How Voters Actually Decide: The Power of Party Identification

the immigration example common picture of an election is a marketplace of ideas. Voters are imagined as rational shoppers, weighing a candidate’s tax plan against their own preferences, then casting a ballot for the best match. Political scientists call this the “menu” view of democracy. In October 2025, G. Elliott Morris named this belief the Strategist’s Fallacy: the assumption that voters hold fixed positions and then select the candidate who offers the closest entree. Decades of research suggest this is almost entirely backward. Voters do not think in left and right, at least not in the way strategists hope. Instead, the most powerful force in an election is party identification—a deep, emotional attachment formed early in life, often inherited from family and reinforced by social circles. This group loyalty shapes not just which candidate a person votes for, but which facts they accept, which issues they care about, and how they interpret the entire political world.

The evidence for this is strong. In the 1950s, political scientist Converse found that only about one voter in nine reasons ideologically, meaning they can consistently apply abstract principles like “liberty” or “equality” to specific policies. Later work by Kinder and Kalmoe found that share has barely moved in seventy years. Morris puts the current figure at 10 to 15 percent of the electorate. The vast majority of voters are not ideologues. They are group members. When asked about policy, they do not consult a personal doctrine; they ask what their party says, or more accurately, what their party feels like. This is why a voter can support a policy when a Democrat proposes it and oppose the identical policy when a Republican does. The policy is secondary. The team is primary.

This reality explains why campaign persuasion is so weak. In a landmark analysis of 49 field experiments, Kalla and Broockman found that the best estimate of campaign persuasion in a general election is zero. Ads, door-knocks, and mailers do not move votes in any measurable way. They can mobilize people who already agree, but they rarely convert anyone. The reason is that most voters are not persuadable voters in the strategist’s sense. They are loyal partisans. The small slice of true swing voters—people with weak party ties and low information—are often the least ideological of all. They decide based on vague impressions, current conditions, and who seems more like “us.” This is why the moderation debate, the fight between moderates versus progressives within a party, often misses the point. Moving to the center may appeal to a handful of undecided independents, but it does little to change the underlying structure of group loyalty.

What actually moves elections, then, is not argument but identity and conditions. Broockman and Skovron found that legislators in both parties overestimate how conservative their constituents are, a misreading born from hearing from the loudest, most ideological voices. But the broader public is more tribal than ideological. Morris’s two-input fundamentals model—based on economic growth and presidential 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, a margin smaller than a single congressional district. That is not persuasion; that is the random jitter of a system driven by loyalties and circumstances.

The practical implication for parties is uncomfortable. WelcomePAC’s report Deciding to Win recommends a 2028 Democratic nominee willing to break with progressive advocacy groups, but that advice is about coalition management, not conversion. A party cannot talk people out of their identity. It can only hope to activate its own base and, occasionally, win over the small number of genuinely unanchored voters. The Strategist’s Fallacy persists because campaigns need to believe their words matter. But the research is clear: voters decide first who they are, and then they decide who to vote for. Everything else is commentary.