The gap between how people talk about elections and how elections actually work is wide, and it matters most in the handful of states that decide the presidency. Political science has spent decades studying what voters do, and the findings are often counterintuitive. The most useful frame for understanding this is something called the Strategist’s Fallacy, a term coined by G. Elliott Morris in October 2025. It describes the belief that voters hold fixed positions and match them to candidates like an entree off a menu. In reality, the research suggests, that is not how most people decide.
What political science says about how voters decide starts with a basic fact: voters do not think in left and right. Early work by Philip Converse found that only about one voter in nine reasons ideologically, using a consistent set of abstract principles to evaluate politics. Later research by Kinder and Kalmoe found that share has barely moved over decades. Morris puts it at 10 to 15 percent of the electorate. For the other 85 to 90 percent, ideology is not a driver. Party loyalty is. Most voters are not weighing platforms; they are expressing an identity. Partisanship acts as a lens through which facts, events, and candidates are filtered. This is not a failure of character; it is simply how the human mind handles complexity.
That leaves a small but crucial group: swing voters. These are not necessarily moderates in any policy sense. They are persuadable voters who lack strong partisan attachments or who cross over in specific circumstances. They are the reason campaigns obsess over a few states and a few hundred thousand people. But here is the uncomfortable finding for campaign professionals: persuasion mostly does not work. Across 49 field experiments, Kalla and Broockman’s best estimate of campaign persuasion in a general election is zero. Ads, canvassing, mailers, rallies—none of it moves the needle in a measurable way on average. What campaigns do is not persuade; it mobilizes. They turn out people who already agree with them.
This is where the moderation debate enters. Many strategists argue that candidates must move to the center to win swing voters. The logic seems obvious: if elections are close, you need the middle. But the research complicates that. Legislators in both parties, as Broockman and Skovron found, overestimate how conservative their constituents are. They project a hawkish, moderate-right norm onto districts that are often more liberal on specific issues. Meanwhile, voters do not think in left and right, so a candidate shifting along that line may be communicating in a language most people do not speak. What matters more is perceived group membership, trust, and whether the candidate feels like “one of us” or “one of them.” That is why the debate between moderates versus progressives inside a party can feel so intense: it is a fight over identity signals, not just policy positions.
The practical consequence is that national elections are decided by tiny margins in a few states, not by a national mood swing. Morris’s two-input fundamentals model—based on economic conditions and approval—forecast Harris at about 48 percent of the two-party vote in 2024; she got 49.3 percent. On conditions alone, Trump wins about nine times in ten. The three decisive states turned on 229,766 votes. That is not a landslide; that is noise around a stable baseline. The fundamentals set the range, and the last few thousand voters in Pennsylvania, Wisconsin, and Michigan choose the winner.
This is why strategy debates get so heated. A report from WelcomePAC, titled Deciding to Win, recommends a 2028 Democratic nominee willing to break with progressive advocacy groups. The argument is that swing voters respond to signals of independence from the party’s activist wing. But the research suggests that such moves may matter less than who turns out. If persuasion is near zero, then the only lever is mobilization, and mobilization is driven by enthusiasm, not by triangulation. The tension is real: party bases want purity, swing voters want authenticity, and the fundamentals do not care about any of it. In the incumbent’s economy end, the election is decided by a rounding error in a few states, and everyone else is left to argue about why.
