Why you’ll never establish the existence of distinct “generations” in American society

An update from Pew, today’s thoughts, and then another data exercise.

Pew response

After sending it the folks in charge at the Pew Research Center, I received a very friendly email response to our open letter on generation labels. They thanked me and reported that they already had plans to begin an internal discussion about “generational research” and will be consulting with experts as they do, although the timeline was not given. I take this to mean we have a bona fide opportunity to change course on this issue, both with Pew (which has outsized influence) and more widely in the coming months. But the outcome is not assured. If you agree that the “generations” labels and surrounding discourse are causing more harm than good, for researchers and the public, I hope you will join with me and 140+ social scientists who have signed the letter so far, by signing and sharing the letter (especially to people who aren’t on Twitter). Thanks!

Why “generations” won’t work

Never say never, but I don’t see how it will be possible to identify coherent, identifiable, stable, collectively recognized and popularly understood “generation” categories, based on year of birth, that reliably map onto a diverse set of measurable social indicators. If I’m right about that, which is an empirical question, then whether Pew’s “generations” are correctly defined will never be resolved, because the goal is unattainable. Some other set of birth-year cutoffs might work better for one question or another, but we’re not going to find a set of fixed divisions that works across arenas — such as social attitudes, family behavior, and economic status. So we should instead work on weaning the clicking public from its dependence on the concept and get down to the business of researching social trends (including cohort patterns), and communicating about that research in ways that are intelligible and useful.

Here are some reasons why we don’t find a good set of “generation” boundaries.

1. Mass media and social media mean there are no unique collective experiences

When something “happens” to a particular cohort, lots of other people are affected, too. Adjacent people react, discuss, buy stuff, and define themselves in ways that are affected by these historical events. Gradations emerge. The lines between who is and is not affected can’t be sharply drawn by age.

2. Experiences may be unique, but they don’t map neatly onto attitudes or adjacent behaviors

Even if you can identify something that happened to a specific age group at a specific point in time, the effects of such an experience will be diffuse. To name a few prominent examples: some people grew up in the era of mass incarceration and faced higher risks of being imprisoned, some people entered the job market in 2009 and suffered long-term consequences for their career trajectories, and some people came of age with the Pill. But these experiences don’t mark those people for distinct attitudes or behaviors. Having been incarcerated, unemployed, or in control of your pregnancy may influence attitudes and behaviors, but it won’t set people categorically apart. People whose friends or parents were incarcerated are affected, too; grandparents with unemployed people sleeping on their couches are affected by recessions; people who work in daycare centers are affected by birth trends. And, of course, African Americans have a unique experience with mass incarceration, rich people can ride out recessions, and the Pill is for women. When it comes to indicators of the kind we can measure, effects of these experiences will usually be marginal, not discrete, and not universal. (Plus, as cool new research shows, most people don’t change their minds much after they reach adulthood, so any effects of life experience on attitudes are swimming upstream to be observable at scale.)

3. It’s global now, too

Local experiences don’t translate directly to local attitudes and behavior because we share culture instantly around the world. So, 9/11 happened in the US but everyone knew about it (and there was also March 11 in Spain, and 7/7 in London). There are unique things about them that some people experienced — like having schools closed if you were a kid living in New York — but also general things that affected large swaths of the world, like heightened airline security. The idea of a uniquely affected age group is implausible.

4. Reflexivity

Once word gets out (through research or other means) about a particular trait or practice associated with a “generation,” like avocado toast or student debt, it gets processed and reprocessed reflexively by people who don’t, or do, want to embody a stereotype or trend for their supposed group. This includes identifying with the group itself — some people avoid it and some people embrace it, and some people react to who does the other things in other ways — until the category falls irretrievably into a vortex of cultural pastiche. The discussion of the categories, in other words, probably undermines the categories as much as it reinforces them.

If all this is true, then insisting on using stable, labeled, “generations” just boxes people into useless fixed categories. As the open letter puts it:

Predetermined cohort categories also impede scientific discovery by artificially imposing categories used in research rather than encouraging researchers to make well justified decisions for data analysis and description. We don’t want to discourage cohort and life course thinking, we want to improve it.

Mapping social change

OK, here’s today’s data exercise. There is some technical statistical content here not described in the most friendly way, I’m sorry to say. The Stata code for what follows is here, and the GSS 1972-2018 Cross-Sectional Cumulative Data file is free, here (Stata version); help yourself.

This is just me pushing at my assumptions and supplementing my reading with some tactile data machinations to help it sink in. Following on the previous exercise, here I’ll try out an empirical method for identifying meaningful birth year groupings using attitude questions from the General Social Survey, and then see if they tell us anything, relative to “empty” categories (single years or decades) and the Pew “generations” scheme (Silent, Baby Boom, Generation X, Millennials, Generation Z).

I start with five things that are different about the cohorts of nowadays versus those of the olden days in the United States. These are things that often figure in conversations about generational change. For each of these items I use one or more questions to create a single variable with a mean of 0 and a standard deviation of 1; in each case a higher score is the more liberal or newfangled view. As we’ll see, all of these moved from lower to higher scores as you look at more recent cohorts.

• Liberal spending: Believing “we’re spending too little money on…” seven things: welfare, the environment, health, big cities, drug addiction, education, and improving the conditions of black people. (For this scale, the measure of reliability [alpha] is .66, which is pretty good.)
• Gender attitudes: Four questions on whether women are “suited for politics,” working mothers are bad for children, and breadwinner-homemaker roles are good. High scores mean more feminist (alpha = .70).
• Confidence in institutions: Seven questions on organized religion, the Supreme Court, the military, major companies, Congress, the scientific community, and medicine. High scores mean less confidence (alpha = .68).
• General political views from extremely conservative to extremely liberal (one question)
• Never-none: People who never attend religious services and have no religious affiliation (together now up to about 16% of people).

These variables span the survey years 1977 to 2018, with respondents born from 1910 to 1999 (I dropped a few born in 2000, who were just 18 years old in 2018, and those born before 1910). Because not all questions were asked of all the respondents in every year I lost a lot of people, and I had to make some hard choices about what to include. The sample that answered all these questions is about 5,500 people (down from almost 62,000 altogether — ouch!). Still, what I do next seems to work anyway.

Clustering generations

Once I have these five items, I combine them into a megascale (alpha = .45) which I use to represent social change. You can see in the figure that successive cohorts of respondents are moving up this scale, on average. Note that these cohorts are interviewed at different points in time; for example, a 40-year-old in 1992 is in the same cohort as a 50-year-old in 2002, while the 1977 interviews cover people born all the way back to 1910. That’s how I get so many cohorts out of interviews from just 1977 to 2018 (and why the confidence intervals get bigger for recent cohorts).

The question from this figure is whether the cohort attitude trend would be well served by some strategic cutpoints to denote cohorts (“generations” not in the reproductive sense but in the sense of people born around the same time). Treating each birth year as separate is unwieldy, and the samples are small. We could just use decades of birth, or Pew’s arbitrary “generations.” Or make up new ones, which is what I’m testing out.

So I hit on a simple way to identify cutpoints using an exploratory technique known as k means clustering. This is a simple (with computers) way to identify the most logical groups of people in a dataset. In this case I used two variables: the megascale and birth year. Stata’s k means clustering algorithm then tries to find a set of groups of cases such that the differences within them (how far each case is from the means of the two variables within the group) are as small as possible. (You tell it k, the number of groups you want.) Because cohort is a continuous variable, and megascale rises over time, the algorithm happily puts people in clusters that don’t have overlapping birth years, so I get nicely ordered cohorts. I guess for a U-shaped time pattern it would put young and old people in the same groups, which would mess this up, but that’s not the case with this pattern.

I tested 5, 6, and 7 groups, thinking more or fewer than that would not be worth it. It turns out 6 groups had the best explanatory power, so I used those. Then I did five linear regressions with the megascale as the dependent variable, a handful of control variables (age, sex, race, region, and education), and different cohort indicators. My basic check of fit is the adjusted R², or the amount of variance explained adjusted for the number of variables. Here’s how the models did, in order from worst to best:

Cohort variable(s)	Adjusted R2
Pew generations	.1393
One linear cohort variable	.1400
My cluster categories	.1423
Decades of birth	.1424
Each year individually	.1486

Each year is good for explaining variance, but too cumbersome, and the Pew “generations” were the worst (not surprising, since they weren’t concocted to answer this question — or any other question). My cluster categories were better than just entering birth cohort as a single continuous variable, and almost as good as plain decades of birth. My scheme is only six categories, which is more convenient than nine decades, so I prefer it in this case. Note I am not naming them, just reporting the birth-year clusters: 1910-1924, 1925-1937, 1938-1949, 1950-1960, 1961-1974, and 1975-1999. These are temporary and exploratory — if you used different variables you’d get different cohorts.

Here’s what they look like with my social change indicators:

Shown this way, you can see the different pace and timing of change for the different indicators — for example, gender attitudes changed most dramatically for cohorts born before 1950, the falling confidence in institutions was over by the end of the 1950s cohort, and the most recent cohort shows the greatest spike in religious never-nones. Social change is fascinating, complex, and uneven!

You can also see that the cuts I’m using here look nothing like Pew’s, which, for example, pool the Baby Boomers from birth years 1946-1964, and Millennials from 1980 to 1996. And they don’t fit some stereotypes you hear. For example, the group with the least confidence in major institutions is those born in the 1950s (a slice of Baby Boomers), not Millennials. Try to square these results with the ridiculousness that Chuck Todd recently offered up:

So the promise of American progress is something Millennials have heard a lot about, but they haven’t always experienced it personally. … And in turn they have lost confidence in institutions. There have been plenty of scandals that have cost trust in religious institutions, the military law enforcement, political parties, the banking system, all of it, trust eroded.

You could delve into the causes of trust erosion (I wrote a paper on confidence in science alone), but attributing a global decline in trust to a group called “Millennials,” one whose boundaries were declared arbitrarily, without empirical foundation, for a completely unrelated purpose, is uninformative at best. Worse, it promotes uncritical, determinist thinking, and — if it gets popular enough — encourages researchers to use the same meaningless categories to try to get in line with the pop culture pronouncements. You get lots of people using unscrutinized categories, compounding their errors. Social scientists have to do better, by showing how cohorts and life course events really are an important way to view and comprehend social change, rather than a shallow exercise in stereotyping.

Conclusion

The categories I came up with here, for which there is some (albeit slim) empirical justification, may or may not be useful. But it’s also clear from looking at the figures here, and the regression results, that there is no singularly apparent way to break down birth cohorts to understand these trends. In fact, a simple linear variable for year of birth does pretty well. These are sweeping social changes moving through a vast, interconnected population over a long time. Each birth cohort is riven with major disparities, along the stratifying lines of race/ethnicity, gender, and social class, as well as many others. There may be times when breaking people down into birth cohorts helps understand and explain these patterns, but I’m pretty sure we’re never going to find a single scheme that works best for different situations and trends. The best practice is probably to look at the trend in as much detail as possible, to check for obvious discontinuities, and then, if no breaks are apparent, use an “empty” category set, such as decades of birth, at least to start.

It will take a collective act of will be researchers. teachers, journalists, and others, to break our social change trend industry of its “generations” habit. If you’re a social scientist, I hope you’ll help by signing the letter. (I’m also happy to support other efforts besides this experts letter.)

---

Note on causes

Although I am talking about cohorts, and using regression models where cohort indicators are independent variables, I’m not assessing cohort effects in the sense of causality, but rather common experiences that might appear as patterns in the data. We often experience events through a cohort lens even if they are caused by our aging, or historical factors that affect everyone. How to distinguish such age, period, or cohort effects in social change is an ongoing subject of tricky research (see this from Morgan and Lee for a recent take using the GSS) , but it’s not required to address the Pew “generations” question: are there meaningful cohorts that experience events in a discernibly collective way, making them useful groups for social analysis.

Very important social scientist says — wait, Brad Wilcox did what now?

Last year Brad Wilcox gave a talk at Saint Vincent College, a monastic college just 8 minutes from the glass-lined tanks of old Latrobe. The talk was structured as a critique of Stephanie Coontz and her decadent view that marriage should be freely chosen.

I was skipping through it looking for his current definition of the “Gold Standard” family (“intact, married, biological family”), when I found this amazing piece of sophistry masquerading as incompetence (or maybe the other way around, hard to know). To lighten things up around here, I thought I’d share it.

This is the clip (It’s here on the full video), followed by the transcript:

Stephanie Coontz, who is a very prominent, progressive author, writer and scholar, has argued that ‘marriage has become more joyful, more loving, more satisfying.’ So what’s the evidence tell us about what actually happened? Well, looking just at those folks who were married, from the 70s to the 80s, and the 90s, and actually the 2000s [he’s looking up at his own graph here], there’s no evidence that marriage is becoming more joyful or more loving among those folks who managed both to get married, and in this case, to remain married. So we see a decline. It’s actually particularly precipitous for women, from the 70s to the 2000s. And when you factor in the fact that fewer and fewer folks were getting married, that’s particularly interesting. What’s this is telling us is that a smaller and smaller share of American adults both men and women were in a very happy marriage from the 70s to the 2000s. So, I don’t think there’s a lot of evidence to sort of back up her idea that sort this newer, post-70s model, this more contingent model, was more successful.

Here’s a screengrab of the slide he’s looking at:

Wow, that sure makes it seem like the percent of adults in a very happy marriage is a lot lower than it used to be, although rebounding strongly this decade. It also just doesn’t look anything like the many graphs of marital happiness I have seen, and made, from the General Social Survey, which is the data he’s using here. So, how exactly is Brad Wilcox completely wrong this time?

Two amazingly incompetent and/or disingenuous ways. First, he included people who aren’t married in his denominator, so the decline in “happy marriage” he shows is due to the decline in marriage altogether, not to any change within marriage. That’s like saying the percentage of people who love reading handwritten parchment scrolls has plummeted since the Middle Ages.

Second, that huge drop in the 2000s he shows is entirely due to the fact that most people weren’t asked the question in survey years 2002, 2004, and 2006, and he included them as not in “very happy” marriages. Brad could just as accurately have said the 19,000 happily married people in the GSS are the only happily married people on earth.

Here is the actual trend, with what he showed (which is trivially easy to reconstruct), showing the individual years instead of the decade grouping he used:

I can’t think of something better to say than just showing this. (I didn’t bother with the gender breakdown, which isn’t the point.)

I hope this helps. Please stay at home if you can, and support those who can’t stay home as best you can.

---

Update: Stephanie Coontz writes on Facebook:

Thanks to Philip Cohen for showing how Brad Wilcox misuses data to criticize my claim that modern marriages are “more loving, more satisfying” than in the past. What I actually say is that when they work WELL, modern marriages are more intimate, fair & loving than marriages of the past. When they DON’T work well, they are more disappointing and seem less bearable because of our higher expectations. This makes our current crisis especially challenging, because we have to expect more of our partners, and can expect less from others, than usual.

The continuation of babies

There is no guarantee that a happy, healthy, equal, and harmonious population wants to produce enough children to maintain or grow its total size.

Anna Louie Sussman wrote an essay in the New York Times, given the unfortunate title “The End of Babies” (about which more below). I like a lot of it, and I have substantial disagreements with the framing.

It’s about falling fertility and capitalism. This is a great summary, though I would replace “not necessarily a bad thing” with “usually a very good thing”:

Declining fertility typically accompanies the spread of economic development, and it is not necessarily a bad thing. At its best, it reflects better educational and career opportunities for women, increasing acceptance of the choice to be child-free, and rising standards of living.

At its worst, though, it reflects a profound failure: of employers and governments to make parenting and work compatible; of our collective ability to solve the climate crisis so that children seem a rational prospect; of our increasingly unequal global economy. In these instances, having fewer children is less a choice than the poignant consequence of a set of unsavory circumstances.

Sussman sees the “bigger picture” as this:

Our current version of global capitalism … has generated shocking wealth for some, and precarity for many more. These economic conditions generate social conditions inimical to starting families: Our workweeks are longer and our wages lower, leaving us less time and money to meet, court and fall in love. Our increasingly winner-take-all economies require that children get intensive parenting and costly educations, creating rising anxiety around what sort of life a would-be parent might provide. A lifetime of messaging directs us toward other pursuits instead: education, work, travel.

This paragraph uses a sort of 1% versus 99% framing with is exaggerated but not unreasonable. This, however, is just exaggerated:

It seems clear that what we have come to think of as “late capitalism” — that is, not just the economic system, but all its attendant inequalities, indignities, opportunities and absurdities — has become hostile to reproduction. Around the world, economic, social and environmental conditions function as a diffuse, barely perceptible contraceptive.

Lost in this, by now, is all the good parts about falling fertility mentioned previously. Remember, contraceptives are good, and most people use them deliberately to help control their lives, and they do it because social and environmental conditions have made it possible to have more control over one’s life than ever before, while offering unprecedented opportunities for women beyond child-rearing.

In short, I agree with Sussman’s description of how some people in rich societies would like to have more children than they do, I just don’t think it’s anything like a universal or even general experience in our era. And there is a puzzle confounding the premise: within rich countries, or at least the USA, privileged people, who presumably have more control over their lives and destinies, still have fewer children than those who are more powerless. I once wrote:

There is an argument that Americans are having fewer children than they want to because of our stone age work-family policies, especially poor family leave support and the high costs of good childcare. I’m sure that’s happening to some degree, but it’s still the case that more privileged people, who should be able to overcome those things more readily — people with college degrees and Whites — have lower fertility rates than people who are getting squeezed more.

Like a lot of work in this area, Sussman’s assessment that people want more children — which generates the image of the “reproductive malaise [that] has settled over,” in this case, Denmark — is based on surveys showing people’s “ideal” family size is larger than the average number of children actually born per family. But the interpretation of this gap is not so straightforward. Maybe people think three is the ideal number of children, but they also think a PhD is the ideal amount of education, and so they compromise, with some having one kid and a PhD, and some having three kids and a no college degree. This is an empirical question. What’s historically unprecedented and still so new that we don’t know what to make of it socially is the fact that this is a choice at all for so many people.

As I previously reported, the proportion of US women whose “ideal” number of children is higher than they number they had by age 40 has risen, from less than 15% among women born in the 1930s to almost a quarter for women born in the early 1970s. If you break that trend down by BA/no-BA education level, you can see that women with BA degrees are pushing it upward:

So maybe college graduate women are having fewer than their ideal number of children like I’m earning less than the ideal amount of money — I think I could be making more money, but then I wouldn’t be able to sit around in my pajamas blogging with my dog, so I compromise. Of course, like some of the people in Sussman’s piece, a lot of people are justifiably unhappy about this, feeling they can’t compromise between forces pulling them in opposite directions. And so the result is dissatisfaction, maybe even malaise. I just don’t think we know how many people feel that way, or even whether the feeling is much more prevalent than it used to be.

Denmark

Sussman uses Denmark as one case study. This is her summary:

If any country should be stocked with babies, it is Denmark. The country is one of the wealthiest in Europe. New parents enjoy 12 months’ paid family leave and highly subsidized day care. Women under 40 can get state-funded in vitro fertilization. But Denmark’s fertility rate, at 1.7 births per woman, is roughly on par with that of the United States. A reproductive malaise has settled over this otherwise happy land.

But where is the evidence for this malaise? Denmark’s fertility rate has been low and relatively stable, while it is the USA’s that has plummeted since 2007, which is why the countries are now at the same level. The malaise that is settling is here — Denmark’s has already settled.

To elaborate on Denmark: There was a rapid drop in fertility from the mid-1960s through the mid-1980s, followed by a rebound, and then relative stability for about 25 years. During that time, as the population continued to grow slowly, women were reaching age 40 with between 1.8 and 1.9 children on average. Rather than slipping into a chasm, it looks more like the affluent people of Denmark have settled into a moderately low-fertility regime.

“Replacement” fertility, of about 2.1 births per woman, doesn’t mean a society is healthy or happy. Maybe late capitalism with a decent welfare state is not “hostile to reproduction,” maybe it just doesn’t quite get to 2.1.

How bad is that? Like the USA (see my last projections), Denmark will have population decline if they keep on this path, discounting immigration. Because they have been at low fertility for a while, the country is close to seeing actual decline based on birth rates alone. Here is what would happen over the next hundred years if current trends persist and there are no immigrants: The population would eventually contract 31%.

A 31% population drop a century from now would make for a pretty different Denmark (as will another few feet of sea-level rise). But there is time to get there — the drop would only be 6% in the next three decades. And of course if they don’t want this, they could easily cushion the fall with immigration. In any event, there is nothing here that suggests the “end of babies” or the abandonment of reproduction — families would continue having an average of 1.8 children each, as they have for the last several decades.

A population below replacement fertility might seem diseased, but it might also just be the aggregation of a lot of people exercising their newfound freedoms in newly discovered ways, including having fewer or no children. I agree with Sussman when she writes:

The problem, to be clear, is not really one of “population” …. Hundreds of thousands of babies are born on this planet every day; people all over the world have shown they are willing to migrate to wealthier countries for jobs. Rather, the problem is the quiet human tragedies, born of preventable constraints — an employer’s indifference, a belated realization, a poisoned body — that make the wanted child impossible.

To the extent those tragedies occur, we should prevent or ameliorate them. And to the extent they are concentrated among people or groups who already experience marginalization, isolation, or exploitation, it’s a social problem that’s part of our burgeoning inequality suite. Healthcare, housing, education, and family leave all come to mind as helpful, even if they can’t solve the existential crisis of late capitalism. But two cautions. First, I’m not convinced such tragedies are more common than they used to be, just because people are having fewer children than they used to. Remember, we also have fewer people trapped into having large families they don’t want (forced-birther policies notwithstanding).

And second, crucially, even if we address these issues of self-determination, there is no guarantee that a happy, healthy, equal, and harmonious population wants to produce enough children to maintain or grow its total size. We may eventually have to learn to live with fewer people, locally and globally, even if we’re all happy with the number of children we have.

What comes around

In the meantime, I think it’s confusing and ultimately unhelpful to confound what are essentially orthogonal issues. We should care about the problems Sussman raises regardless of population trends.

And that brings me to an aside on New York Times coverage. It was just 11 years ago, in 2008, that a different New York Times story about the existential threat of falling fertility, this one in the Magazine and titled “No Babies?”, singled out Scandinavian countries — with total fertility rates of 1.8 — as positive examples, bucking the trend toward “lowest-low” fertility demonstrated by Southern European countries, due to their “vigorous social-welfare systems.” That’s the same social welfare system, and the same total fertility rate, that Sussman characterizes as a “reproductive malaise” in Denmark today.

And there are illustrations of children playing alone in both cases. Because “the end of babies” and a world with “no babies” is best illustrated with a picture of the last child on earth alone in a playground. Great illustrations — just not of our societies.

That said…

You can’t really pin sudden fertility swings on things like “late capitalism,” which are decades in the making. It was just February of 2009 that I was writing one of my first blog posts, “Why Are American Women Having More Children?” as U.S. total fertility rose to 2.1 for the first time since 1971. I think it was late capitalism, too, but the U.S. TFR was 13% higher than Denmark’s (they are now the same), and everyone was talking about American mothers “opting out” of the labor force to stay home with their four children. On the other hand is socialist Finland — a country with a lot of what I want from social policy, including low inequality and poverty, and lots of family leave — which has seen a fertility decline since 2010 that could reasonably be called a crash. The government estimates the TFR in 2019 is 1.32-1.34, down from 1.86 a decade ago!

Here are the trends in select countries:

Does this mean the people in Finland are suddenly much less happy  relative to those in Denmark, which has seen a recent uptick in fertility rates? I have no idea. I can make population projections if you tell me the fertility rate, but I can’t tell you what the fertility rate will be in the future (and neither can you). A tiny bit humbling, honestly.

Wilcox and colleagues plagiarized my work in the New York Times

In the New York Times yesterday, W. Bradford Wilcox, Jason S. Carroll and Laurie DeRose published an Op-Ed with the ridiculous title, “Religious Men Can Be Devoted Dads, Too.” In it they included this figure:

In 2015 I wrote a post titled, “That thing about Republican marriages being happier (isn’t true),” which included this figure:

There are trivial differences between these figures. Theirs is from the General Social Survey for 2010-2018, mine was for 2010-2014. Theirs used political views while mine used party identification. Theirs is just women, and controls for age, education, and race; mine included men and women while controlling for gender, and I also controlled for income and religious attendance. (And they used gray for the middle bar, instead of purple.) However, in a subsequent post, from 2017, I redid the analysis for the years 2012-2016, using political views instead of party identification, in a post titled, “Who’s happy in marriage? (Not just rich, White, religious men, but kind of).” The results are almost identical to theirs in the Times (on the right, here):

Did they know about my pieces? I am certain they did, though I can’t prove it. It’s relevant that my first post, “That thing about Republican marriages…” was a critique of a post by Wilcox and Nick Wolfinger, which had only reported that Republicans were slightly happier in marriage than Democrats, which they called “The Republican Advantage in Marital Satisfaction.” My post was a correction, showing the U-shape the emerged when you broke out the categories — the change Wilcox and colleagues have now adopted. My follow-up post was reported by Bloomberg (and carried in the Chicago Tribune), and the Daily Mail. Both of my posts were tweeted by popular journalists who work in this area. I expect that would claim they never noticed my little blog posts.

You also could split hairs on the definition of plagiarism to try to defend this unethical behavior. The relevant passages of the American Sociological Association Code of Ethics:

(b) In their publications, presentations, teaching, practice, and service, sociologists provide acknowledgment of and reference to the use of their own and others’ work, even if the work is paraphrased and not quoted verbatim.
(c) While sociologists utilize and build on the concepts, theories, and paradigms of others, they may not claim credit for creating such ideas and must cite the creator of such ideas where appropriate.

But no one can seriously argue they shouldn’t have referenced my work.

Wilcox has done much worse, of course, most importantly leading a conspiracy to gin up research to turn the Supreme Court against same-sex marriage and then lying about his role in that conspiracy (the subject of a chapter in my book Enduring Bonds). And this is not a very important idea (their explanation is very flimsy, and I have no real explanation or theory to explain the pattern.) But this one goes on the list somewhere.

Why?

Why do I care? Is this just petty partisanship and even jealousy because Wilcox paid himself $80,000 of right-wing foundation money in 2016, and continues to publish low-quality research in important outlets like the New York Times? Draw your own conclusions. Of course his views are noxious to me. But more than that, in the game of trust that is the research ecosystem, reputations matter a lot. Once someone is tenured, and funded by unaccountable political actors, our options for defending the system are limited. The norms of publishing, especially outside academia, don’t require research transparency (like their current report, made to order for conservative funders, not the research community or peer review). If someone says, “This is my finding,” publishers (like the Times) usually vet the researcher instead of the research. 

I don’t believe in lifetime bans, and I don’t care about atonement for research ethics. My question is, “Can we trust this person’s research?” Before we can answer that affirmatively, we need to have an accounting of past malfeasance that makes clear future work will be clean. Until then, I don’t mind spending a few minutes now and then reminding people that Wilcox (like Mark Regnerus) is not trustworthy.

Sibship size and educational attainment update

I just read this Demography paper by Benjamin G. Gibbs, Joseph Workman, and Douglas B. Downey for my work on the new edition of The Family (don’t hold your breath, but I’m working on it). It seeks to modify the traditional “resource dilution” model of explaining why children with more siblings end up with lower educational attainment, arguing that the effect depends on who else is sharing in the child rearing. Worth a read.

Anyway, since their 2016 paper only used the General Social Survey through 2010, I figured I could make a quick update through 2018. So this is based on their model, but a lot simpler — I didn’t impute missing values, and I didn’t include all the background variables (most notably parents’ education), or get into the religious context, which is the whole point of their analysis.

So, this is my result, showing proportion of people age 25 and older who have a BA or higher, by the number of siblings they have and the decade of their birth. These are average marginal effects from a model that includes race and age, as well as region, family structure (married parents), and religion at age 16.

I find no difference between 0 and 1 sibling, a little lower odds for those with 2 siblings, and then the proportion with a BA or higher drops off. The sibling effect has decreased, at least proportionally — for example, the 3-sibs to 1-sib ratio fell from 1.9-to-1 for those born in the 1910s to 1.4-to-1 for the 1940s to 1960s cohorts and even lower for the 1970s. However, the gaps crept back up in the years after Gibbs’ et al’s analysis ended, back to 1.5-to-1 for the 1990s cohort.

No doubt this is relevant to the decisions people make about how many children to have, discussed in yesterday’s post on fertility trends, and in another Atlantic piece today.

Some people who say getting married is good for kids so you should do it are also more-kids proponents. They should make their priorities clear.

I put the Stata code for this on the Open Science Framework, here.

Fertility rate implications explained

(Sorry for the over-promising title; thanks for the clicks.)

First where we are, then projections, with figures.

For background: Caroline Hartnett has an essay putting the numbers in context. Leslie Root has a recent piece explaining how these numbers are deployed by white supremacists (key point: over-hyping the downside of lower fertility rates has terrible real-world implications).

Description

The National Center for Health Statistics released the 2018 fertility numbers yesterday, showing another drop in birth rates, and the lowest fertility since the Baby Boom. We are continuing a historical process of moving births from younger to older ages, which shows up as fewer births in the transition years. I illustrate this each year by updating this figure, showing the relative change in birth rates by age since 1989:

Historically, postponement was associated with reduction in lifetime births — which is what really matters for population trends. When people were having lots of children, any delay reduced the total number. With birth rates around two per woman, however, there is a lot more room for postponement — a lot of time to get to two. (At the societal level, both reduction and postponement are generally good for gender equality, if women have good health and healthcare.)

This means that drops in what we demographers call “period” fertility (births right now) are not the same as drops in “completed” fertility (births in a lifetime), or falling population in the long run. The period fertility measure most often used, the unfortunately named total fertility rate (TFR), is often misunderstood as an indicator of how many children women will have. It is actually how many births they are having right now, expressed in lifetime terms (I describe it in this video, with instructions).

Lawrence Wu and Nicholas Mark recently showed that despite several periods of below “replacement” fertility (in terms of TFR), no U.S. cohort of women has yet finished their childbearing years with fewer than two births per woman. Here is the completed fertility of U.S. women, by year of birth, as recorded by the General Social Survey. By this account, women born in the early 1970s (now in their late-forties by 2018) have had an average of 2.3 children.

Whether our streak of over-two completed fertility persists depends on what happens in in the next few years (and of course on immigration, which I’ll get to).

Last year at this time I summed up the fertility situation and concluded, “sell stock now,” because birth rates fell for women at all ages except over 40. That kind of postponement, I figured, based on history, reflected economic uncertainty and thus was an ill omen for the economy. The S&P 500 is up 5% since then, which isn’t bad as far as my advice goes. And I’m still bearish based on these birth trends (I bet I’ll be right before fertility increases).

Projection

It is very hard to have an intuitive sense of what demographic indicators mean, especially for the future. So I’ve made some projections to show the math of the situation, to get the various factors into scale. My point is to show what the current (or future) birth rates imply about future growth, and the relative role of immigration.

These projections run from 2016 to 2100. I made them using the Census Bureau’s Demographic Analysis and Population Projection System software, which lets me set the birth, death, and migration rates.* I started with the 2016 population because that’s the most recent set of life tables NCHS has released for mortality. Starting in 2018 I apply the current age-specific birth rates.

First, the most basic projection. This is what would happen if birth rates stayed the same as those in 2018 and we completely cut off all immigration (Projection A), or if we had net migration running at the current level of just under +1 million each year, using Census estimates for age and sex of the migrants (Projection B).

From the 2016 population of 323 million, if the birth rates by age in 2018 were locked in, the population would peak at 329 million in 2029 and then start to decline, reaching 235 million by 2100. However, if we maintain current immigration levels (by age and sex), the population would keep growing till 2066 before tapering only slightly. (Note this assumes, unrealistically, that the immigrants and their children have the same birth rates as the current population; they have generally been higher.) This the most important bottom line: there is no reason for the U.S. to experience population decline, with even moderate levels of immigration, and assuming no rebound in fertility rates. Immigration rates do not have to increase to maintain the current population indefinitely.

Note I also added the percentage of the population over age 65 on the figure. That number is about 16% now. If we cut off immigration and maintain current birth rates, it would rise to 25% by the end of the century, increasing the need for investment in old age stuff. If we allow current migration to continue, that growth is less and it only reaches 23%. This is going up no matter what.

To show the scale of other changes that we might expect — again, not predictions — I added a few other factors. Here are the same projections, but adding a transition to higher life expectancies by 2080 (using Japan’s current life tables; we can dream). In these scenarios, population decline is later and slower (and not just at older ages, since Japan also has lower child mortality).

Under these scenarios, with rising life expectancies, the old population rises more, to between 27% and 29%. Generally experts assume life expectancies will rise more than this, but that’s the assumed direction (now, unbelievably, in doubt).

Finally, I’ve been assuming birth rates will not fall further. If what we’re seeing now is fertility postponement, we wouldn’t expect much more decline. But what if fertility keeps falling? Here is what you get with the assumptions in Projection D, plus total fertility rates falling to 1.6, either by 2030 or 2050. As you can see, in the 1.6 to 1.8 range, the effects on population size aren’t great in this time scale.

Conclusion: We are on track for slowing population growth, followed by a plateau or modest decline, with population aging, by the end of the century, and immigration is a bigger question than fertility rates, for both population growth and aging.

Perspective

In a global context where more people want to come here than want to leave (to date), worrying about low birth rates tends to lend itself to myopic, religious, or racist perspectives which I don’t share. I don’t think American culture is superior, whites are in danger of extinction, or God wants us to have more children.

I do not agree with Dowell Myers, who was quoted yesterday as saying, “The birthrate is a barometer of despair.” That even as some people are having fewer children than they want, or delaying childbearing when they would rather not. In the most recent cohort to finish childbearing, 23% gave an “ideal number of children for a family to have” that was greater than the number they had, and that number has trended up, as you can see here:

Is this rising despair? As individuals, people don’t need to have children any more. Ideally, they have as many as they want, when they want, but they are expensive and time consuming and it’s not surprising people end up with fewer than they think “ideal.” Not to be crass about it, but I assume the average person also has fewer boats than they consider ideal.

And how do we know what is the right level of fertility for the population? As Marina Adshade said on Twitter, “Did women actually have a desire for more children in the past? Or did they simply lack the bargaining power and means to avoid births?”

However, to the extent that low birth rates reflect frustrated dreams, or fear and uncertainty, or insufficient support for families with children, of course those are real problems. But then let’s name those problems and address them, rather than trying to change fertility rates or grow the population, which is a policy agenda with a very bad track record.

---

* I put the DAPPS file package I created on the Open Science Framework, here. If you install DAPPS you can open this and look at the projections output, with graphs and tables and population pyramids.

Do rich people like bad data tweets about poor people? (Bins, slopes, and graphs edition)

Almost 2,000 people retweeted this from Brad Wilcox the other day.

Brad shared the graph from Charles Lehman (who noticed later that he had mislabeled the x-axis, but that’s not the point). First, as far as I can tell the values are wrong. I don’t know how they did it, but when I look at the 2016-2018 General Social Survey, I get 4.3 average hours of TV for people in the poorest families, and 1.9 hours for the richest. They report higher highs (looks like 5.3) and lower lows (looks like 1.5). More seriously, I have to object to drawing what purports to be a regression line as if those are evenly-spaced income categories, which makes it look much more linear than it is.

I fixed those errors — the correct values, and the correct spacing on the x-axis — then added some confidence intervals, and what I get is probably not worth thousands of self-congratulatory woots, although of course rich people do watch less TV. Here is my figure, with their line (drawn in by hand) for comparison:

Charles and Brad’s post got a lot of love from conservatives, I believe, because it confirmed their assumptions about self-destructive behavior among poor people. That is, here is more evidence that poor people have bad habits and it’s just dragging them down. But there are reasons this particular graph worked so well. First, the steep slope, which partly results from getting the data wrong. And second, the tight fit of the regression line. That’s why Brad said, “Whoa.” So, good tweet — bad science. (Surprise.) Here are some critiques.

First, this is the wrong survey to use. Since 1975, GSS has been asking people, “On the average day, about how many hours do you personally watch television?” It’s great to have a continuous series on this, but it’s not a good way to measure time use because people are bad at estimating these things. Also, GSS is not a great survey for measuring income. And it’s a pretty small sample. So if those are the two variables you’re interested in, you should use the American Time Use Survey (available from IPUMS), in which respondents are drawn from the much larger Current Population Survey samples, and asked to fill out a time diary. On the other hand, GSS would be good for analyzing, for example, whether people who believe the Bible is the “the actual word of God and is to be taken literally, word for word” watch TV more than those who believe it is “an ancient book of fables, legends, history, and moral precepts recorded by men” (Yes, they do, about an hour more.) Or looking at all the other social variables GSS is good for.

On the substantive issue, Gray Kimbrough pointed out that the connection between family income and TV time may be spurious, and is certainly confounded with hours spent at work. When I made a simple regression model of TV time with family income, hours worked, age, sex, race/ethnicity, education, and marital status (which again, should be done better with ATUS), I did find that both hours worked and family income had big effects. Here they are from that model, as predicted values using average marginal effects.

The banal observation that people who spend more time working spend less time watching TV probably wouldn’t carry the punch. Anyway, neither resolves the question of cause and effect.

Fits and slopes

On the issue of the presentation of slopes, there’s a good lesson here. Data presentation involves trading detail for clarity. And statistics have both have a descriptive and analytical purpose. Sometimes we use statistics to present information in simplified form, which allows better comprehension. We also use statistics to discover relationships we couldn’t otherwise — such as multivariate relationships that you can’t discern visually. The analyst and communicator has to choose wisely what to present. A good propagandist knows what to manipulate for political effect (a bad one just tweets out crap until they get lucky).

Here’s a much less click-worthy presentation of the relationship between family income and TV time. Here I truncate the y-axis at 12 hours (cutting off 1% of the sample), translate the binned income categories into dollar values at the middle of each category, and then jitter the scatterplot so you can see how many points are piled up in each spot. The fitted line is Stata’s median spline, with 9 bands specified (so it’s the median hours at the median income in 9 locations on the x-axis). I guess this means that, at the median, rich people in America watch about an hour of TV per day less than poor people, and the action is mostly under $50,000 per year. Woot.

Finally, a word about binning and the presentation of data (something I’ve written about before, here and here). We make continuous data into categories all the time, starting from measurement. We usually measure age in years, for example, although we could measure it in seconds or decades. Then we use statistics to simplify information further, for example by reporting averages. In the visual presentation of data, there is a particular problem with using averages or data bins to show relationships — you can show slopes that way nicely, but you run the risk of making relationships look more closely correlated than they are. This happens in the public presentation of data when analysts are showing something of their work product — such as a scatterplot with a fitted line — to demonstrate the veracity of their findings. When they bin the data first, this can be very misleading.

Here’s an example. I took about 1000 men from the GSS, and compared their age and income. Between the ages of 25 and 59, older men have higher average incomes, but the fit is curved with a peak around 45. Here is the relationship, again using jittering to show all the individuals, with a linear regression line. The correlation is .23

That might be nice to look at but it’s hard to see the underlying relationship. It’s hard to even see how the fitted line relates to the data. So you might reduce it by showing the average income at each age. By pulling the points together vertically into average bins, this shows the relationship much more clearly. However, it also makes the relationship look much stronger. The correlation in this figure is .65. Now the reader might think, “Whoa.”

Note this didn’t change the slope much (it still runs from about $30k to $60k), it just put all the dots closer to the line. Finally, here it is pulling the averages together in horizontal bins, grouping the ages in fives (25-29, 30-34 … 55-59). The correlation shown here is .97.

If you’re like me, this is when you figured out that reducing this to two dots would produce a correlation of 1.0 (as long as the dots aren’t exactly level).

To make good data presentation tradeoffs requires experimentation and careful exposition. And, of course, transparency. My code for this post is available on the Open Science Framework here (you gotta get the GSS data first).

The happiness scamcession

While we’re making up overwrought terms.

W. Bradford Wilcox and Lyman Stone, both fellows who are fellows at the Institute for Family Studies (still waiting for the acknowledgment for correcting their data error), have a bad new post up called “The Happiness Recession.”

I made some annotations on their essay using the excellent Hypothesis tool. Now you can read the essay with my comments here. Click on the little arrow at the top right to expand the comments. Here are the highlights, with figures.

Note: Get a free Hypothesis account and you can annotate any page on the web, including in closed groups if you want to, and you can share your Hypothesis profile so people can see all the things you’ve annotated, like mine. You can also comment on SocArXiv papers this way.

The main problem is that there is no evidence of a happiness recession. They make their case with this figure from the General Social Survey:

To people not familiar with the General Social Survey, like, apparently, whoever edited this piece at the Atlantic, that 2018 drop in happiness might look dramatic. A lot of publications are used to mentioning margins of error when they report survey results. But maybe because WilcoxStone reminded the editor in the opening paragraph that the GSS is “a key barometer of American social life,” they didn’t bother. It seems so authoritative (and it’s a great, indispensable resource, publicly funded, and freely available for all researchers). But it’s a sample survey, and it’s not that big. Here are the actual numbers of men ages 18-34 who answered the happiness question on the survey:

It’s a little funky because of the weights I’m not showing, but in round numbers, if 8 young men had said “very happy” instead of “pretty happy” in 2018, WilcoxStone wouldn’t have been able to write this article — the percentage wouldn’t have changed. So, it’s a great survey, but you have to know what you’re doing with it, and you have to be honest about it.

There are different ways to assess significance, and no one rule, but I made this figure showing the percentage of young adults describing themselves as “very happy” in each year of the survey, with a simple p<.05 significance test for whether each year is different from 2018.

The 2018 level is the lowest point estimate, but it’s not distinguishable from about half the previous survey years at conventional levels of statistical significance. Eight pretty happy men doing all the work here.

A lot of what’s ridiculous about the post follows from this simple manipulation — pretending an insignificant change is very important. One other thing was totally wrong, though.

WilcoxStone have some incoherent theory about how friendship might play a role in the happiness recession. Maybe they were expecting to see a decline in friendship, to support their get-married-in-church-and-stay-there predetermined conclusion. Anyway, they produce this wrong figure:

I say “wrong” because, although they never define “regularly,” and neither does the GSS, men in the survey spend more evenings with their friends, and nothing seems to match these numbers. Maybe you can figure it out. Here are the distributions for young men and women in 2018 — see if you can see how they could get 64% for women and 35% for men, the last points in their figure:

In response to my tweets about this, linking to the annotated post, Stone responded that they “just use a different definition of regular,” before clarifying that they “just cut it a different way,” and finally acknowledging that “maybe there was a mislabeling in that,” and promising to look into it.

Anyway, they conclude their analysis with some counterfactual models to explain the “happiness trend.” No details are provided, and definitely nothing like standard errors or significance tests (and if the friend contact variable is screwed up, it’s useless). Here are the results:

If you look carefully, you can see that the biggest difference comes from adjusting for sexual frequency — which has declined in the sample, and sexual frequency is associated with higher happiness scores. That dotted light blue line ends at 29%. And the simulations start at 28%. So, if sexual frequency were held constant, they proclaim, happiness among young adults would be a point higher. Just kidding, they don’t give the number, they just say, “If Americans still had sex like they did in 2008, or even 2012, we might be a much happier country.” OK then.

After apparently reading my annotations, Stone was effusive on Twitter, delighted that, “And I think anyone who reads his annotations will see that he basically doesn’t refute anything we say. :)”

I appreciate @familyunequal annotating our work! It’s a good form of interaction! And I think anyone who reads his annotations will see that he basically doesn’t refute anything we say. 🙂 https://t.co/UEMY2iQZPo

— Lyman Stone 石來民 (@lymanstoneky) April 6, 2019

I guess you could say:

Anyway, see if you can figure out this conclusion:

Thus, while most of the decline in happiness [there is no decline in happiness -pnc] is about declining sex, that’s not the end of the story. Declining sex is at least partly about family and religious changes that make it harder for people to achieve stable, coupled life at a young age. If we’d like more young adults to experience the joy of sex, we will have to either revive these institutions or find new ways to kindle love in the rising generation.

Your guess is as good as mine. But if it just means go to church, get married, and then have sex, and you’ll be happy, I don’t think they needed a bologna statistical analysis to reach that.

Trust

Publications like the Atlantic put out a lot of content, much of which they get for free or low cost. A lot of that cheap material comes from academics (which Wilcox sort of is). And some comes from think tank snake oil salespeople (which Wilcox definitely is), who use their right-wing, tax-subsidized foundation money to slime the public square. It might be too much to expect that publications with little budget for writers would somehow have the ability to vet statistical analyses. So they run on trust. People like Wilcox, who has a long proven record of lying about research, and Stone, prey on the credulity of the economically precarious media. People who know better should speak up and help limit the damage.

Equal-education and wife-more-education married couples don’t have sex less often

In my review of Mark Regnerus’s book, Cheap Sex, I wrote: “The book is an extended rant on the theme, ‘Why buy the cow when you can get the milk for free?’ wrapped in a misogynist theory about sexual exchange masquerading as economics, and motivated by the author’s misogynist religious and political views.”

Someone just reposted an old book-rehash essay of Regnerus’s called, “The Death of Eros.” In it he links to my post documenting the decline in sexual frequency among married couples in the General Social Survey. In marriage, Regnerus writes, “equality is the enemy of eros,” before selectively characterizing some research about the relationship between housework and sex. (Here’s a recent analysis finding egalitarian couples don’t have sex less.)

But I realized I never looked at sexual frequency in married couples by the relative education of the spouses, which is available in the GSS. So here’s a quick take: Married man-woman couples in which the wife has equal or more education don’t have sex less frequently.

I modeled sexual frequency (an interval scale from “not at all” = 0 to “4+ times per week” = 6 as a function of age, age-squared, respondent education, respondent sex, decade, and relative education (wife has lower degree, wife has same degree, wife has higher degree). The result is in this figure. Note the means are between 3 (“2-3 times per month”) and 4 (“weekly”). Stata code for GSS below.

OK, that’s it. Here’s the code (I prettied the figure a little by hand afterwards):

*keep married people
keep if marital==1

* with non-missing own and spouse education
keep if spdeg<4 & degree<4
recode age (18/29=18) (30/39=30) (40/49=40) (50/59=50) (60/109=60), gen(agecat)
recode year (1970/1979=1970) (1980/1989=1980) (1990/1999=1990) (2000/2008=2000) (2010/2016=2010), gen(decade)
gen erosdead = spdeg>degree
gen equal=spdeg==degree

gen eros=0
replace eros=1 if spdeg<degree & sex==1
replace eros=2 if spdeg==degree
replace eros=3 if spdeg>degree & sex==1

replace eros=1 if spdeg>degree & sex==2
replace eros=3 if spdeg<degree & sex==2

label define de 1 "wife less"
label define de 2 "equal", add
label define de 3 "wife more", add
label values eros de

reg sexfreq i.sex i.agecat i.decade i.degree i.eros [weight=wtssall]
reg sexfreq i.sex c.age##c.age i.degree i.eros##i.decade [weight=wtssall]
margins i.eros##i.decade
marginsplot, recast(bar) by(decade)

Note: On 25 Dec 2018 I fixed a coding error and replaced the figure; the results are the same.

New findings: The widening political divide over science

Gordon Gauchat published a paper in American Sociological Review in 2012 (paywalled), reporting that political conservatives in the United States lost confidence in the scientific community from the 1970s to 2010, while political liberals and moderates did not. As a result, a political divide opened up so that by 2010 conservatives had the lowest level of confidence in science.

I’ve written a short working paper that extends the trends through the 2016 General Social Survey. I find that political conservatives, Republicans, and Americans who attend religious services regularly, all demonstrate falling levels of confidence in the scientific community. Further, for the most recent period (2012-2016) educational attainment, for conservatives, is not associated with increased levels of confidence in science, except at the graduate degree level (and even there confidence is lower among conservatives). It’s as if conservative views on science are impervious to education.

I have only a few sentences about politics in the paper, but the gist of it is that this divide is bad, it undermines democracy, and plays into Trump’s authoritarian hands.

Here’s the figure showing the Republicans falling off the trust cliff: the trend in expressing “a great deal of confidence” in the scientific community, with controls, by party identification:

 And here’s the figure from the recent period, showing that conservatives have lower levels of confidence regardless of education level — and education has no relationship to confidence in science for them, up to the level of graduate degree at least.

I welcome your feedback.

The paper is here on SocArXiv, and the Stata code is here.