Six months ago, the outcome of the November 3 midterm elections would have seemed impossible. Most states and regions were highly polarized politically. So-called purple states were generally a reflection of a balance of voters from different areas of the state that offset each other rather than a collection of moderates and independents.
Yet, states made enormous swings. I spent almost three decades living in Texas (and only moved away last year). A state that had not elected a single Democrat to a statewide office since 1994 saw every statewide office taken by Democrats. Similar stories occurred across the country, with Democrats winning in states and districts that had not flipped in decades.
There are three things to note about the election. First, registered Democrats turned out in droves compared with registered Republicans. Second, there were vast differences in polling numbers for different races within states that were often not reflected in the returns. Third, it appears that individual voters switched the party they voted for as well, though this is harder to validate on an individual basis.
What do these trends tell us about human behavior?
Motivation and action
Nationwide, about half of registered voters actually voted. That is a high turnout for a midterm election, but still means that only one out of every two voters cast a ballot. So, what gets people out to vote?
One factor that drives voting is habit. The best predictor of whether you vote in an election is whether you voted in a previous election. When it becomes a habit to vote, then you do it year in and year out in primaries and general elections.
Beyond habit, though, actions are generally driven by a combination of energy and efficacy. That is, someone needs to be energized to engage and needs to believe that voting has an impact on the world.
Energy comes from dissatisfaction. People are energized when there is a gap between where they are and where they would like to be. 2026 saw a high level of turnout in part because many people felt that gap between where their lives are (and where the country is) and where they (and the country) should be.
That energy only becomes action when people feel that the voting process matters. When people believe that the choice of candidate is important and that getting out to vote influences the outcome, they vote. When you’re trying to engage people in any behavior, then, you want to energize them by emphasizing the gap between where they are and where they want to be and demonstrating that the desired action will help to minimize that gap.
Choice consistency
One surprising finding was that the polls showed more variability than was observed in the actual election. In Texas (which I was following closely), the polls generally showed James Talarico with a significant lead over Ken Paxton, while at the same time they showed incumbent Greg Abbott with a small but consistent lead over Gina Hinojosa.
In the end, both Tallarico and Hinojosa won. This didn’t just reflect an underestimate of the number of people who would vote for Democrats. If it did, you might expect that Talarico would have had a substantially larger margin of victory than Hinojosa. And while Talarico did have a larger margin (55.1% to 43%) than Hinojosa (52.8% to 45.7%), it wasn’t that much different.
Polls tend to focus on individual candidates or at least to present questions about candidates in isolation. As a result, they get a snapshot of people’s thoughts about that particular candidate at that moment. In contrast, when people step into the voting booth, they are making a large number of selections all at once. And when they do that, they often simplify the decision-making to reduce the cognitive effort they have to put in.
A straightforward way to reduce mental effort is to vote by category rather than to treat each race individually. Picking the candidate by party affiliation rather than by their own record is a straightforward approach, and (clearly) many people in the 2026 elections did just that—despite the preferences they expressed in polls. Indeed, in Texas, Democrats didn’t just win a couple of statewide races—they won every single statewide office.
Outer certainty masks inner chaos
One thing that stumped political analysts over the past decade in the age of Donald Trump was that supporters of the president typically had a seemingly unshakable level of support for him. The prevailing wisdom was that Republicans would support the president despite the high gas prices, rising inflation, and the war in Iran.
In the aftermath of the election, everyone is looking for the reason why support for Republicans shifted. The mistake in this question is looking for the reason. Instead, we should think more about the psychology of belief.
People have lots of psychological mechanisms that maintain consistency of beliefs. When there is a belief we want to maintain, we discount new information that contradicts with that belief. We minimize its importance. We focus on information consistent with what we believe in rather than information that is inconsistent. We also tend to believe things that the people around us also believe. As a result, for a long time—even in the face of lots of contradictory evidence—a person will appear to have an unwavering consistency.
But, eventually, if the weight of contradictory evidence gets too great (particularly if the social support for our beliefs also starts to crumble), a person may experience a sudden shift to a new stable configuration of belief. Indeed, their belief may not just waver, it may collapse entirely, and suddenly they find themselves supporting the opposite of their initial belief.
While there are clearly many people who continue to support the president and the Republican Party in general, a significant number of people underwent this kind of belief change in which they went from being supporters of the Republicans to being supporters of the Democrats.
But, this consistency masks a slow accretion of information under the surface. People don’t display progressively less confidence in their support of one party followed by gradual support of the other. Instead, at some point there is too much counterevidence for them to maintain their prior affiliation, and they switch. That tendency to shift wholesale is another reason why voters ultimately selected candidates by party affiliation when they got in the voting booth.
So what?
It is rare to get a decision made by so many people where so much data is collected both in the run-up to the choice and about the choice itself. So, it is valuable to draw lessons about human behavior from these situations. Certainly, these factors are valuable to know about if you’re in the business of understanding voters. But, they are also valuable in daily life. You may not be trying to affect the way people vote, but you are likely to be trying to influence the behavior of people at work and at home. It is worth thinking about how you might use these insights to influence decision-makers in your world.