Finding 5: Message Clarity Favored Yes on 8

How many times the average voter saw Yes on 8 ads and No on 8 ads


Yes on 8 Ads

No on 8 Ads










Los Angeles






Palm Springs



San Diego






San Francisco



Santa Barbara



The number of advertisement views can potentially over- or understate message penetration. Yes on 8 focused most of its ads around one clear message, increasing its potential for message penetration or even discourse domination. No on 8’s apparent parity in number of ad views in some markets was undercut by the multiplicity of messages in its ads.
Yes on 8’s ads were clear, direct, and repetitive. As a result, their message penetrated. No on 8’s ads were vague, inconsistent, and de-gayed, reducing their power to persuade.

Above and beyond the power of the specific arguments, the message discipline of Yes on 8 maximized the memorability and impact of its communications. By contrast, the more scattered and less disciplined messaging of No on 8 meant that it forfeited some of its potential impact.

·         Yes On 8’s message was consistently clear. It focused on the untrue idea that kids were in danger.

·         No on 8’s message was neither consistent nor clear.

·         Very late in the campaign, in the final nine days, No on 8’s ads offered a clear, consistent discrimination message. But for swing voters, the message wasn’t memorable or credible.

·         No on 8 struggled with message discipline for two reasons. The first was the late change in decision makers. The second was the tension between the desire to describe LGBT people honestly and the impulse to de-gay the campaign. The result was message tentativeness and gay avoidance in the later No on 8 ads.

·         De-gaying tested better in No on 8’s polls than it worked in reality. It was a dangerous choice because it allowed Yes on 8 to define what LGBT people are like. But the issue is not simple.


Comparing Yes on 8 and No on 8’s Message Clarity

Yes on 8’s Message was Consistently Clear

Yes on 8 focused not only “Princes” but all of its ads to make three clear points:

·         You as a voter and parent are losing control over your kids (every ad);

·         A pro-gay change is being imposed “whether you like it or not” (driven home hardest and most explicitly in “Newsom,” and reiterated in “Princes,” “Massachusetts,” “Field Trip,” and “Closer”);

·         The pro-gay change poses a real, immediate danger to children, including your children (the centerpiece of “Princes,” “Massachusetts,” and “Field Trip,” and part of “Closer”).

Yes on 8 knew what it wanted to say and it said it, again and again. The Yes on 8 ads were easy to grasp, and they constantly repeated a few fundamental points.

The term used to describe this combination of clarity and repetition is message discipline. The message discipline of Yes on 8 meant that the average voter in California heard the same basic message twenty to forty times in paid TV advertising alone. (Whether a voter heard it twenty or forty times or something in between depended upon where the voter lived, their demographic makeup, and their television viewing habits.) The rule of thumb upon which media consultants rely is that voters need to be exposed to an ad a minimum of five times, and perhaps as many as ten times, before it penetrates and has the potential to be memorable. Yes on 8 exceeded that threshold easily.

The highly repetitive TV ads represented the bulk of the communications sent by the Yes on 8 campaign and received by most voters. The message discipline in these communications almost certainly increased memorability and impact of the ads.

No on 8’s Message was Neither Consistent nor Clear

In contrast, the No on 8 ads that made it on the air in large enough TV buys so that voters saw them were not clear, consistent, or tightly focused. Instead, the widely seen No on 8 ads—“Thorons,” “Unfair,” “O’Connell,” “Internment,” and “Obama”—made a variety of much more loosely related points. Most of these ads made multiple points, a doubtful practice for 30-second spots because of the sacrifice of clarity. Some ads made points supported by the less frequently aired “Feinstein” and “No for Latinos” ads.

·         The Equality Argument: We should allow gay people to marry because we should treat them the same as everyone else (“Thorons”).

·         The Kitchen Sink: Well-known, respected people and groups, including Obama, Schwarzenegger, and Feinstein, say Prop 8 is bad for many reasons. The key intended takeaway is that the sheer number of problems raised by endorsers and negative adjectives featured in the ads should give voters a distaste for Prop 8 (“Unfair,” “No for Latinos,” and “Obama,” and to a lesser degree, “Feinstein”).

·         The Rights Argument: It’s wrong to take away rights from any group of people (“Unfair”).

·         The “Don’t Treat People Differently” Argument: It’s wrong to treat people differently under the law (“Unfair”).

·         Rebuttal to the Kids and Schools Argument: Prop 8 will not lead to teaching about gay marriage in schools (front and center in “O’Connell,” mentioned in passing in “No for Latinos” and “Obama”).

·         The Discrimination Argument: Discrimination is wrong (“No for Latinos”); also expressed as “No matter how you feel about marriage, discrimination against any group of people is wrong (“Feinstein,” “Internment,” “Obama”).

To many LGBT people and our allies, all of the above arguments may feel very interconnected and self-reinforcing. To us, being for equality and against discrimination feel like the same thing, or so close to the same thing that they might as well be the same thing.

But for some voters who are not already strongly on our side, these arguments are separate ideas, not identical, and not equally appealing. For example, there are a group of voters who don’t want to “treat people differently” when it comes to marriage—but they don’t see denial of marriage as amounting to discrimination. One way we know this is that among the least-firmly decided voters, the cluster of arguments made by the No on 8 ads do not rate identically when they are tested in polls. Some voters agree with one argument but not with another.

From the point of view of persuadable voters—those who could have gone either way, and those who changed their mind over the course of the campaign—the No on 8 ads made a multiplicity of points. In some key media markets, no voter heard any one of these messages as frequently and clearly as they heard the Yes on 8 message. See the table below for the details. In some markets, however, including Los Angeles and San Francisco, No on 8 did achieve repetition parity with Yes on 8. The failure to break through in those cases probably reflects two factors: (1) timing: No on 8 achieved most of its repetition in the closing days, when between a third and half of all voters had already voted absentee; and (2) weakness of the discrimination message, discussed in more detail below.

In addition, among the individual No on 8 arguments, some were clear but others were not. For example, the “Thorons” ad was clear, and polling found it memorable to a larger group of voters than any other No on 8 ad (more on this in Finding 3 above). But the equality message of “Thorons” was not repeated in subsequent ads.

Similarly, the non-campaign-approved ads that aired before the election had no measureable impact on voters.  For a discussion of the two ads, the widely broadcast “Garden Wedding” and the much less broadcast Courage Campaign advertising, see Appendix F, Non-campaign Media.

An In-Depth Look at No on 8’s Problems with Message Clarity

How many times the average voter saw Yes’s Kids Argument and No’s Discrimination Argument


Yes’s Kids Argument

No’s Discrimination Argument










Los Angeles






Palm Springs



San Diego






San Francisco



Santa Barbara



Kids Argument = “Princes” + “Massachusetts” + “Field Trip”

Discrimination Argument = “Feinstein”+ “Internment” + “No for Latinos” + “Obama”

No on 8’s reliance on Endorsers was Not an Adequate Substitute for Message Consistency and Clarity.

Starting with “Unfair,” No on 8’s TV ads cited several of its many endorsers as part of the campaign’s attempt to persuade. Different ads mentioned different endorsers. Sometimes an ad explained why a particular endorser had given their endorsement to No on 8; others were listed with no specific reason for their position. The consultant team assembled by the second campaign manager thought these endorsers would serve as validators. The hope was that some voters would respond either to a particular endorsement that they trusted or be impressed by the totality of endorsements for No on 8.

In campaigns in general, this reliance on endorsements has its greatest value in four specific circumstances:

·         When a ballot measure is only vaguely understood;

·         When the other side doesn’t offer clear reasons for its position;

·         When the other side fails to run a strong campaign; and

·         When an endorser is so widely seen as expert and trusted that voters will think it makes sense to substitute the expert’s judgment for their own.

Unfortunately, none of those were the case in Prop 8. The measure was widely publicized. From the point of view of most voters, the topic of same-sex marriage is easily understood. The Yes on 8 campaign offered clear reasons for voting Yes and ran a $40 million campaign with widely broadcasted ads. From the Yes on 8 ads voters knew that not every expert or endorser was on the same side. No one person in American society was viewed as so expert on marriage or LGBT people that their endorsement by itself would have the power to persuade.

It’s less clear whether one person in America was widely enough respected that his or her endorsement would have moved a significant slice of voters. Possibly then-Presidential-candidate Obama enjoyed that kind of popularity and credibility on this issue among African-Americans or other groups of voters; the data is insufficient to judge. But No on 8 hardly used the Obama endorsement prior to November 1, and starting November 1 used it only in one ad in a very brief way, with one still photo that was quickly replaced by a similar shot of Gov. Schwarzenegger. To the extent that No on 8 was attempting to use endorsements strategically, then it bypassed the most obvious possibility.

There is one additional reason to doubt the persuasive power of endorsements in an election like Prop 8. Polling and focus group research has over a long period of time come to suggest that voters view themselves as experts on marriage. They know what marriage is. When they watch ads like “Unfair” or “Feinstein,” they may be skeptical that the Teachers Union or Sen. Feinstein has more expertise on the matter of marriage than they themselves do. The polling does not tell us whether one or both of these reasons (or other reasons) dominated the thinking of voters as they viewed these ads, but the Lake polling shows that neither ad moved voters toward No on 8.

Late in the Campaign, No on 8’s Ads Offered a Clear, Consistent Discrimination Message, but for Swing Voters, the Message Wasn’t Memorable or Credible for Seven Reasons.

The No on 8 campaign chose one message—the discrimination message—for most of its ads in the final week of the campaign. Discrimination became the dominant message when “Feinstein” and “Internment” were on the air from October 30 through November 4. For seven reasons, however, the message wasn’t effective.

The first problem was late introduction of the message. No on 8 had largely avoided the word “discrimination” prior to this very late phase of the campaign. The discrimination message was neither reinforced by nor did it reinforce what had come before—the first reason why it was unlikely to be memorable for voters.

The second problem was lack of clarity. No on 8’s discrimination message was encumbered with a major distraction: there was no mention of the group being discriminated against.

The lack of transparency was related to a third problem: the expression of the message may have unintentionally undercut itself. Both the “Feinstein” and “Obama” ads pose a rather ominous idea: “No matter how you feel about marriage, vote against discrimination, and vote No on 8.” Some voters might have wondered why they had to ignore their feelings about marriage to take a stand against discrimination. The statement at some level suggests that if voters noticed their feelings they might choose differently.

If the third problem, above, was that voters might have difficulty decoding the message; the fourth problem was that voters might succeed too well at decoding it. The message is meant to keep “gay” in the background; yet when voters receive an ambiguous communication, they put the ad in context. At this late point in the campaign, days away from Election Day, voters very likely realized that “how you feel about marriage” really meant “how you feel about gay marriage” or “how you feel about gay people.” Each voter was being asked to disregard how they feel about “gay”—the missing word they themselves had to supply to make sense of the ad—to take a stand against discrimination. The takeaway for some voters might well have been that gay people are so unpopular that even those who support them can’t speak of them. Gay people are the people who must not be named. The ads offer no assurance that gay is OK, or that same-sex marriage is tolerable. The possible result: voters got a message that was the opposite of what was intended: “gay” and the unpopularity of “gay” was now front-and-center in their thinking.

Stimulating ambivalence may have been the intention of the ad creators and the new No on 8 decision makers; yet quite possibly this is a fifth problem. Ambivalence is the emotional opposite of the equality argument made in No on 8’s first ad, “Thorons.” The decision makers must have decided to target voters unsure of whether they could tolerate LGBT equality, or were ambivalent about gay people. The ad, by validating these feelings, was seemingly intended to connect the voters to the rest of the sentiment—that in spite of how they feel about gay people, their even stronger feelings against discrimination should lead them to vote No on 8. But it’s not obvious why heightened ambivalence makes the totality of the No on 8 message appealing.

To the extent that the campaign decision makers at this point meant to stimulate both ambivalence and also positive feelings about LGBT people, they were unwilling to pull the trigger. There is no explicit, positive statement about LGBT people, nor any mention of the word “gay.” The only hint of this intention was the fleeting, unlabeled images of LGBT people in the closing frames of “Internment.” Even LGBT viewers would have been hard pressed to recognize that these quick shots depicted gay people. The impulse to include these images was admirable but not meaningful. They were too brief and unclear to have an impact.

The problem of lack of clarity recurs in No on 8’s other major antidiscrimination ad, “Internment.” Watching this, voters had to decide what was meant by the phrase “discrimination against any group of people is wrong.” By itself, the idea seems simple and appealing. But the idea was not being offered by itself. Instead, “Internment” offered this language to explain Prop 8 and the need to vote No on 8 at the same time that the ad showed provocative images of shameful past discrimination against Japanese-Americans and Armenians. The apparent hope of the ad creators was that viewers of the ad would vote no due to their strong, generalized antipathy to discrimination.

Yet here we encounter the sixth problem: a lack of relevance or credibility. The voters viewing the ad had already received at least dozens—and more likely hundreds—of communications about Prop 8. Most had very likely gotten the message across that this vote involves gay people, and lesbian and gay couples getting married. Every Yes on 8 ad, every news story, every conversation over a water cooler at work told voters that Prop 8 is about “gay.” So voters watching “Internment” didn’t watch “Internment” as though it’s all they knew. Instead, they put two and two together and realized that Prop 8 was not a vote on racial or ethnic discrimination; it was a vote on “the gays.” And as part of voting on the gays, they’re also considering, after having seen the Yes on 8 kids message ten to twenty times, how they feel about their kids learning about gay people, perhaps even how they feel about their kids growing up to be gay. For voters experiencing this thought process, “Internment” could not conceivably have been a persuasive ad, and the tracking polling shows it wasn’t persuasive. More likely it struck swing voters as irrelevant or misleading.

There is one final, seventh problem with the discrimination argument, particularly as raised in “Internment.” In its polling on a variety of same-sex marriage ballot measures, Lake Research had found that many voters define “discrimination” as treating someone wrongly, not simply treating them differently. At this point in history, almost everyone agrees that internment of Japanese-Americans during World War II amounted to treating them wrongly. It was both different treatment and wrong treatment.

But the inability of same-sex couples to marry is not nearly as universally seen as wrong. Many who voted Yes on 8 may have viewed it as different treatment of LGBT people but not wrong treatment of them. To the extent that that’s true, the marriage issue does not strike them as “discrimination,” and the analogies to other types of discrimination seem unconvincing. This may well be why both “Internment” ads—the one created by No on 8 in 2008 and the virtually identical one created ten years earlier, in 1998, by the pro-LGBT No campaign in Hawaii’s ballot measure on marriage—had no apparent persuasive impact. To view both ads, go to the Media section.

It is always possible that some as yet untried or unknown version of the discrimination argument could overcome these seven problems. An effective form of the discrimination argument might exist, and we simply do not yet know how to make it. Alternatively, the discrimination argument may work better in the future than it does now if more voters come to see unequal access to marriage as a form of discrimination. At the present time, however, “discrimination” appears to be an argument that appeals to many who already support same-sex marriage but not to those whose who have yet to vote with us. At least so far, when we make this argument in the hopes that it will persuade more people to vote with us, we are only talking to ourselves.

The equality argument may suffer from some of the same defects as the discrimination argument in the eyes of swing voters.

The discrimination argument may not be the only one that is both intuitively appealing to those who already support same-sex marriage and yet irrelevant or worse when we seek to persuade more voters to stand with us. There may be a similar problem with the equality argument that was favored earlier in the No on 8 campaign. Third Way points out in its report “Moving the Middle on Marriage: Lessons from Maine and Washington” that the equality argument has not yet worked in these ballot measure campaigns and that most voters don’t agree with the underlying rationale. Only 32% of Maine voters in the moveable middle agreed with the argument that “Separate is not equal; everyone, including gay and lesbian couples, should be treated equally under the law, including laws governing marriage.” This is better than the 22% who believe that “denying gay and lesbian couples the right to marry is discrimination,” but it may be too low to help us compete for most of the moveable voters.

Particularly in a competitive election, message discipline matters.

Yes on 8 did better than No on 8 at having a consistent message that it hammered home, ad after ad. No on 8 might have been better served if it had taken any one of its messages and stuck with it, but we can’t know that from the data because that isn’t the approach that was tried.

The approach taken by No on 8 is a better strategic match in a campaign that is itself a mismatch, eg a campaign where a proposition on the ballot is mostly ignored or where the opposing side doesn’t run a vigorous campaign or where the issue is one where feelings don’t run so high. But the lack of message discipline and the problems with message clarity are most costly in the kind of situation epitomized by Prop 8: where a controversial measure, on which public opinion is closely divided, is hotly contested by two equally well-funded campaigns.

Two factors contributed to No on 8’s difficulties with message clarity and message discipline.

·         The late change in decision makers led to a change in media consultants and in thinking about the message. Ads planned and aired by the first regime were very different from those created and aired by the second.

·         Polling tended to support de-gaying the campaign. Clear arguments about LGBT people and use of the word “gay” tested less well than abstract arguments and vagueness. Although the No on 8 executive committee resisted the pressure and insisted on using the word “gay” when they were operating as decision makers, the tension between the two impulses resulted in message tentativeness and gay-avoidance apparent in the body of the later No on 8 ads.

Most campaigns view it as a given that message discipline—clarity and repetition—is essential for success. Yet No on 8 sacrificed clarity and some of the power of repetition in an attempt to avoid facing the centrality of “gay” to the question on the ballot. Let’s examine why the campaign made this trade-off.

Message continuity suffered from the late change in regime.

On Sept. 29, No on 8 changed its campaign decision-making process and its campaign decision makers. Given the late date and the scale of the change, the quick transition was remarkably smooth. Externally, news coverage of the new campaign manager in the Advocate was only positive. Internally, the new manager brought order to the challenging situation: he recruited a new group of paid and pro bono consultants to advise and assist him; created a campaign “boiler room” in San Francisco, where his new team met face-to-face; and he focused on fundraising, which had been slumping since early August. Everyone with whom I met who was involved in the No on 8 leadership acknowledges that these improvements stabilized and focused the campaign.

One downside, however, was that No on 8 was not prepared for Yes on 8’s aggressive campaign. The “Newsom” ad went on the air the same day the new manager officially took over. He was recruiting his new team even as the first tracking polls were showing the extent of the toll the “Newsom” ad was taking. Fundraising required immediate attention, but so did a wide variety of matters.

Another downside was that the newly assembled team had limited time to assimilate the information and assess the choices made by their predecessor consultants, and to consider the political environment now altered by Yes on 8. None of the new ads continued or built on the message of “Thorons” and “Conversation,” the two No on 8 ads that were on the air from September 22 through October 6. In departing from the earlier choices, the new team necessarily ran the risk that the “Thorons” message would either be muddied or lost, replaced by a new message. For a detailed discussion of the timline of the campaign, see Appendix C.

The No on 8 Campaign’s De-Gayed Message

The polling supported a message minimizing “Gay”.

It is not clear whether the new No on 8 team was familiar with the polling done before they arrived on the scene. But if they had been, it would have supported their approach. Arguments sidestepping “gay” polled the best in the Lake benchmark poll of May 19 through 27. Of fourteen potential No on 8 message arguments that asked voters whether and how strongly they agreed, just two of the five (#1 and #4) used the word “gay”, and they used it only to dismiss its importance and relevance to the campaign decision voters would face (one begins “regardless of how I feel about gay marriage”; the other begins “I may not agree with gay marriage”). The other three made no mention of “gay” at all.

None of the messages that tested best in the polling dealt directly with the issue on the ballot. The first said “people should not be treated unfairly under the laws,” but it does not make the case that “gay people should not be treated unfairly.” The third says “we should support committed couples” but does not signal that it could include committed gay and lesbian couples. The fifth says “we should not single out one group for discrimination” but gives no clue what the one group might be.

In the same poll, Lake also tested fifteen additional statements, asking which were convincing and to what degree. These were more realistically related to voters’ likely understanding of the issue by Election Day.

This battery found that relatively direct arguments using the word “gay” tested about as well as vague statements. But the overall levels of convincingness of any of the arguments in the second battery—with percentage of agreement in the mid-50s—was less encouraging than the first battery, where voter agreement with the de-gayed statements ranged from the mid-60s to the high 70s.

Follow-up testing in July provided more support for messages omitting clear mention of “gay.” Then, Lake Research found that among undecided voters, the most convincing argument by over 12% made no mention of gay people.

For all of the polling information on message testing, see Appendix J.

The Danger: De-gayed arguments test better in a poll than they work in the real world.

From the May and July message testing in the polling, one argument tested the best. It began, “We do not need more government in our lives.”

Yet on the face of it, this was a risky argument for the pro-LGBT side. The potential risk was that the voters we needed to persuade, even if they nodded along with it beforehand, would not find this argument credible once the campaign was joined and Yes on 8 was putting forth its arguments.

To see the danger readers, put yourselves in the voters’ shoes at two key moments. Moment one: when one of our targeted voters was on the phone, answering the pollster’s questions in May or July, or participating in a focus group in July or August. Moment two: much later, when a voter (not the same person) began to really think about how they were going to vote. The latter moment is when our targeted voters—the undecided, the persuadable, the most wavering on either side, the most ambivalent, and the least informed voters, for better or worse, these are likely to be among the voters who will make the difference between our winning and losing—started to pay attention to the campaign, and began to absorb (sometimes even to seek out) information from the larger political environment and from the Yes and No campaigns.

At the earlier moment, when voters were on the phone answering a pollster’s questions—especially in an early benchmark poll—they had much less external information coming their way about the ballot measure and about the issue of same-sex marriage. The public parts of the campaign, including TV ads, hadn’t begun. The voters hadn’t heard from both sides. They hadn’t gossiped with their friends, chatted with their coworkers, overheard their neighbors talking about it in the supermarket checkout line. The continual stream of miscellaneous information in which we all swim hadn’t yet bathed the voters with information on the topic.

Our targeted voters may had thought little about Prop 8 and may not have yet formed an opinion about how they would vote. About the issue of same-sex marriage, perhaps they had thought little or much; either way, they very likely had an opinion.

At this early point, when talking on the phone with the pollster, few of these voters were going to disagree with a bromide like:

“We do not need more government in our lives. The government has no role telling two committed and devoted adults who they should marry. Government has no business telling people who can and cannot get married just like it cannot tell us what we can read or say or do in our private lives. We do not need a constitutional amendment that gives the government more say in our lives.

At the moment of the poll, it got 72% agreement. (The amazing thing is that 28% didn’t agree.)

Now consider the later moment. As the targeted voters started to consider how they would vote, more information had come their way. Most knew that the issue in some way had something to do with gay people, and they may have known that it related to marriage for gay and lesbian couples.

Here’s a specific example. The average targeted voter in LA who waited until Election Day to vote saw the Yes on 8 ads thirty times. Of those thirty, twenty exposures repeated the same message. Of the thirty, twenty were the Yes ads “Princes,” “Massachusetts,” and “Field Trip,” which were all about kids in danger of learning about gay people and same-sex marriage in school.

The same average voter also saw the No on 8 ads forty times. But they had very little repeated exposure to one consistent No on 8 message. Instead, the average LA voter saw:

·         “Thorons” five times (parents of a lesbian daughter);

·         “O’Connell” seven times (rebutting “Princes”);

·         “Internment” seven times (one of several No on 8 ads to make some version of a discrimination argument, though the only one to make the argument by analogy; it compared discrimination against LGBT people explicitly to discrimination against Japanese-Americans interned during World War II and against Armenian-Americans);

·         A mix of other No on 8 ads twenty-one times.

In the final week there was greater consistency, though nothing like the discipline exercised by Yes on 8. In the closing week, the average voter saw No on 8 ads fifteen times, and those fifteen included all of the views of “Internment.” Even in the final week in LA, however, No on 8 broadcast eight different ads.

(All of the specific numbers of ad exposures used in the above paragraphs come from data on the actual Yes on 8 and No on 8 ad buys in the Los Angeles media market. See Appendix E for the backup data and more details.)

In TV ads alone, Yes on 8 exposed the average LA voter twenty times to explicit anti-LGBT propaganda, to the false idea of pro-gay indoctrination of children. At this moment, after exposure to the Yes on 8 ads and other information, it was no longer likely that 72% of the voters would find the poll-tested argument (“government has no business telling people who can and cannot, ” etc.) compelling as they decided how to vote on Prop 8.

Instead, the poll-approved message was probably an irrelevancy, not part of a meaningful political dialogue. The Yes side had taught the voter that “gay” is the issue at hand. The Yes side had defined what “gay” is and, implicitly, what kind of people gay people were. Even though “Internment” is a beautiful piece of film art, under these circumstances its message could easily have felt like a non sequitur to a voter who had been exposed to all of the above.

In other words, voters’ agreement with platitudes in a poll does not predict whether they will find those same ideas persuasive in a competitive campaign. The appeal of veiled references and euphemisms may poll better than depictions of actual LGBT people and use of the word “gay”, and yet have zero persuasive power once voters learn that LGBT people are in fact the ones they are voting on and that “gay” has everything to do with the issue on which they’re voting. The success of the de-gayed “persuasion” argument in a poll is therefore most probably an artifact of the polling process itself.

No on 8, particularly in the late stages of the campaign, chose a de-gayed message for its TV ads.

A reasonable inference to draw from watching all of the No on 8 ads (except for the first ad, “Thorons”) is that the ad creators made a decision to avoid the word “gay.” The principal exception was the much more direct “Thorons” ad as aired (though not in its first version; as initially screened by the No on 8 executive committee, “Thorons” did not include the word “gay” and was inscrutable; the EC pushed to have the word added). The only other use of “gay” in a No on 8 ad is Superintendent O’Connell’s assurance that Prop 8 would not lead to teaching gay marriage in schools, ie Prop 8 was not in this respect pro-gay. To the extent that the “O’Connell” spot uses the word “gay,” “gay” is not portrayed as something positive or neutral; it is assumed to be negative. Go to. See Appendix E to see all the No on 8 ads and for the full timetable of when the various No on 8 ads aired and with what size buy.

By contrast, No on 8’s mail to voters made more favorable use of the term “gay” and also provided much greater clarity that Prop 8 was about LGBT people.  Mail was, however, a tiny part of No on 8’s voter communications.  The mail is therefore worth examining primarily because it takes a different approach that No on 8 could have pursued, and that strikes me as more promising as well as better executed.  See Appendix N for the mail pieces developed for the campaign.

The decision whether to depict and describe LGBT people clearly in ads or to de-gay is not a simple one.

As much as I would like to say for certain that de-gaying is bad, and “gaying” will get us a better election result, as a community we currently have too little experience to know. It is possible that a persuasion argument that accurately depicts or describes LGBT people would do better than a de-gayed argument; it is possible it would not. I do not possess data that definitively provides us an answer.

The reason for our lack of knowledge is that only a few pro-LGBT campaigns have tried depiction or description of LGBT people in TV ads and measured whether and how it moves us forward in the heat of a campaign. No on 1 in Maine (2009) and No on 36 in Oregon (2004) are the best-funded campaigns to date to create and air ads depicting and intending to define LGBT people. Pro-LGBT campaigns in Utah and Texas went even further but faced very uphill odds and broadcast their ads only in marginal and submarginal buys respectively.

Evidence is beginning to develop suggesting that more direct and non-de-gayed arguments exist and poll promisingly. They are at the moment largely untried. Third Way, for instance, recommends messaging that shows that gay couples “will honor and respect the tradition of marriage”; want to make “a lifetime commitment”; and want to “join” the institution of marriage, not change it. Third Way specifically concluded that based on its data, “Talking to a gay person helps to convince the [moveable] middle that gay couples want to get married for the same reasons straight couples do.” But given the blind spots in message testing polling on same-sex marriage, it is important to recognize that encouraging poll findings alone do not make the case that the arguments will work.

Progress begins, however, with our recognition that polling tends to overstate the power of de-gayed arguments.

What’s clear is that we run the risk of misleading ourselves when we focus our polling on de-gayed arguments. Implicit in each of these arguments is an unwarranted assumption that we so fully control voters’ view of the ballot measure that we can shield them from realizing the centrality of LGBT people in it. Each de-gayed argument insufficiently simulates the reality established when we face an organized campaign like Yes on 8. Lake Research erred when it relied so heavily on testing de-gayed arguments in its message battery. See Appendix I for additional discussion of this topic.

When the pro-LGBT campaign is de-gayed, it opens up an opportunity for the anti-gay side to define what gay people are like. It’s dangerous to give the anti-LGBT side that opportunity. The damage may (1) in the short run, jeopardize the ability of the pro-LGBT side to win the immediate election, and (2) in the long run, perpetuate or add longevity to anti-gay stigma. The pro-LGBT side should investigate much more fully both of these possibilities. If either is true, the LGBT community should urgently seek alternatives to the de-gayed campaign messages.

See Appendix I for a more detailed discussion.

The key point to add is this: whether or not we feel comfortable with a de-gayed campaign, whether or not a de-gayed campaign polls is our best alternative, and whether or not a de-gayed campaign gives us our best chance to win, the strategy to de-gay runs a serious risk.

The risk is that the reputation of LGBT people as decent, trustworthy, ethical, and honorable people will be diminished or damaged—that the campaign in its totality, including the communications by both sides, will lead a significant number of voters to see us as less than human, and therefore justify our treatment as less than human. Our failure to answer defamation of our character may be very costly. It would be smart to consider and investigate this possibility. Then, if it turns out that damage is done whenever we de-gay, we need to find healthier ways to advance our legislative agenda and reduce anti-gay prejudice.