Some caution is in order in interpreting the recent Quinnipiac University survey results on the Supreme Court's recent decision restricting the use of race in assigning students to public schools. As Orin notes in his post, Quinnipiac asked respondents whether they approved of the Supreme Court's recent decision "that public schools may not consider an individual's race when deciding which students are assigned to specific schools." A massive 71% majority said that they "agree" with the Court's decision.

However, the true level of public opposition to affirmative action preferences in education is likely much lower than this. Public opinion scholars have known for years that most survey respondents will express hostility to anything described as a "racial preference" or as racial discrimination. This is particularly true if the question at issue - like Quinnipiac's - fails to distinguish between affirmative action and traditional racial discrimination against minorities. Many of the Quinnipiac respondents probably assumed that the Supreme Court forbade old-style racial discrimination against minorities.

By contrast, a recent ABC/Washington Post poll that described the decision as "restrict[ing] how local school boards can use race to assign children to schools" and noted that "Some argue this is a significant setback for efforts to diversify public schools, others say race should not be used in school assignments." It found that 56% "disapproved" of the decision, while only 40% said they approved. A July Newsweek poll (hat tip: VC commenter Tim Dowling) described the decision as "limit[ing] the use of race for school integration plans," and found that 32% of respondents supported the decision, while 36% opposed it (a statistical tie).

The ABC/Washington Post and Newsweek questions have their own flaws (e.g. - "restricting how local school boards can use race" is very vague, and "integration plans" introduces a term - "integration" - with positive associations). But the contrast between their results and Quinnipiac's is nonetheless striking.

More generally, strong majorities favor programs described as "affirmative action." For example, this 2003 Pew survey found that 57% of Americans support "affirmative action programs that give special preferences to qualified blacks, women, and other minorities, in hiring and education," while only 35% oppose them. Note that a large majority - at least in this survey - supported "affirmative action" even in a question that defined AA as giving "special preferences" to women and minorities (albeit perhaps only to "qualified" ones). The same survey found that 60% said that "affirmative action programs designed to increase the number of black and minority students on college campuses" are a "good thing," while only 30% said that they were "bad." Similarly, this 2005 USA Today poll found that 49% of American support "affirmative action programs for racial minorities," with only 43% opposed. As a general rule, the majority of the public will express support for a program defined as "affirmative action" for minorities or women, but will oppose anything described as a "preference" or as "discrimination."

While political elites and others in the know use terms such as "racial preference" and "affirmative action" interchangably and have clear, stable views on the issue, much of general public has far less clear opinions and fails to understand the connections between them. Some of this is due to a genuine desire many people have to support "affirmative action" while at the same time rejecting racial "preferences." Some is probably due to widespread rational political ignorance, which results in many people not understanding the implications of common political terms. Whatever the cause, we must be very cautious in interpreting polls on affirmative action and racial preferences. Small differences in wording can have a big impact on results.

UPDATE: While it is not directly relevant to the subject of this post, it's also worth noting that the Quinnipiac question described the Court's decision incorrectly. The Court emphatically did not hold that the government "may not consider an individual's race when deciding which students are assigned to specific schools." Instead, Justice Anthony Kennedy's controlling opinion in the close 5-4 decision clearly indicated that some uses of race are in fact permissible, just not the very flagrant ones at issue in these particular cases. Kennedy emphasized his disagreement with Chief Justice Roberts' "all-too-unyielding insistence that race cannot be a factor in instances when, in my view, it may be taken into account."

Several commenters on my last post make the perfectly valid point that strong public support for referendum questions banning racial "preferences" in the generally liberal states of California, Washington, and Michigan is an indicator of public opposition to affirmative action programs. True enough. However, all three of these referenda were worded as banning "preferences." In a 1997 Houston referendum, opponents succeeded in getting the city government to reword an otherwise similar ballot question as banning "affirmative action" rather than "preferences" (see here for an account critical of the City's decision). Sure enough, the initiative was defeated by a 55-45 majority, even in relatively conservative Houston. In referenda as in polls, whether the public supports affirmative action policies depends on how the question is worded.

James Taranto of the Wall Street Journal's OpinionJournal.com website has written an interesting response to my posts on affirmative action and public opinion polling (see here and here).

Taranto acknowledges that whether or not survey respondents support affirmative action policies depends on how the questions are worded. Strong majorities will say they support "affirmative action," but reject "discrimination" and "racial preferences." This pattern occurs not just in polls, but also on in referenda on anti-affirmative action ballot questions, with propositions banning "preferences" winning in California, Michigan, and Washington, and one banning "affirmative action" getting defeated in Houston.

However, Taranto claims that the "affirmative action" wording is misleading because "changing 'racial preferences' to 'affirmative action' is a change of meaning, not just wording. 'Affirmative action' is not only a euphemism for discrimination in favor of minorities; it is also a blanket term that encompasses other, less controversial policies." For example, Taranto quotes Justice Kennedy's concurring opinion in the Seattle school case that set off this debate, as an illustration of "non-preference" affirmative action policies:

School boards may pursue the goal of bringing together students of diverse backgrounds and races through other means [than racial preferences in admissions], including strategic site selection of new schools; drawing attendance zones with general recognition of the demographics of neighborhoods; allocating resources for special programs; recruiting students and faculty in a targeted fashion; and tracking enrollments, performance, and other statistics by race.

Taranto's argument is not without some merit. But there are two major problems with it. First, as I pointed out in my initial post on this issue, using the term "preferences" is also misleading because it will lead many voters to think of traditional invidious discrimination against minorities rather than of remedial affirmative action preferences.

Second, most of the supposedly nonpreferential programs listed by Taranto and Justice Kennedy are in fact themselves racial preferences. If, for example, school boards "pursue the goal of bringing together students of diverse . . . races" by trying to increase the percentage of minorities at certain schools, they are still intentionally taking race into account in their decisionmaking and still trying to ensure that the schools in question have some particular percentage of minority students. Similarly, "strategic site selection" of schools based on the racial composition of the surrounding neighborhood is surely a form of race-based decisionmaking that advantages areas with one type of racial composition at the expense of others with a different type. The fact that officials are not using formal racial classifications but are instead using "facially neutral" means does not change the essential nature of what they are doing. Racial balancing cloaked in seemingly neutral language is still racial balancing. As I explained in this post on Texas' "ten percent" college admissions plan:

[I]f it is morally wrong to aim for a given racial balance in a state university student body by using explicit racial preferences, why is it not equally wrong to intentionally try to achieve the same effect through indirect, facially "neutral" means? In the days of Jim Crow, southern states often used facially neutral policies such as literacy tests, poll taxes, and peonage laws to disadvantage blacks. Few today would argue that these policies were somehow morally superior to those Jim Crow laws that discriminated against blacks through explicit racial classifications. If, as critics of affirmative action claim, explicit affirmative action preferences are morally wrong for the same reason that Jim Crow laws were wrong, then "facially neutral" affirmative action systems such as the Texas ten percent are wrong for the same reasons that the facially neutral means of propping up Jim Crow were.

I do not contend that the majority of voters are clearly in favor of affirmative action. Rather, my view is that many of them don't know much about the issue, have not thought about it carefully, and therefore have unclear and often internally contradictory views. This is not in any way unusual, and is a natural consequence of voters' rational ignorance about public policy, which extends to a wide range of issues including some that are far more important than affirmative action. For that reason, I also question Taranto's claim that the voters in the California, Michigan, and Washington referenda reached a clear decision indicative of their true opinions because "they had the opportunity to hear both sides of the argument." While it is certainly true that they had the "opportunity" to hear both sides (something that they also had during the previous 25 years of intense public debate over affirmative action), it is unlikely that very many actually used that opportunity to study the issue in any detail.

UPDATE: Taranto has a rejoinder to this post here. I would first like to apologize to James Taranto for incorrectly referring to him as "Jim." I have corrected this in the original post. However, I remain unpersuaded by his substantive arguments.

In his new post, Taranto claims that initiatives that ban "discrimination" or "preferences" do not mislead voters by leading them to think of traditional discrimination against minorities because they "are intended to apply, to both forms of discrimination." It is true that they are intended to do that, but an unsophisticated voter reading the text is more likely to think of traditional discrimination when seeing the word "preference" or "discrimination" than of affirmative action.

Taranto also notes that his argument "is not that the rewritten language was deceptive but that it changed the meaning of the proposed law by expanding it to encompass nondiscriminatory forms of 'affirmative action.'" I don't think that this distinction undermines my point that questions banning "discrimination" or preferences will lead voters to think of traditional discrimination against minorities. Moreover, as I explain in my post above, the "nondiscriminatory" forms of affirmative action Taranto refers to are in fact preferences themselves.