Welcome back to The Argument’s monthly poll series, where we press Americans on the issues everyone’s fighting about. We’ve done free speech, AI, and now immigration. Our full crosstabs are available for paying subscribers here and our methodology can be read here.
Americans have gotten used to thinking of politics as a battle of the sexes where women are consistently more liberal than men and female voters are Democrats’ strongest soldiers. Young women’s beliefs in particular have moved sharply to the left on issues like climate, racial equality, and gun laws.
So it immediately struck me that, in The Argument‘s most recent poll, women were not consistently registering left-wing opinions on immigration.
Asked about Trump’s performance as president, 60% of women somewhat or strongly disapprove compared with 48% of men. But when we asked the following questions about immigration policy, male and female respondents appeared indistinguishable:
Should immigration be decreased, kept at its present level, or increased? (39% of women and 37% of men said decreased, respectively.)
Should law enforcement be allowed to demand proof of citizenship? (24% of men and 24% of women responded “strongly agree” and 32% of men and 32% of women said strongly disagree.)
Which illegal immigrants should be deported? (8% of men and 8% of women said none while 38% of men and 32% of women said all.)
Remarkably, when asked which party better reflects their values on immigration, women give Democrats just a four-percentage-point advantage. Compare that to the generic ballot, where women support Democrats by an 18-percentage-point lead1.
This is a break from past poll results that showed women consistently to the left of men on immigration. In 2018, 40% of women told Pew Research Center that the priority for dealing with illegal immigration should be creating a pathway to citizenship. Just 27% of men said the same.2
Likely, this shift is due to concerns about public safety. As Lakshya Jain, The Argument’s director of political data pointed out in his own piece on Monday, there is a “pattern, where throughout recent American history, anti-immigrant sentiment has risen during times of societal unrest and spikes in crime.”
In our poll, public safety was the only dimension where respondents had a net-negative view of immigration’s local impact.
I think a reasonable hypothesis is that women, who have greater concerns around personal safety, are cross-pressured on their support for liberal immigration policy when crime rates spike and prominent stories about victimization spread.
This small polling anomaly matters because it exposes how fragile and contingent our gender stories really are.
Political memories are short. That women are expected to vote with left-leaning parties is a relatively new fact about the world — conventional wisdom used to be that women were a conservative voting bloc. One of the strangest things about discussions over gender differences in political beliefs is how inevitable people imagine those differences to be.
When I’ve raised the fact that women used to be seen as a conservative voting bloc people often register surprise: “Haven’t women always been more left wing?” No!
It actually wasn’t until 2000 that women’s leftward shift was broadly appreciated. Political scientists Ronald Inglehart and Pippa Norris famously described a new phenomenon in advanced industrial societies: By the 1990s, women had become more left-wing than men.
As Inglehart and Norris detail, “women were found to be more apt than men to support center-right parties.”
In the United States, the parties were not well sorted by ideology (lots of conservative Democrats and liberal Republicans, that is) until roughly the 1960s, so it’s hard to draw a clean story from the data. A recent economics paper looked at the 1920 congressional elections, which took place immediately following the ratification of the 19th Amendment, and found suggestive evidence of women’s conservatism.
Inglehart and Norris quoted a 1963 academic book by Princeton University Press, which claimed: “Wherever the consequences of women’s suffrage have been studied, it would appear that women differ from men in their political behavior only in being somewhat more frequently apathetic, parochial, [and] conservative … Our data, on the whole, confirm the findings reported in the literature.”
Things change; now wherever the consequences of women’s suffrage have been studied, it would appear that women are more engaged, highly ideological, and liberal.
Demographics aren’t destiny
It can be tempting to assume differences between genders are driven by immutable or biological traits. It can also be tempting to assume that those differences will persist forever. But women are people and people change their minds — both as the facts of the world change and as they are persuaded to view things differently.
Now, voters are not clay to be molded by any smooth-talking politician or halfway-decent algorithm, but I do think political observers regularly veer into the type of essentialism that is regularly trounced by history.
For instance, Democrats once convinced themselves that America’s increasingly diverse population was delivering them an emerging majority; then Latino support for the party fell 14 percentage points between 2012 and 2024. Women’s support for the Democratic Party — and for that matter, men’s support for the Republican Party — is mutable.3
Unfortunately, we’ll probably never be free of some version of Men are from Mars and Women are from Venus. But no matter who is doing it, treating sex as political destiny is a fool’s errand.
The current essentialism en vogue is contained in the increasingly popular term feminization. Large parts of the center and right have become enamored with the thesis that the rise of women’s participation in important roles across media, business, and academia is to blame for a variety of social ills. Arguments about how women are determined to oppose technological progress or how women are single-handedly responsible for all cancellations ever have moved from just fringe ideas on the corners of the internet to the mainstream.
Take Helen Andrews’ viral essay, “The Great Feminization,” where she makes the totalizing argument that “Everything you think of as wokeness involves prioritizing the feminine over the masculine: empathy over rationality, safety over risk, cohesion over competition.”
Stereotyping people due to immutable characteristics like race or sex is bad for a lot of reasons. For one, it’s quite rude.4 For another, it leads you to make embarrassing errors.
Let’s take one of the specific claims in Andrews’ thesis, that female risk aversion helps explain wokeness. This is one of those things that everybody knows but there is actually lots of evidence suggesting that gender differences in risk are wildly overblown. A 2022 IZA study reviewed the evidence and found that gender can only explain a small fraction of the variance in risk attitudes. Yet the academic and public conversation still reads as if female risk aversion is the key to unlocking why women lag in the labor market.
People who cling to stereotypes often feel like their points are obvious: “I’m just looking at the world around me!” But it’s easy to mold stereotypes to fit basically any outcome. For instance, Andrews posits that women’s affinity for cohesion makes them opposed to open conflict and debate. But I could just as easily construct an internally coherent argument for why masculine norms are the root cause of wokeness. Here goes:
Wokeness is a punitive system that reflects masculine norms of authoritarian order. Just as men support the death penalty at higher rates than women, they support stronger sanctions on norm violations. If women were in charge, we would see fewer cancellations because they would support restorative justice practices that bring communities together rather than harshly enforce discipline.
As someone who spent many, many sleepless nights at college debate tournaments, you come to realize pretty quickly that the ability to construct an internally consistent argument does not necessarily correlate with truth-seeking behavior. Things can feel true without actually being true. This is why social science and randomized controlled trials are useful, because it’s easy to spin up narratives that support your views.
Stereotypes are models of the world, and even the best models are mere heuristics. Typically, a dedicated learner begins by adopting a very simple model, then complicating it. For instance, when I was very young, I asked my father what pi was, and he told me 3.14. A few weeks later in school, when the teacher asked if anyone knew what pi was, another student gave the correct answer that it was something to do with the ratio of a circle’s circumference to its diameter.
I proudly interrupted and said “No, my father told me pi is 3.14!!!”
This is why I have empathy for people who hold such overly simplistic models of how gender differences drive political beliefs. I remember what it was like to be in elementary school, thinking that everything I needed to know about pi was contained in three decimal points.
The women in our poll who are now more aligned with men on immigration — despite remaining strongly Democratic — are a perfect illustration of why feminization arguments fail. They are not automatons predestined to support immigration because of their overactive empathy5; they are responding to shifting circumstances and priorities.
Decades from now, when what it means to be a liberal or a conservative has flipped a few more times, and subgroups have shifted back and forth between various political positions and parties, arguments like Andrews’ will read as silly as those who argued that women’s suffrage wouldn’t really change politics much because they would just copy their husband’s votes.
“How will [women] vote on election day?” Mr. George Gallup himself once asked rhetorically. “Just exactly as they were told the night before.” How quaint.
When undecided voters are pushed.
An analysis by Gallup showed young women shifting 12 points in favor of increasing immigration levels from 2008-2016 to 2017-2024 while young men only moved three percentage points. (For women and men over 30 that was seven and four percentage points respectively).
None of this is to say that subgroup analysis is useless or that current trends can’t help inform what might happen in the short term. For instance, while Latino support for Democrats fell precipitously over the course of four elections, from election to election, voting patterns looked relatively stable. Change takes time.
Ugh, women are obsessed with tone policing
Or is it that women are predestined to oppose immigration because of fears of public safety?
Excellent reminder of the fluidity of political preferences of groups over time. Obviously Prohibition is probably the first major effect women had on national policy once they won the right to vote, and was indicative of the fact that women tended to be more socially conservative than men in the 20th century - higher religiosity, more interest in regulating vices. You see this as late as the classic Tipper Gore crusade against certain genres of music!
I wonder how much of the liberal shift in women's voting patterns is driven by the decline in religiosity. I am sure right wing cultural critics would argue that the religion of the left is now pc/wokism, and there's probably a kernel of truth to that, although as with all right wing angst about political correctness it's overblown and totally uninterested in examining right wing cultural rot.
" . . . the ability to construct an internally consistent argument does not necessarily correlate with truth-seeking behavior."
This is elegantly stated. The link to "randomized controlled trials" is certainly the ideal way to illustrate the hypothetical nature of purely intellectual models. Ordinary people don't have the advantage of social science training (me neither) and simulacra of objective data bombard us when we "do our own research." Ms. Demsas's watchword can be, and in less elegant form often is, deployed by those who have adopted conspiracy theories when responding to clear and cogent "normie" arguments.
The problem, I think, lies in the plasticity of our understanding of "truth-seeking." "Truth" is always contingent (a "truth" I've accepted while understanding many others don't), and for any truth-seeking statement to be accepted as reasonably objective there must always be a foundation of subjective trust in the source. When the practice of social science leads to a "replication crisis," as has been the case recently, I think it's possible to see how longstanding norms of practice have undermined trust and, through inattention, allowed that loss to spread to hard sciences, like medicine. It's not hard to find parallels in socio-political practice.