Just how much have gender attitudes really changed?
Just how much have gender attitudes really changed?
In which I rethink my priors on the rise of egalitarianism
If you’ve ever taken a college-level writing class, you’re probably familiar with the concept of “motivating” a piece of academic prose. To draw a reader in, it’s critical to establish the stakes, introduce a puzzle, or otherwise find a way to convince them early on that there’s a good reason to keep slogging through your text.
In my own academic prose, I’ve noticed—with some concern—that I tend to deploy a similar motivating puzzle across texts. Here it is in the abstract for an article I published in 2020: “Despite widespread support for gender-egalitarianism, men’s and women’s household contributions remain strikingly unequal.”
Here’s another version from a grant application I submitted just last week: “Contemporary marriage ideals emphasize equality and mutuality…Yet different-gender married and cohabiting relationships are often marked by labor and power imbalances largely favoring men.”
In less formal language, I’m basically saying ‘Our gender ideals have changed, but our gendered behaviors haven’t. Isn’t that weird?’ If I’ve done my job right, the reader is now wondering about the apparent contradiction between ideology and action and eager to read on to learn how I make sense of that disjuncture.
The source of concern is not my lack of originality. All scholars have a particular intellectual beat, and mine is trying to understand what’s changed, what hasn’t, and why in the contemporary U.S. gender structure. No, the problem is that I’ve recently been having doubts about the first part of my motivating premise: have our gender ideals really changed?
This whole election cycle – really, the whole eight-year saga from 2016 to now – plays a large part in my growing doubt. If half the country is apparently ready to reelect a noted misogynist and sexual assaulter to the highest office, just how “widespread” can gender-egalitarian ideals really be?
The rise of the so-called manosphere, exemplified by influencers like Andrew Tate, is also concerning. These (mostly) men peddle retrograde gender roles at best, and horrifying sexism at worst, and they are apparently immensely popular. Are their millions of viewers hate-watching, or do they perhaps buy into a relationship ideal marked by something decidedly different from “equality and mutuality”?
Then there are the scholars in my own field, like my friend Joanna Pepin, who argue that the ideological progress narrative I so often rely on just doesn’t fit the story emerging from recent data. In a 2020 paper about high schoolers’ gender attitudes, for instance, Pepin and coauthor Brittany Dernberger found that young people continued to prefer “conventional arrangements” (i.e., man as primary breadwinner, woman as primary caregiver) for their family lives. So much for a new generation of egalitarians coming of age?
All of which has led me to a mini existential crisis in which I’m questioning not only whether gender attitudes have changed but also, What even is this thing we call “gender ideology” (or, sometimes, “gender ideals”, or “gender attitudes”)? And how well are we actually measuring it?
I suspect this is going to become an ongoing preoccupation, so for this week I’ll just focus on a) explaining how we typically measure gender attitudes, and b) identifying some of the problems and limitations of these measures.
How do we measure gender attitudes?
Most of the time when I (and other scholars) write that Americans have become more egalitarian in recent years, I cite one of a handful of big, nationally representative surveys to back up my claim: the General Social Survey (GSS), say, or the World Values Survey (WVS).
Each of these (and many other similar) surveys offers respondents a series of gender-related statements and asks them to rate their level of (dis)agreement or (dis)approval. Here are a few from the GSS, my go-to source:
Most men are better suited emotionally for politics than are most women.
A working mother can establish just as warm and secure a relationship with her children as a mother who does not work.
It is much better for everyone involved if the man is the achiever outside the home and the woman takes care of the home and family.
Because of past discrimination, employers should make special efforts to hire and promote qualified women.
Survey designers typically include multiple statements because “gender ideology” is complex in the same way religious and political ideologies are: just because I support the right to abortion doesn’t necessarily mean I will support increasing taxes on the wealthy. Statistically speaking, if I endorse one plank of liberal ideology, I will probably endorse the other. But just asking people how liberal they are would lead me to overlook important distinctions between social and economic liberalism that are only captured when you break “liberalism” down into multiple items.
Similarly, the various survey items related to gender capture (at least in principle) multiple dimensions of gender ideology. How does the respondent feel about women in public life? What about gender roles at home? About the existence of essential differences or a hierarchical relationship between men and women?
Researchers can then take the survey results, do some sophisticated quantitative analysis, and identify categories of respondents who tend to answer similarly across multiple questions.1 Perhaps one group dislikes the idea of women working and prefers that men be political and business leaders; we might call them “Traditionals.” Another group is fine with gender equality in the public sphere but prefers a more old-school arrangement in family life; maybe we label them “Neotraditionals.”
To understand how gender ideology is changing, we can then chart the shifting proportion of respondents who fall into each category over time: maybe in 1972 most survey-takers were Traditionals, but by 2023 they’d fallen into the minority.
What’s wrong—or at least, limited—with those existing measures of gender ideology?
Every survey does it slightly differently: As I learned when trying to put together a gender ideology questionnaire recently, there are approximately 5,000 ways to ask about gender attitudes and little scholarly consensus about which questions are best for any given purpose. If I want to understand someone’s beliefs about working mothers, should I go with the GSS and ask them whether “a preschool child is likely to suffer if his or her mother works”? Or, should I follow the ISSP and ask if “family life suffers when the woman has a full-time job”? These are similar questions, yet different enough that they may be capturing different underlying beliefs.
Question wording hasn’t always kept up with changes in contemporary vernacular: Consider this doozy from the GSS: “There's been a lot of discussion about the way morals and attitudes about sex are changing in this country. If a man and woman have sex relations before marriage, do you think it is always wrong, almost always wrong, wrong only sometimes, or not wrong at all?” Setting aside the fact that the question wording leaves no room for those who think it is always right, who says “sex relations” in 2024?! The fix is seemingly simple: just change the question to something about “premarital sex” or “sex before marriage,” terms in more widespread usage today. The catch is that by doing so you would risk losing the possibility of measuring change over time. And so survey designers always face a tradeoff between continuity and contemporary resonance.
Social desirability bias, aka political correctness, is always lurking: Even when they’re taking an anonymous survey, people tend to want to look good, or at least not like total assholes. Deep down, I might believe that men are better suited for politics than women, but I might be afraid to admit that out loud if I’m worried that’s not the “right” way to think, or that I’ll be mistaken for a sexist. That’s always a problem for surveys, but particularly so when it comes to measuring opinions on controversial topics like race or gender.
Interpretation is inherently subjective: Consider the graph below, which shows the % of women (38%) and men (24%) who strongly disagree with the statement that it’s better for a man to work and a woman to stay home. (The GSS Data Explorer doesn’t make it easy to aggregate “strongly disagree” and plain old “disagree,” but when you do that, you see that in 2022 79% of women and 69% of men responded in one of those two categories.)
What’s pretty unambiguous here is that most people disagree with the statement, and that the disagreers have grown in number over time. But do you focus on that majority, or on the fact that 21% of women and 31% of men agree with it? Furthermore, what do you make of the pretty sharp dip from 2021 to 2022? Is that a meaningful shift, indicative of a resurgence in traditional beliefs, or just a blip? It’s relatively easy to frame the same “objective” data in different ways depending on the story you want to tell.
We don’t do a good job distinguishing mere tolerance from actual preference: Most of the attitudinal measures I’ve mentioned so far are rather general, in the sense that they get at people’s tolerance of a variety of personal and professional choices other people, or people in general, might make. That is not quite the same thing as asking someone what they personally prefer. For example, I might think it’s perfectly fine for some other family to pursue a dual-breadwinner arrangement. I don’t want to yuck their yum! But for my own family, I might strongly prefer to have one parent at home. This is the insight that Dernberger and Pepin made in the 2020 paper I referenced above: they took advantage of a unique data set that asked high school seniors whether a variety of division of labor configurations were unacceptable, somewhat acceptable, acceptable, or desirable. Is it desirable for a husband and wife to both work full time, or merely acceptable? What about for a wife to work full-time and a husband to stay home? When they allowed people to choose this way, across a gender-balanced set of configurations, they found widespread tolerance for multiple scenarios. But the most desirable situation among 2014 seniors was still “husband full-time, wife at home,” as it had been consistently since 1976. And 53% of the 2014 high schoolers said that having a husband at home while the wife works full-time was unacceptable.2
My hunch, given the above issues, and others I haven’t space to mention, is that while the ideological progress narrative I so often rely on isn’t entirely wrong, it’s not quite right either. Alas, actually figuring out what a better narrative might be will have to wait for another time.
There are lots of examples of this kind of analysis. In a 2017 paper I often cite, Carly Knight and Mary Brinton analyze seven questions from the WVS (plus some other surveys) and come up with four classes of respondents: traditionals, liberal egalitarians, egalitarian familists, and flexible egalitarians.
I’m really describing two issues here: the tolerance versus preference thing, but also the fact that most big surveys don’t ask about both men’s and women’s paid and unpaid work roles; they tend to focus more on women. And we know, from lots of other research, that social norms around what women can do have expanded much more than those around what men can do.