The ballot sits on the kitchen counter for three days, wedged between the grocery list and a field-trip permission slip. When she finally fills it in, it takes under a minute, and the choice feels less like a calculation than a recognition. The box she ticks simply looks like the kind of person she is. Across town her older sister, married, two children, a mortgage that renews next spring, ticks the opposite box just as quickly and with the same sense of obviousness.

Neither woman did much research that week. Both are intelligent. So what happened in those sixty seconds? The question of why are young women more liberal than they were a generation ago has become one of the most argued-over puzzles in modern politics, and most of the arguing skips the part that is actually interesting: what the thinking brain does when it settles on a belief, and why it so often settles where the people around it already stand. Anyone who has read our piece on why media shapes how we see ourselves will recognize some of the machinery.

The numbers behind the headline

Start with what is not in dispute. In its long-running social series, Gallup found that between 2017 and 2024 an average of 40 percent of American women aged 18 to 29 described themselves as liberal, roughly fifteen points more than young men of the same age. Young men have barely moved in twenty-five years; about a quarter called themselves liberal in 1999 and about a quarter still did in 2023. The movement has come almost entirely from one side. Gallup's own write-up, Gallup's analysis of young women's leftward shift, traces most of the change to the years after 2017, with abortion and the environment carrying the largest swings.

Then the 2024 election produced a twist that the headline writers mostly missed. The gap between men and women was large, but no larger than in 1980, 2012 or 2020. The gap that stood out was the marriage gap in voting. Married women backed Donald Trump by a narrow margin in the national exit poll. Unmarried women backed Kamala Harris by more than twenty points. For the first time on record, unmarried women made up a larger share of voters than married women did.

The household question

In 2002 two economists, Lena Edlund and Rohini Pande, published a paper in the Quarterly Journal of Economics with a title that would be hard to get past an editor today: "Why Have Women Become Left-Wing? The Political Gender Gap and the Decline in Marriage". Their argument was economic, and it was blunt. For most of the twentieth century, marriage worked as a private transfer of income from men to women. As marriage rates fell and divorce rose, that transfer shrank. Men, on average, came out richer and women poorer, and women's political preferences moved toward the party promising public transfers instead.

Using state-level data, they found that women living with a higher risk of divorce tended to vote further left than their income alone would predict. A follow-up study with Laila Haider in 2005 found a similar pattern across Europe. Conservative writers have long put the same idea in plainer language: when a husband is no longer the expected safety net, the state applies for the job.

For a mind deciding how to vote, this matters because the thinking brain is, among other things, an insurance broker. Much of what feels like political conviction is a running estimate of a practical question: if things go wrong, who pays? A woman whose answer is "my husband, my parents, my church" reasons from one set of premises. A woman whose honest answer is "probably no one, unless the government does" reasons from another. Both are being rational about the world as they see it from their own kitchen.

How a belief actually gets built

Economics explains the incentive. It does not explain the feeling of certainty, the sense that the other box is not merely different but faintly unthinkable. For that, the research on how minds form beliefs is more useful.

In 1977 psychologists Lynn Hasher, David Goldstein and Thomas Toppino ran a simple experiment. They gave volunteers a list of trivia statements, some true and some false, and asked them to rate how true each seemed. Two weeks later they did it again, slipping some of the earlier statements back in. Statements people had seen before were rated as truer, whether or not they were. Repetition alone moved the needle. The finding, now called the illusory truth effect, has been replicated many times. In 2015 Lisa Fazio and colleagues showed it works even when people already know the correct answer; familiarity quietly outvotes knowledge.

Political messages are repeated more than almost anything else a person hears. Whichever slogans fill a woman's classroom, office, group chat and evening television will start to feel true for that reason alone, and that is just as true of a man, or of a household that watches only one cable channel.

Repetition has a partner. Yale law professor Dan Kahan has spent years documenting what he calls identity-protective cognition: people process evidence in whatever way protects their standing in the groups they belong to. In one well-known experiment, Kahan and colleagues gave participants the same table of numbers framed either as a skin-cream trial or as a gun-control study. The most numerate participants read the skin-cream data accurately and then misread the identical data when it pointed against their politics. Being clever did not protect them. It made them better at arriving where their group already stood.

Much of what feels like political conviction is a running estimate of a practical question: if things go wrong, who pays?

The case critics make about classrooms and screens

Put those two mechanisms together and the argument many conservatives make starts to look less like a slogan and more like a hypothesis. If belief follows repetition and belonging, then whoever controls the repetition and the belonging has enormous influence over young minds. Critics point to decades of curricula, campus culture, entertainment and social feeds that, in their view, lean consistently one way, and to the long hours young women spend inside those spaces before they marry, if they marry. Our earlier piece on parents and the modern classroom covers one part of that debate.

The evidence here is honestly mixed. Gallup tested whether the rising share of women in college explained the shift and found that young women had moved left whether or not they held a degree, which suggests the campus alone is not the engine. Statistician Allen Downey, reanalyzing the General Social Survey, has argued that the growth in the gap may be smaller than the Gallup series implies. And studies on how far media exposure actually changes votes, as opposed to reinforcing what viewers already believed, have produced modest and contested results. What the psychology does support is narrower and more useful: in any environment where one view is repeated constantly and dissent carries a social cost, that view will feel more true to the people inside it.

Why the brain borrows its beliefs

It is tempting to read all this as a story about gullibility. It is closer to a story about efficiency. No one can personally verify the economics of tariffs, the science of climate models and the history of every foreign conflict before November. A brain that trusted only what it had checked itself would never decide anything. So it uses shortcuts that work most of the time: what has been heard often is probably true, and what trusted people believe is probably sensible.

For nearly all of human history, those shortcuts pointed home. The voices repeated most often were a girl's mother, father, grandmother and pastor. The group whose approval mattered was her family and the families around it. Borrowing beliefs from that circle tied a young woman's convictions to people who knew her, depended on her, and would still be there in thirty years.

The machinery has not changed. The inputs have. A twenty-six-year-old woman who lives alone, works remotely and spends her evenings on her phone is running the same ancient shortcuts, but the voices repeated most often are now a feed, and the group whose approval matters is often a network of peers she may never meet. The brain is doing exactly what it was built to do. It is simply drawing from a different table.

Views move with the seasons of life

None of this is fixed. Political views tend to drift as circumstances change, and the shifts track the same household question. Marriage, the arrival of children, a first mortgage and a growing stake in a neighborhood all change a person's answer to "who pays if things go wrong?" Researchers have long noted that married parents, as a group, lean more conservative than single adults of the same age, and that people often move right as they acquire property and dependents. Gallup also found women over 65 moving left, which its analysts attribute largely to liberal Baby Boomers aging into that bracket, a reminder that generations carry their formative repetitions with them.

Workload plays a part too. A mother juggling two jobs and three schedules has little attention left for weighing arguments, which leaves the familiar slogan and the trusted friend to do more of the deciding. Our piece on the silent cost of the second income looks at how that pressure reshapes family life more broadly.

Where the evidence pushes back

An honest account has to include the objections. Many women on the left would say their votes rest on specific issues, abortion law after the 2022 Dobbs decision above all, along with health care and the environment, and Gallup's data does show those issues carrying the biggest shifts. Reducing those positions to a search for a substitute husband underestimates the women who hold them, just as dismissing conservative married women as voting their husbands' preferences underestimates them.

Causation is also unsettled. Edlund and Pande's critics note that the arrow may point both ways: women who already hold progressive views may marry later or less often, which would produce a marriage gap without marriage itself changing anyone's mind. And the overall gap between men and women in 2024 was about the same size it has been since 1980. The striking change is concentrated among the young and the unmarried, not spread evenly across all women.

Go back to the two sisters and their ballots. Neither one was being manipulated in any dramatic sense, and neither was reasoning from nothing. Each mind looked around its own life, asked who would be there if the floor gave way, leaned on the voices it had heard most often, and produced an answer that felt obvious. Seen that way, a vote is less a verdict on the world than a portrait of the household behind it, and the most revealing question to ask about any woman's politics, including our own, is not what she believes but whose table she learned it at.