How Did Abortion Ballot Initiatives Impact the Presidential Vote in 2024?

This seems like a lifetime ago, but the 2004 election was actually a really interesting one in terms of religion and politics. George W. Bush was seeking a second term and his Democratic challenger was John Kerry. The Republicans devised a strategy to goose turnout in a bunch of battleground states. Same-sex marriage was an incredibly salient issue at the time and the majority of the public was opposed to allowing gay couples to get married. So, Republican parties in thirteen states managed to get the question put on the ballot.

In all thirteen states, the ballot measures that would enshrine marriage between a man and a woman passed with a majority. That was even true in the state of Oregon, where 57% of voters said yes to this statement, “Amends Constitution: Only Marriage Between One Man and One Woman Is Valid or Legally Recognized as Marriage.” Did that strategy actually propel Bush to a victory in 2004? According to this paper from David Campbell and Quin Monson, it appears that white evangelical turnout increased in states with those amendments on the ballot. But they also found that non-religious voters turned out at much lower rates, and they argue this actually was more instrumental in Kerry’s loss.

Similarly, the 2024 election faced a situation that echoes the circumstances of 2004. In the Dobbs decision, the Supreme Court essentially turned the question of abortion regulation back to the states. That means that the ballot initiative/referendum has become the instrument through which states can set limits (or not) on abortion access. And in total, ten states had a ballot measure that directly dealt with the issue of abortion in 2024.

So, here’s my question - how much was Trump’s vote share correlated with those abortion measures?

In an effort to make this post more condensed, I won’t go into the nuances of each abortion measure. Some would have extended the window for access to a legal abortion, others removed any language about a specific week limitation. They were all just a little bit different. This CNN page is really helpful in summarizing each of them. That’s also where I got my county level data from to perform this analysis.¹

Let me start by just showing you the maps of a couple of states on their abortion access measures. This first one is Florida, which currently has a six week ban in place on the procedure. Amendment 4 would have increased that to the point of fetal viability.

There’s this well worn saying about Florida politics - the farther north you go, the more south it gets. That’s most certainly the case when it came to Amendment 4. The panhandle of the state was the least supportive of expanding abortion access. In fact, you hardly find any red at all when working your way down the peninsula. The areas that were the most supportive of expanding abortion were the counties around Tampa on the west side and then the area surrounding the Miami metro.

In total, about 57% of voters in the state cast a ballot in favor of Amendment 4. That means that it failed, as the threshold for passage is 60%. Here’s a fun fact - for decades the bar for passage was a simple majority. Then, in 2006, an initiative was placed on the ballot to increase it to 60%. It passed with 57.8% of the vote. I love democracy.

Let me show you one more state - Colorado. The story there was much different.

Colorado Amendment 79 would “establish a right to abortion at any stage of pregnancy.” It passed with nearly 62% of the vote. You can clearly see the geographic differences in the politics of the state. It’s like a huge blob of blue paint was dropped in the center of the map and then it just splattered on all the surrounding counties. In the Denver area, support was two-thirds of voters. Even in Colorado Springs, which is often considered to be the epicenter of evangelicalism, 52% of votes were in favor of expanding abortion access.

Here’s a link to a state map for all the rest and quick descriptions of the outcome.

• Missouri - 52% of voters were in favor of expanding access to abortion.

• Arizona - 62% in favor.

• South Dakota - 41% in favor. ← Failed

• Maryland - 75% in favor.

• Montana - 58% in favor.

• Nevada - 64% in favor.

• New York - 62% in favor.

Then, Nebraska - which was a weird one. There were actually two initiatives on the ballot. Initiative 439 would have expanded access to the point of viability - here’s a map of that one. It got 49% of the vote. Initiative 434 would have enshrined the current standard, access through 12 weeks. It passed with 55% of the vote.

But, let’s get to the point of this whole thing now - how well did those abortion votes track the presidential vote on November 5th? Here’s what I did - a basic scatterplot. On the x-axis is the share of each county that voted in favor of expanding abortion access. On the y-axis is the share of the county vote for Donald Trump.

As you can undoubtedly see - the trend line is the negative direction. More votes for expanding abortion rights led to fewer votes for Donald Trump. Social science is shocking sometimes! But how much did those two things relate to each other? Almost perfectly. In a regression model, for every percentage point increase in support for abortion access, Trump’s vote share dropped by .975%. So, essentially it was a 1 for 1 drop.²

But if you study the scatter plots colors a bit, you can see that some states tend to have a bit of a different pattern than others. So, I broke this down into ten different models - one for each state.³

Of course, those same basic trend lines are here - more pro-choice votes equates to fewer ballots cast for Donald Trump. But I also included the slope of each line in the top right corner of each graph. For instance, take a look at Arizona. For every 1% increase in abortion access votes, Trump did 1.4% worse. In Maryland it was a bigger difference - Trump lost 1.7 points for every 1 point increase in the pro-choice vote.⁴

There were some states that ran below the national average, though. For instance, in Missouri there were a lot of folks who voted for Donald Trump and against abortion access. That was also true in Nebraska, as well. But beyond the Maryland and Arizona graphs, there aren’t many examples of abortion mattering a lot more (or a lot less) than the overall average.

The more I poked around on this data, the more I began to wonder if county population had any bearing on these results. We always assume that Republicans do well in rural places, Democrats dominate urban areas, and the suburbs are a toss-up. I divided the counties into four equally sized quartiles based on the number of votes cast on the abortion measure and then did the same scatter plots.

The differences in the slopes of the lines is very minor, really. In the absolute smallest counties, a 1% increase in pro-choice votes yielded .8% fewer ballots for Donald Trump. That moved up to .9% in the middle two quartiles. In the largest of the counties, that slope did get a bit steeper. For every one point increase in vote for expanding abortion access, Trump’s share dropped by 1.1 points. In other words, support for abortion appears to have hurt Trump more in suburban and urban areas. It hurt him a bit less in rural parts of these states.⁵

How about we take a look at some places where there was clear divergence in the presidential vote and the abortion vote? Here’s what I did - I only considered the largest 25% of the counties in the data (otherwise you get some real weirdness when <1,000 votes were cast). The table below is the counties where Trump ran the farthest ahead of the pro-life vote. These are places with the highest concentration of pro-choice Republicans.

Note the recurring colors: green for Maryland and yellow for Florida. Those are two places where abortion votes in populous areas don’t really track with Trump votes. For instance, in Cecil County, Maryland, Trump got 72% of the vote but only 37% of ballots were in favor of restricting abortion access. Monroe County in Florida is an interesting one. It’s the southernmost county in the state and encompasses the Keys. About 60% of ballots were for Trump, but just 36% of folks wanted to restrict access to abortion.

What about the other direction? I went hunting for counties that had larger shares of votes for restricting abortion than votes for Donald Trump. I didn’t find hardly any in those larger counties. In fact, there were only 2 out of 143 were there was a significant difference.

In this entire dataset of 143 counties, there is exactly one where the pro-life vote share ran ahead of the pro-Trump vote share by at least one percentage point. It was Monroe County, NY. That’s Rochester. And even there, the difference was just 1.3 points.

Many counties on this list show similar percentages for pro-life votes and Trump support. A whole bunch of them are in the state of New York and a handful are in Colorado. But the upshot is pretty clear - you don’t find a lot of counties that are super pro-life that don’t like Donald Trump.

Let me put some actual numbers on this.

In 2% of counties, pro-life beat pro-Trump by at least 1 point.

In 95% of counties, pro-Trump beat pro-life by at least 1 point.

A couple of years ago I wrote a piece for Religion News Service with the headline, “Abortion just isn't the motivating issue for evangelicals it once was,” based on 2019 data from PRRI. Among white evangelicals, just a quarter of them said that their candidate “must share their views on abortion.” In other words, 75% of evangelicals would consider voting for a pro-choice candidate.

The data tells a clear and compelling story: abortion is not a deal-breaker for many Republicans. Appealing to those abolitionists who want to jail women for seeking the procedure is a politically losing strategy. Those on the far right are going to vote for the GOP no matter what. Trying to endorse their positions will only alienate the huge chunk of voters who are uncomfortable with the idea of abortion but take a libertarian position.

If there’s one enduring lesson from 2024, it’s this - the abortion discussion is largely irrelevant at the national level. The states are going to decide - which may have been the best outcome all along.

Code for this post can be found here.

1. When I scraped this data, all the votes weren’t in. However, in the vast majority of counties, at least 85% of the votes were tabulated. I think that these results will shift just a little bit as those last tranches of votes get counted, but I can’t imagine it will change the overall conclusions of what I do here.

2. The R-Squared for this simple bivariate model was .77.

3. For Nebraska, I included the Initiative 439 - that’s the one that would have expanded abortion access beyond 12 weeks.

4. The R-Squared for each state are as follows:
Arizona: .73, Colorado: .96, Florida: .87, Maryland: .88, Missouri: .75, Montana: .91, Nebraska: .76, Nevada: .84, New York: .93, South Dakota, .83

5. Here’s the share who voted in favor of abortion access by population size:
Bottom 25%: 34% - Second 25%: 39% - Third 25%: 43% - Top 25%: 57%