
In 2024, Will Republicans Run on Abortion or Run Away From It?
What we can glean from the midterm polling.

Why did Ron DeSantis, Floridaās Republican governor, do so well in the midterms? One reason, some pundits speculate, is that DeSantis found the sweet spot in the abortion debate. In April, two months before the Supreme Court handed down its ruling in Dobbs v. Jackson Womenās Health Organization, DeSantis signed a ban on abortion after 15 weeks of gestation. By steering clear of the first trimester, he allowed more than 90 percent of pregnancy terminations to remain legal. According to this theory, if DeSantis were to go furtherāif he were to outlaw abortions earlier in pregnancy, as pro-lifers are urging him to do, and as his legislature is reportedly consideringāhe would pay a steep political price. Some commentators have suggested that the 15-week line is a politically important threshold:
Florida's current abortion law - which allows abortions until 15 weeks - is the perfect solution politically for red-state governors with national ambitions. Tightening restrictions beyond that will dramatically raise the salience of DeSantis' anti-abortion views in 2024.
ā James Surowiecki (@JamesSurowiecki) November 13, 2022
All my instincts tell me this theory is true. And in surveys of the midterm electorate, Iāve found lots of evidence that Dobbs triggered a backlash against Republicans. But I canāt find evidence that the DeSantis 15-week position is relatively safe. Among Republican governors who ran for re-election in 2022, surveys show no apparent relationship between the time frame of a stateās abortion restrictions and the magnitude of the backlash. The network exit poll taken during the midterms covered five states in which Republican governors sought re-election. Among respondents who said abortion should be legal in most but not all casesāthe most likely target audience for a 15-week banāDeSantis got 42 percent of the vote. His neighbor, Gov. Brian Kemp of Georgia, who had signed a ban on abortions at six weeks, got 40 percent of these mostly pro-choice voters. Thatās a marginal difference at best.
In Texas, Gov. Greg Abbott, who had signed a ban on abortion at conception, got only 36 percent of the mostly-pro-choice vote. In New Hampshire, Gov. Chris Sununu, who supports abortion rights, got 62 percent of the mostly-pro-choice vote. If you confine the sample to these four states, you can try to draw a line that correlates vote share with each stateās abortion laws. But the pattern collapses when you turn to Ohio, whose ban on abortion at six weeks notoriously forced a 10-year-old girl to leave the state to end her pregnancy. In the midterms, Mike DeWine, the Republican governor who signed that law, got 61 percent of the mostly pro-choice vote. Ohio isnāt the only problem with the theory. The two-point surplus for DeSantis over Kemp turns into a four-point deficit when you account for the candidatesā overall performance. Kemp did 13 points worse among mostly-pro-choice voters than among voters overall. DeSantis did 17 points worse among these voters than among voters overall. Abbott, at 19 points, wasnāt far behind. Alternatively, you can test the sweet-spot hypothesis by measuring the percentage of each stateās electorate that voted against a Republican governor specifically because of abortion. If you start with people who said abortion was their most important voting issue, and if you then narrow that sample to the subset who voted for the Democratic gubernatorial candidate, Florida comes out even with Georgia and statistically tied with, or marginally higher than, Ohio. In other words, the abortion-driven vote against DeSantis was as big, percentage-wise, as the abortion-driven vote against Kemp and DeWine. If you then factor in the relative popularity of abortion rights in each state, DeSantis comes out marginally better than Kemp but worse than DeWine.
To broaden the analysis, letās turn to AP VoteCast, which surveyed additional states in which Republican governors ran for re-election. Here, the percentage of voters who said abortion was their most important issue was lower than in the network exit poll, because more issues were offered in the VoteCast questionnaire. If we run the same analysisācalculating what percentage of each stateās voters cited abortion as their top issue and voted for Democrats, relative to the stateās overall tilt on abortionāFlorida comes out at the lower end, but statistically about even with South Carolina (where a six-week ban is currently blocked in the courts) and with Texas and Alabama, which prohibit abortion throughout pregnancy. VoteCast data published by the Wall Street Journal show how midterm voters cast their ballots in congressional races, not gubernatorial races. But when you analyze these numbers by stateālooking at the percentage of ālegal in most casesā voters who chose GOP candidatesāRepublicans running in Florida performed slightly better than Republicans in Georgia (38 percent) and Iowa (41 percent), equal to New Hampshire (43 percent), slightly worse than Ohio (46 percent), and significantly worse than Texas (51 percent), Alabama (53 percent), and South Carolina (61 percent). Again, thereās no sign that the DeSantis position was safer. The best measure of abortionās impact in gubernatorial races comes from two supplementary questions added to VoteCast by the Kaiser Family Foundation. The foundation has posted its findings in two reports, and it ran additional calculations at the request of The Bulwark. The two questions asked whether āthe Supreme Court decision overturning Roe v. Wade had a major impact, a minor impact or no impactā on 1) the respondentās decision to vote at all and 2) āwhich candidates you support.ā By focusing on people who voted for Democratic gubernatorial candidatesāand comparing the percentage in each state who said the courtās ruling had a āmajor impactā on that decisionāwe can gauge the extent to which abortion drove the midterm vote against each Republican governor. By this measure, DeSantis did no better than his Republican colleagues. Among Floridians who cast their ballots for his Democratic opponent, the percentage who cited Dobbs as a major factor in their decision to vote at all was statistically indistinguishable from the percentage who cited the same factor as their basis for turning out in other states. Republican governors who banned abortion at six weeksāand, for that matter, Republican governors who banned abortion at conceptionādid no worse than DeSantis, on average, in terms of the extent to which abortion drove the turnout against them. This remains true even when you weight the data to account for abortion rights being less popular in some states than in others. When you look at vote choice, as opposed to turnout, the DeSantis numbers are even less auspicious. Among people who voted for Democratic gubernatorial nominees, the percentage who cited Dobbs as a major factor in their selection between the candidates was higher in Florida than it was, on average, in states where abortion is banned from conception. This gap goes away when you adjust the numbers to account for abortion rights being less popular in some states. But the result of this adjustment is a complete muddle, with all three categories of statesāthose with bans at conception, those with bans at six weeks, and Floridaāin a statistical tie. By this measure, only Sununu, the lone pro-choice Republican governor, comes out ahead. This isnāt what I expected to find when I began to look into the numbers. The idea of 15 weeks as a relatively safe compromise, politically, makes sense to me. I thought the exit polls would back it up. They donāt. Iām still betting that the prediction about DeSantis is correct: If he keeps pushing on this issue, heāll raise its salience, brand himself with it, and thereby damage his chances of winning a general election for president. But from the evidence of the midterms, I canāt prove it.