On June 24, the Supreme Court issued a ruling in Dobbs v. Jackson declaring that the right to an abortion was not constitutionally protected and could be blocked by states. The ruling, which was deeply unpopular across the nation, set off a firestorm of debate on both sides regarding legislation and electoral prospects. Questions abounded as to whether Democrats could experience an electoral bounce that could help stave off what was looking to be an increasingly catastrophic November, and unsurprisingly, the most prudent advice was (and remains) to wait and see.
But a month has now passed since the ruling, and between polls, fundraising data, and a special election in Nebraska, we now have a decent amount of data to assess the electoral landscape in the wake of the decision. And the overall picture does suggest a small bounce towards Democrats in its aftermath, a picture Split Ticket has confirmed privately with Republican and Democratic internals on both sides.
On June 25, the polling average on FiveThirtyEight was R+2.3, and on RealClearPolitics, it was R+2.8. Since then, FiveThirtyEight’s average has dipped to R+1.1, as of July 21, and RealClearPolitics’ to R+1.8. Both sides suggest a consistent bounce of approximately a point in margin, going by the polling aggregates.
What about individual polls, excluding partisan ones such as Democratic and Republican internals? FiveThirtyEight provides data on poll affiliations and sponsors, and using the data listed on the site, we can do a before-and-after comparison of several pollsters that polled the race in the month before and after Dobbs to see how the picture has changed, if at all.
The criteria for the polls we’ll use is as follows: any nonpartisan poll that polls at least 800 registered or likely voters (to get a good sample size) will be included on this, except for YouGov/The Economist, which showed an R+4 -> D+3 bounce that was less of a function of the race shifting and more because of a data quality issue with their previous sample that makes actual comparisons impossible to trust. When a firm polls both registered and likely voters, we’ll use the likely voter number.
|AVERAGE||R +0.2||D +2.1||D +2.3|
In the six weeks prior to Dobbs, the polling average among these pollsters was R +0.2. In the four weeks since, the raw average among the nine pollsters listed above is D +2.1, for a D +2.3 shift overall.
This is not to say that the national environment will be Democratic in the fall. Indeed, the past two Republican waves have seen the Democratic support fall off sharply come Labor Day, and this means that any assessment of the race must be couched in significant uncertainty and understanding that things may very well degrade sharply for Democrats if independents do break against the President’s party, as has happened in the past. Split Ticket already predicts the House to flip Republican, and while the Senate is a pure tossup, we would probably tilt towards Republicans if you forced us to make a choice today.
But caveats aside, what the data above does tell us is that regardless of the permanence of the effect, polling has consistently shown a shift towards Democrats from before the ruling to now. Seven of the nine polls listed show a change between the pre and post-Dobbs environments, and when a sample size this large shows a shift of approximately two points in magnitude for the nation, it’s probably worth adjusting our priors a bit.
Why might this be happening? For starters, legislation that makes abortion illegal in most or all cases, which the ruling opens the door to, is usually extraordinarily unpopular, as Split Ticket discussed a while back. Moreover, the ruling in this case involves taking away a right that many Americans already had, which is usually quite unpopular.
It’s also possible that this is a function of nonresponse bias, and that Democrats are responding to polls at a higher rate than Republicans immediately in the wake of the ruling. But it’s worth pointing out that nonresponse environments really do look different in 2022 than they did in 2020, as both Nate Cohn of the New York Times and Elliott Morris of The Economist seem to say.
There are two other datapoints that add evidence to any argument. The first is small-dollar fundraising, which can serve as a proxy for enthusiasm and engagement. Here, the Democratic Party has seen a massive influx of donations in the wake of the ruling, which further compounds the huge advantage in small-dollar donations they enjoyed even before the decision. Money certainly isn’t everything, and the advantage Democrats have in fundraising is likely at least partly due to their coalition being far more college-educated. But even then, the surge since the ruling does suggest a currently higher rate of Democratic engagement, which the party will pray holds in November to stave off a potential electoral cataclysm.
The second is the special election in Nebraska’s first congressional district, held just four days after the Dobbs ruling, to replace scandal-plagued incumbent Jeff Fortenberry. Here, Democrats stunned analysts and observers with a massive overperformance relative to expectations, losing by 5.4% in a seat that Joe Biden lost by 11 points in 2020. Some of this was likely down to district and campaign effects, and it is worth noting that this district has a lot of college-educated whites, a left-trending demographic that tends to turn out more in irregular elections anyways. But an overperformance of nearly six points from Biden’s 2020 margins, combined with the fact that Democrats saw a solid surge in election day turnout, suggests that something else was at play, and it is entirely possible (and perhaps likely) that the post-Dobbs surge probably did help them here.
Between the special election, the polling, and the fundraising numbers, it is probably fair to say that the evidence we do have suggests there has been a material change in the environment in the wake of the Dobbs ruling, and that Democrats probably are experiencing an overall national shift of two points in margin, at least temporarily (though we expect that there will be significant variation by demographic and by state on both persuasion and turnout effect).
This does not mean the effect will hold; if the economy worsens, the issue may fade in salience down the stretch, meaning that it is worth keeping an eye on the environment to see if the effect holds. But the question we are trying to answer today is whether the ruling has had an electoral effect up until this point. And in our opinion, the data suggests that this is more likely than not to have been the case.
I’m a software engineer and a computer scientist (UC Berkeley class of 2019 BA, class of 2020 MS) who has an interest in machine learning, politics, and electoral data. I’m a partner at Split Ticket, handle our Senate races, and make many kinds of electoral models.