the academic and civic expectations of the university.

Whereas the defenders of affirmative action in the US have invested much of their arguments in the notion of a critical mass, and the educational benefits of a diverse student body, the opposition has placed their bets on a rather simplistic notion of merit: for example, if X students have certain test scores and grades, they must be superior in merit to Y students with lower test scores and grades.

Notions of student engagement

A path for shaping more effective admissions policies in the US and other parts of the globe – and more generally, in creating a student body suitable for the mission and ideals of the public university – can be found by investigating more fully how admissions standards and practices correlate with actual student academic performance.

The difficulty of disentangling the specific effects of the college experience and engagement from personal influences and the background of students has long been recognised. The variables are many and familiar: a person’s early experience, family relationships and friends, social origins and advantages.

“The effects of higher education,” wrote the sociologist Martin Trow, “may be very subtle and difficult to measure: effects on mind, character, sensibility, competence, horizons and ambitions – effects, that is to say, on the whole range of moral, emotional and intellectual skills and qualities that a person takes with him into his adult life.”

Yet it is possible to decipher some of the influences on what can be termed academic and civic engagement of students – to measure how their prior achievements and social backgrounds influence their learning and experience once they began their undergraduate careers.

A study of undergraduates at the University of California, one of the most highly selective public university systems in the US, using data from the Student Experience in the Research University (SERU) Consortium and Survey, attempted to explore this link and generated some intriguing findings.

First, it found that the socio-economic and ethnic background of students was much more diverse than previously thought. Some 55% from a representative sample stated that they had at least one parent who was an immigrant; on the two most selective campuses within the system, Berkeley and Los Angeles, the figure was approximately 65%. About 25% learned another language before learning English, and another 25% grew up speaking English in addition to another language.

Although students reported a median parental income of $72,000, about 25% of the student body was drawn from families whose incomes were below $35,000. About 25% of the students were the first generation of their families to enter college, and about 30% identified their families as ‘working class’ or ‘low income’.

Most important for this discussion, students from relatively disadvantaged backgrounds were likely to rank higher on indicators of academic engagement (for example, time studying, class attendance, interaction with faculty, completion of assignments) than students from more affluent backgrounds.

They also had similar and sometimes better time-to-degree rates than the more affluent. All of this was true despite the fact that the less privileged cohort was more likely to be working more hours to finance their education. They also typically had more family-related responsibilities than their more affluent counterparts.

Test scores and merit?

How did standardised test scores predict academic success and levels of engagement?

Critiques of America’s higher education institutions suggest that students at selective and other institutions are ‘academically adrift’, marginally engaged in their studies and learning little in their academic careers.

But we know that there is a great variety of academic and other forms of engagement valued by, for example, major research universities, and many influences on the behaviours and learning outcomes of students, including students’ immersion in their discipline or major.

The SERU data and other studies indicate that students with high test scores are often less engaged than the students with scores near the median and below; they do not take advantage of their academic opportunities once past the gauntlet of a standardised test-focused admissions process, and often have narrow skills – valuable, but narrow.

Are they predisposed to being ‘academically adrift’? From this and other analyses on the validity of test scores, one might surmise that a place like UC Berkeley should avoid admitting students with very high test scores. But that is not the point here; rather, the lesson is that student talent and potential comes in many forms.

Then there is the relative value of test scores versus the grade point average (GPA) in determining merit. Within the University of California’s multi-campus system, studies going back to the 1950s repeatedly showed that GPA is the best and most consistent predictor of academic success – better than test scores.

Further, UC has viewed course requirements as a primary tool for influencing the curriculum of secondary schools, and the behaviours and preparation of prospective students for a university education.

But another important finding is that both GPA and test scores, in the end, have historically only marginally explained the variance in, for example, GPA performance once at university.

In the aftermath of the affirmative action battles in California during the 1990s, one university study indicated that high school GPA explained only 14.5% of the variance in the GPA of freshmen at UC; the SAT explained 12.8%; and subject-based tests were only a little bit better, but still not as good as grades in high school.

A study in the early 1960s conducted by UC’s academic senate showed similar results and was, in fact, one reason that UC did not require the SAT until 1968 largely as a diagnostic tool, and did not incorporate the SAT and ACT into actual admissions decisions until 1979.

UC was one of the last major universities to require standardised tests or use them in admissions precisely because of their lack of predictive power or what is termed their ‘validity’.

Seeking nuance

These observations emphasise the importance of a more nuanced understanding of merit. What are public universities attempting to achieve in their admissions practices?

In part, the goal should be to distribute access (when choosing among students) to those who are not only talented but who will be the most engaged, who will have the most to gain or contribute to the academic milieu.

In choosing among many, one threshold is who will have a reasonable chance of success at the university, a principle incorporated by the University of California almost from its founding, although it came later for many other state universities.

One might also assume that universities, public and private, must also take some chances and provide exceptions to the general rules of admissions.

The goal is to seek admissions policies that can take into account the great variety of human talents and abilities, the desire for learning, and the potential contribution of a student to the larger academic community, which includes the ideals of representation (but not as an edict), and ultimately balances individual merit with the larger needs of society.

Hence, the court, and the public, should consider the complexity of merit and reconsider the value of test scores and even GPA in high school as insufficiently narrow measures for predicting collegiate success. Test scores alone, it can be argued, do not equate to merit worthy of trumping all or any other factor in an admissions decision.

They also need to see that the best place for such judgment on the difficult question of access to a highly sought public good is at the institutional level, not in the courts.

* Dr John Aubrey Douglass is senior research fellow in public policy and higher education at the Center for Studies in Higher Education (CSHE) at the University of California – Berkeley. This article is a version of a paper published online as part of the CSHE’s Research and Occasional Paper Series, “Affirmative Action, the Fisher Case, and the Supreme Court” and is based in part on themes and portion of the narrative in “Perils and Opportunities: Autonomy, merit, and privatization”, in John Aubrey Douglass, The Conditions for Admission: Access, equity and the social contract of public universities. Stanford University Press, 2007. See here.