Kansas Demonstrates the Difference Between Polls and Votes

How to interpret the surprising results of the most important abortion vote since Dobbs

One of the most difficult, but necessary, elements of opinion writing is providing an objective analysis of issues you care deeply about. There’s a constant pressure (both internal and external) to “get on the team,” to deny the other side any “ammunition” in the never-ending war of ideas.

There’s perhaps no debate more impacted by this pressure than the American abortion debate. The stakes are so high—and emotions run so hot—that it’s hard to drill down and answer the key question: What’s really going on?

Well, I’m going to try to do that today. I’m writing about the Kansas abortion referendum, where voters in a solidly red state resoundingly rejected a constitutional amendment that would have empowered the legislature to tightly regulate (and perhaps ban) abortion in the state. Instead, voters preserved the status quo, which protects the right to abortion up to 22 weeks of pregnancy.

The outcome doesn’t surprise me. The margin and turnout do. To understand why, we’ve got to go deep into public opinion on abortion. Buckle up. This is going to get complicated.

Let’s deal first with the outcome. Abortion polling is often a mess. For example, pay no attention to polls asking whether voters support or oppose overturning Roe v. Wade. Voters aren’t lawyers, and they don’t know what Roe truly held—apart from broadly upholding the right to an abortion.

That’s how you can get incompatible polling results. Most voters oppose the Dobbs v. Jackson Women’s Health Organization decision and support upholding Roe, but in the same poll most voters support abortion bans after 15 weeks. Yet 15-week bans were incompatible with Roe. The best abortion polling doesn’t ask about Roe alone, nor does it ask generalized or vague questions about abortion. The best abortion polling gets very specific, and when polling gets specific you learn a surprising reality about the abortion debate in the United States: Both the activist pro-life and activist pro-choice positions are minority positions.

Let’s take the Harvard CAPS / Harris Poll I linked above. My pro-life position, which would prohibit abortion except to preserve the life or physical health of the mother or except in cases of rape or incest, is held by only 37 percent of Americans. That’s unpopular. A supermajority of Americans disagree. That means I have to do real work to persuade my fellow citizens of my position.

But you know what’s less popular than my position? The most extreme pro-choice positions. Only 10 percent of Americans support abortion rights up to nine months. Only 28 percent of Americans would permit abortion up to 23 weeks. The most popular polled position is a 15-week ban, which garners a collective 72 percent support.

I single out the Harvard CAPS / Harris Poll because its precision captures the true state of the abortion debate better than most polls I’ve seen. The activist pro-life position is a minority position, but it’s still bigger than the activist pro-choice position, and that has granted the pro-life position an electoral advantage.

The 2021 Virginia gubernatorial race is a prime example of this phenomenon. The Democratic candidate, Terry McAuliffe, poured money into ads attacking his opponent, Glenn Youngkin, for his position on abortion. The ads were so prominent that The New York Times, NPR, and The Washington Post all highlighted the issue as a key test of voter sentiment.

McAuliffe’s ads fizzled. After he lost, exit polling indicated that only 8 percent of voters listed abortion as the “most important issue” facing Virginia. And of that 8 percent, almost 60 percent were pro-life. An 8 percent issue is the very definition of an activist issue, and there were more activist pro-life voters than pro-choice.

Here’s the truly important question—did Dobbs reverse this dynamic? In other words, has the end of Roe increased the number of voters who prioritize abortion rights? Issue polling is ambiguous. The Kansas results are not, but questions remain.

Let’s stick with the Harvard CAPS / Harris Poll. Even though it demonstrates that the pro-life position is a minority position, that same poll indicates that abortion is much less prominent in voters’ minds than multiple other issues. For example, inflation is the most important concern, followed by the economy, guns, and immigration. When voters are asked to list their greatest and second-greatest concerns, inflation tops the list at 62 percent, followed by energy prices (a subset of inflation) at 28 percent, and then abortion is third at 25 percent, with only 14 percent of voters listing abortion as their top concern.

Moreover, that same poll found that the Dobbs decision would have a “neutral” impact on the midterms. An equal percentage of voters indicated that the decision would make them “more likely” to vote Democratic or Republican.

But that’s mere issue polling. Nothing is at stake when a voter answers an issue poll. The Kansas referendum was as real as it gets, and turnout was through the roof. More voters voted in the referendum than voted in the Republican and Democratic primaries combined, and the result wasn’t close—which meant that the margin had to include Republican crossover votes. As Philip Bump noted in The Washington Post, every single Kansas county voted to the left of its 2020 election results.

What does this all mean? I have some preliminary thoughts:

First, the Kansas numbers indicate that issue polls might be underestimating pro-choice turnout and intensity. Whenever virtually the entire political commentariat is taken by surprise—including those people who sleep, eat, and breathe polling data—that suggests that even the best pollsters are failing to capture public sentiment. Is pro-choice intensity greater than polls indicate? The Kansas results suggest the answer is yes.

Second, the pro-life position may be most vulnerable when abortion is literally on the ballot. A vote on a referendum is fundamentally different from a vote for a candidate. A referendum asks you about a single issue, and if the pro-life position is the minority position, then issue voting is going to be challenging everywhere except perhaps the country’s most conservative states. In fact, a New York Times analysis suggests that four out of five states would support abortion rights “in a similar vote.”

Third, abortion rights are rarely decided via referenda, and we still don’t truly know voters’ priorities in candidate elections in this post-Dobbs world. As my Dispatch colleague Sarah Isgur noted in her newsletter, voters often “show fundamentally different priorities on ballot measures than when voting for candidates.” In other words, pro-choice Republicans still vote for Republicans even when they’re pro-choice, or swing voters may vote Republican because they care more about sending a message over inflation than they care about abortion.

Finally, over the long term, the side that’s least willing to compromise on anything is most apt to lose everything. Activists on both sides should remember that they don’t command majority support. There’s a strange dynamic that we’ve seen from Democrats in the Senate: On the one hand, they recognize the unpopularity of the Dobbs decision. But they proposed legislation so extreme (it went beyond Roe in protecting abortion access) that it’s even less popular than an abortion ban.

I’ve never been under any illusion that my position is the majority position. I’m willing to support incremental changes in abortion policy. An all-or-nothing political position (from either side) will often mean “nothing,” except in the deepest red or deepest blue jurisdictions. Yet if Kansas teaches us anything, it teaches us that all commentary is conjecture—until the voting starts.

One final note: I'm truly overwhelmed by the response to last week's newsletter. I received more kind notes than I could possibly respond to, and I forwarded many of them to my wife, the hero of the story. While I couldn't respond to every note, I want you to know that I read all of them. Thank you.

If you don't know what I'm talking about, I'd urge you to read the piece. But not because of me or my thoughts. Rather it's a story of my wife and her courage. A moment that seemed profane turned sacred. It was unforgettable.