How Divided Are We, Really?

It’s not even a question, right?

Wealth inequality

Political violence

Generational disdain

Civil War raking in the box office love

All the -isms fueling culture wars

There’s no chance of common ground when everyone is constantly reminded of how polarized the country is today. Is there?

Beneath all the discord and agenda-driven flashpoints, I believe most Americans do in fact have a largely similar worldview. And I can prove it.

Building on the work of Harry Guild at BBH London, where he took a quantitative view of like-mindedness among different segments of the UK population to refute the myth of generational homogeneity, we’ve taken a deep look at the U.S. to see if the discourse about divisiveness is borne out in the data.

TL;DR – We’ve created a custom Cohesion Score to assess just how like-minded a group is. The higher a group’s Cohesion Score, the more they agree on worldview. This lets us not only identify the baseline cohesion of the U.S. – how large of a group has a cohesive set of values and beliefs – but also compare different subsets of the population against that score to see if a subset is more or less cohesive; more or less “tribal”; and can be expected to hold similar beliefs, values, and behaviors or to simply process things similarly.

---

Long version – skip ahead unless you’re down to get nerdy

The goal is to identify a quantitative measure of like-mindedness, essentially a score for how much people agree with how they (1) see the world around them and (2) react to it in how they see themselves.

We used MRI-Simmons data as the primary input for our analysis because:

1. Scale – It is the largest nationally representative panel available in the U.S.
2. Longevity – It allows us a common core of data that can be analyzed going back decades.
3. Specificity – It offers a deep library of demographic, psychographic, behavioral, and consumption data for our analysis.

Our Cohesion Score is based on finding the degree to which people agree in worldview. We wanted to understand like-mindedness in the broadest sense, so we took a wide view of the views we’d analyze for agreement.

Using the MRI-Simmons Winter 2024 U.S. dataset, we measured the size of the majority viewpoint across 412 lifestyle statements. These statements range from the mundane (“People who overanalyze are annoying”) to the profound (“Religion should be a pillar of society”) to the philosophical (“Makeup is my passion” or “I believe cryptocurrency will replace traditional currency”). We keep this very simple – for every statement, we’re looking for the total average “agree” or “disagree.”

These statements will elicit conflicting opinions in every group, but close-knit, homogenous groups (e.g., Mormons) should have larger majorities than weaker ones (e.g., left-handers); therefore, a group with more cohesion will have scores that deviate much further from the midpoint (either agree more or disagree more – we don’t care which, just the degree to which the group agrees or disagrees consistently).

The group’s Cohesion Score is simply the average deviation from the midpoint.

Further analysis lets us not just show how cohesive a group is but also quantitatively evaluate how far that group’s core beliefs and values differ from the total U.S. baseline – essentially, not just how tribal is a group, but how contrarian that group is from the norm or from another group.

---

What did we find? Are we really that divided?

Our baseline Cohesion Score for the U.S. in 2024 is 18.6.

This means that across 412 attitudinal and value statements, over two-thirds of Americans have demonstrably similar viewpoints. Two-thirds of Americans would be considered largely like-minded.

I certainly didn’t see that coming. I was expecting a value far lower, and far closer to a 50/50 split in worldview, regardless of where you make that split – politically, by age or gender, by lifestyle or sexual orientation. Given current public discourse, I expected there to be a smaller core group.

Digging deeper, are we growing closer together or drifting further apart over time? Our Cohesion Score shows that the U.S. population is actually coalescing, with cohesion growing from 2016 to 2020 (from 17.2 to 17.6) and taking a much larger leap all the way to 18.6 by 2024. We are concentrating toward a more homogenous worldview.

In the aggregate, more cohesion may be a good thing as a society. It suggests that there are common values and principles we agree on. We may differ in the application of those principles and the degree to which we hold the values dear, but the general shape of our culture is coming from a common core.

But on a smaller, more tribal scale, is greater cohesion a good thing? Does it breed increasing isolation and othering when people don’t adhere to the group’s consistent worldview? We were fascinated to find out what kinds of groups are more cohesive (in good or bad ways) and what attributes of a group most predict collective cohesion. Do stereotypes hold true? Are CrossFitters, Valley Girls, vegans, MAGA faithful, Gen Z, first-gen immigrants, or some other group the most tribal group out there? Are independent thinkers more independent, or does the strength of their independent streak actually make the collective group more homogenous?

To answer this, we look at each segment’s average Cohesion Score’s difference from the national average. Groups with more cohesion will be positive, and those with more variance in perspective among their members will show up as negative relative to the U.S. norm.

First up: politics.

As you split along party lines, you start to see a little more homogeneity increase (with Independents living up to their name by diverging even further), with perhaps a shade extra hivemind for those who identify as Republican. That like-mindedness, as you would expect, grows substantially at the ideological extremes within each group. The most pronounced is visible among people who watch Fox & Friends.

What about age? Are Millennials twinning as much as pop culture mocks?

Rather than some generations being more cohesive by default, the consistent pattern of older generations having higher group cohesion suggests that age itself appears to have a massive effect on like-mindedness. Maybe having a wider range of lived experiences creates a greater chance that people will arrive at a wider perspective, and that creates common understanding. (At least in theory – the alternative is that views calcify with age, but in the U.S., those views just happen to be consistent.) With a +2.8 Cohesion Score over the U.S. norm, that means that more than seven out of every ten members of the Boomer generation have a unified worldview.

Other demographic and lifestyle factors show no consistent pattern of creating more or less group cohesion. Having kids under 17 living at home appears to create more diverging worldviews than living a child-free lifestyle, but household members with fur appear to swing the pendulum back toward consistency.

What other tribes are stereotypically singular? CrossFitters? Not nearly as cohesive as live music lovers. Crypto bros? Far less cohesive than those who choose to eat vegan. But the more niche you go, you find much greater cohesion – vegan CrossFitters clock in with the most by far.

Same same, but different

Cohesion alone doesn’t paint the full picture. Just because two groups have similar Cohesion Scores doesn’t mean they have common views between the groups. It is entirely possible for a group to hold a similar degree of like-mindedness as another group, but elements of those values and beliefs are diametrically opposed to each other. We found a way to illustrate this through a “Spike Score” (the unfortunately sticky name that popped up as the visual below came to represent the concept – more academically, it would be the “collective indices and deviation from the norm” of the group; admittedly, “Spike Score” rolls off the tongue better).

Here is a comparison of the commonality of views between those who identify as Democrats (reminder, +0.3 Cohesion Score) against those who identify as Republicans (+1.0 Cohesion Score):

If each group matched the general population’s common viewpoint across all 412 values and beliefs, this would be a perfectly clean circle. Areas where the group average response deviates from the norm result in the radar pulling away from the circle (pulling inward or pushing out). Read more simply as basic = smooth. Those that start to stretch into a more divergent view will be more spiky. When comparing two groups as we are here, you see where those spikes move counter to one another. We learn from both the positively and the negatively correlated spikes. Each group may have some cohesion among themselves – and most importantly (at least to me) by and large share common values with the overall population – but on a wide range of topics, those beliefs move counter to one another.

The more extreme the views, the spikier the score.

People who watch Fox & Friends are more cohesive as a group than those who simply identify as Republicans, but the common views that Fox & Friends viewers share go well beyond those of the core Republican group (quantitatively, Fox & Friends viewers have a deviation from the norm +892 further than Republicans (3,525 vs. 2,633 Spike Scores) and +2,093 points beyond Democrats (1,432 Spike Score).

For a more innocuous comparison, let’s look at the divergence of views between two wholly uncorrelated groups: those who own pets (0.4 Cohesion Score, 983 Spike Score) and vegans (1.0 Cohesion Score, 5,679 Spike Score).

As you can see, we called it the Spike Score for a reason.

What does this all mean?

First and foremost, it indicates that any commentary suggesting that any group has a universally agreed-to set of beliefs about the world is inherently false. Even our most cohesive groups cap out at around 75% consistency. Fully aligned worldviews are a myth.

Equally untrue is that America is more divided than ever. A significant and growing majority still have a common view of the world, shared values (broadly defined), and common motivation. We can take heart in that.

We can also use these simple, global observations as a reminder that the more abstractly we think about people – defining them with broad labels, as groups, personas, or even simply as people who all do something together – the more wrong we’ll be about the reality of the individuals within those groups. We have to dig deep into what makes each person unique and uniquely interesting.

Most importantly: Stay curious.