Public Opinion and the Wording of Referendum Questions Banning Racial "Preferences" and "Affirmative Action" :

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.

Henri LeCompte (mail):
Which is why so much effort goes into finding and promulgating such euphemisms as "pro-life," or "pro-choice," or "affirmative action," or "racial preferences."

Nobody said voters were rational.
8.16.2007 9:25pm
Ilya Somin:
Nobody said voters were rational.

Well, I have. The problem is that it's rational for voters to be poorly informed.
8.16.2007 9:42pm
guest (mail):

Have you accounted for the fact that within each state, where the proposition did not vary, the polling always showed far more support for AA than was shown behind closed doors? References to Houston can't answer that. You'd need to account for the intra-state gaps between polling and voting, right? And that gap is sizable, right?
8.16.2007 9:53pm
Ilya Somin:
Have you accounted for the fact that within each state, where the proposition did not vary, the polling always showed far more support for AA than was shown behind closed doors? References to Houston can't answer that.

I don't need to account for it, because it doesn't affect my argument that the amount of support for AA varies greatly based on question wording. The Houston example shows that that is true for referenda as well as for poll questions.
8.16.2007 9:58pm
DRJ (mail):
I agree with your premise that words matter but so does timing. This poll was in 1997 when AA was more popular nationwide and in Texas. In addition, AA is better understood today than it was 10 years ago, when it was common to believe that voting against AA was a clear vote for segregation and racism. But maybe that was your point and, if so, I agree.
8.16.2007 10:32pm
Why not be honest and just ask if you are for or aganist AA. Simple. To think that after all this time citizens do not know what AA is all about is a little silly. They also know when their chain is being jerked. Citizens are not as stupid as some think. The general public is far ahead of the Academics. It is like thinking the public does not understand diversity is code word for AA. Or that the public does not understand PC as speech control. Come on we are not living in a cave anymore.
8.16.2007 11:04pm
guest (mail):

However, the true level of public opposition to affirmative action preferences in education is likely much lower than this.

While I agree with the notion that "wording matters," I still think that your original post failed to account for the true level of public opposition to AA as expressed in the ballot box. I also think that the 1997 Houston situation, while interesting, is no trump card on that issue. The repeated phenomenon of a gap between polling and voting needs to be accounted for when we wish to gauge the true level of public opposition to AA.
8.16.2007 11:05pm
Houston Lawyer:
The Houston City Council's rewording of the proposition was about as low as you can go in politics. I have no doubt that I could re-write a referendum that 60% of the public supports in such a way that a majority will vote against it. Try getting approval for authority to hire lesser qualified people based upon the race of the person applying. Local governments fairly openly do this all the time and are allowed to get away with it.
8.16.2007 11:54pm
A Texan:
Call me crazy, but I'd consider the fact that Houston was an urban center with a 25% black population as more likely the decisive factor than the fact that the ballot called the proposal anti-affirmative action. To compare, the 2000 Census demographic data:

Houston: 49.27% white, 25.31% black.
California: 79.75% white, 7.65% black.
Michigan: 83.05% white, 14.92% black.
Washington: 88.64% white, 4.12% black.
8.17.2007 12:21am
Andy Freeman (mail):
Perhaps we should poll on specific practices instead of labels.

For example, should we reserve slots for folks with a given skin color/ethnic heritage? What kind of preference should we give folks with said characteristics?

Would the affirmative action advocates be happy with that kind of question? How about the opponents?
8.17.2007 12:27am
DRJ (mail):
Here's another example of how words can skew the debate.
8.17.2007 3:43am
From Michigan's actual title on the ballot, in bold large print:

8.17.2007 10:18am
chris c:
I believe aff action and racial preferences aren't nec the same (though in practice they usually are.)

for ex, one component of AA is outreach - such as advertising in media popular with blacks or Hispanics. I doubt most people have a problem with this. it would survive referenda like the Mich one quoted in the comments.

it's in the interest of supporters of racial preferences to elide the difference b/w practices like that, which are popular (or at least not unpopular) and don't discriminate against anyone, and preferences, which are widely hated and discriminate on their face.
8.17.2007 10:37am
karrde (mail) (www):
Thanks for the reminder about the use of the words "Affirmative Action" in the Michigan ballot proposal.

Question about the Michigan result:

Blacks number about 14% of the state's residents, but are a majority in the city of Detroit and in certain suburbs.

Has anyone measured whether the outlawing of Affirmative Action was popular or not among voters in those areas?

To my recollection, most of the advertising was mostly of the "Yes on Proposal N" or "No on Proposal N" variety. (There were 7 State Proposals on the ballot in that election, I think the anti-AA Proposal was Proposal 2.) How much was the phrase "Affirmative Action" in the public debate about the proposal?
8.17.2007 11:54am
Libertarian1 (mail):
I don't need to account for it, because it doesn't affect my argument that the amount of support for AA varies greatly based on question wording. The Houston example shows that that is true for referenda as well as for poll questions.

As a non-attorney much of the discussion here genuinely disturbs me. The argument is made let's try and change the actual wording of the referenda or poll to get the desired results. If we just use AA we will get one result and if we say racial preferences we will get another.

Hyoothetical- you come into my office as your physician and I prescribe a brand new medication for you. You ask what are the side effects? In this carefully composed scenerio you are as uninformed of the facts as the average voter is of AA or racial preferences.

What is being suggested here is I should parse the words of my answer to you to get my desired result- you take the meds. But we all insist in medicine on "informed consent". Why not in referenda or even polls? Is it easy to find the correct language? Absolutely not but we must try if the voter can make a ratiomal informed choice.

It seems to me in the case we are discussing both sides believe just saying AA obfuscates the question. For some that is the goal. Maybe just saying racial preferences is not clear enough. But we have some great legal minds here and I would bet if I locked you in a room for a few hours a sentence could be agreed on which would actually tell us what our citizens want.
8.17.2007 12:23pm
just me:
I like how a proposition that essentially clones the Title VII language can be read to cut off preferences, as everyone "knows" Connerly's evil intent, while the original Title VII language itself was open to the preference game, because Lyndon Johnson said so.

I wish, during the GOP Congress era, someone had proposed repealing Title VII and re-enacting it with . . . the identical language. Would floor debate showing that their "intent" somehow mean that they could effect a major change in the law by re-enacting the same darn text?

Too funny.
8.17.2007 12:37pm
My definition of 'affirmative action' and 'preferences' are different. Affirmative action means reaching out to make sure minorities know that a job, for example, is open by advertising directed to minorities, etc. Preferences, on the other hand, mean that when job candidates come along, the minority is hired over the non-minority simply because he is minority.

The former is fine. The latter is racism.
8.17.2007 12:51pm
markm (mail):
Note that Michigan's ballot proposal was to "...BAN AFFIRMATIVE ACTION PROGRAMS THAT GIVE PREFERENTIAL TREATMENT..." It doesn't ban advertising directed at minorities or sending more recruiters to majority-black schools. I don't think there'd be any dispute about such actions, but most AA is racial preferences. Which is racism.
8.17.2007 3:11pm
Here are the county numbers from counties with a large number of blacks (in total votes/yes/no format--"yes" means make AA illegal). Wayne County has by far the most blacks.

WAYNE 622,261 254,545 367,716
SAGINAW 78,596 42,923 35,673
OAKLAND 497,839 284,554 213,285
GENESEE 160,587 93,024 67,563
INGHAM 102,774 49,539 53,235
JACKSON 55,372 34,269 21,103
8.17.2007 3:49pm
Affirmative Action = Affirmative Racism.

Is there any other way to look at it?
8.17.2007 9:44pm
lyarbrou (mail):
There is now a lawsuit involving the wording for a Missouri Affirmative action initiative petition. The group circulating the petition suggested the following language for the ballot:

"Shall the Missouri Constitution be amended to prohibit any form of discrimination as an act of the state by declaring:

"The state shall not discriminate against, or grant preferential treatment to, any individual or group on the basis of race, sex, color, ethnicity or national origin in the operation of public employment, public education or public contracting?"

The office of the Democratic Secretary of State (Carnahan) used this language for the ballot:

"Shall the Missouri Constitution be amended to:

"Ban affirmative action programs designed to eliminate discrimination against, and improve opportunities for, women and minorities in public contracting, employment and education; and

"Allow preferential treatment based on race, sex, color, ethnicity, or national origin to meet federal program funds eligibility standards as well as preferential treatment for bona fide qualifications based on sex?"

The statements would seem to convey very different impressions and appear to be an attempt to "load" the dice for the voting.
8.18.2007 12:19am