Merriam-Webster’s decision to name “polarization” as 2024’s ‘word of the year’ certainly feels spot on. Of course Americans are divided into “two sharply distinct opposites,” right? and of course our beliefs are at “opposing extremes.”
But, as my therapist often reminds me, feelings aren’t facts. As a data scientist who thinks a lot about politics (and is of course extremely fun at parties), I prefer to inspect claims based on evidence. And it turns out that the quest to empirically pinpoint “polarization” in the US yields a much more complicated picture—you know, the kind of nuanced political context we are famously good at talking about.
The first thing we might ask is: who is polarized? Let’s start with our leaders. Thanks to a 1985 political science research project, we can estimate “ideology scores” for legislators based on their voting behavior. These scores can be coded in terms of how “liberal” or conservative” they are. Like all measures, the scores that result are by no means perfect, but do offer helpful clues, such as graphs like this:
In addition to resembling a concerning skin disease, this graph depicts a clear and growing ideological divide between Democrats and Republicans in the House that starts around the early 1990s. (A similar trend is apparent in the Senate.) Notably, Republicans seem to be moving more to the right than Democrats are to the left, and I look forward to all the hate mail that this sentence will generate.
What about us regular people? The evidence suggests there’s been an increase in straight-ticket voting in recent elections, compared to years past when it was more common for people to pick different parties’ candidates for different offices across their ballot. This could be evidence of growing ideological extremes, but it could also be a reflection of the more extreme candidates on the ballot.
Another place we might look is our party identification. If we have more extreme and opposing beliefs, we should expect more Americans to strongly identify as Democrats or Republicans, and perhaps fewer as Independents. Alas, if we consider data from the American National Election Survey (ANES), the longest running survey of American political beliefs, neither theory really pans out:
The most striking thing about the above graph is how boring it is. We see more Independents and slightly more Republicans, but overall, the closest hint to polarization is the proportion of respondents who identify “strongly” as Republicans, as opposed to “weakly”, or simply as “Republicans,” which has increased from 14 percent to 21 percent. This is not nothing, but it also doesn’t really feel like it captures the kind of word-of-the-year-worthy sentiment we all seem fixated on.
Maybe it’s the issues themselves? Additional data from the ANES shows that while there are some issues (notably abortion) for which we see a growing divide among Americans, there are plenty of others, including LGBTQ rights and government spending, where the gap between Democratic and Republican-identified respondents has remained almost stubbornly constant over the past few decades.
(There is even one more thing we all agree on besides polarization, however: trust in government is at an all-time low among Democrats and Republicans. Something fun we can talk about with our families over the holidays!)
It’s worth noting it was not always the case that “democrat” meant “liberal” and “republican” meant “conservative.” In fact, once upon a time in this country it was not uncommon to be a liberal republican or a conservative democrat, two identities that have all but disappeared in America these days.
This trend is at the heart of a 2004 discovery by political scientist Morris Fiorina. He observed that plenty of Americans used to hold conservative views on some issues (say, guns and immigration) but liberal views on others (say, healthcare and government spending). But more recently, he found that if we are liberal on one issue, we’re likely liberal on all of them. The same thing goes for conservatives. Our ideologies, on average, haven’t drifted further apart over time, but we are better sorted—meaning we have fewer, if any, overlapping views. In other words: we don’t actually vehemently disagree on some issues; rather, we mildly disagree on just about all of them.
This is good news and bad news. The good is that for most of us, our beliefs are not nearly as far apart from our fellow citizens as they appear. The bad is that with no overlapping views, it’s harder to talk to people outside our bubbles, let alone empathize with them.
This brings us to the most recent big finding in this area, which is that while our ideologies are not actually notably further apart these days, we do dislike each other more. This phenomenon is called "affective polarization." Research in this area shows that both sides of the political spectrum tend to think of the ‘other side’ as people who are generally bad, closed-minded, hypocritical and worse.
Scientist don’t exactly know what’s behind this rise in mutually unadultered loathing, but one theory suggests our political identities have increasingly become internalized as part of other core identities we might have and hold, as opposed to simply representing a set of preferences. The more sacred our political identities are, the more terrifying it feels when we encounter someone with a different one. Affective polarization doesn’t just affect our politics, either— a growing body of research points to all kinds of influence in economic, social, and even medical spheres.
Thus, I never thought I’d find myself disagreeing with the dictionary, but here we are: The evidence doesn’t really support “polarization” in the way that Merriam-Webster describes it. It’s not the case that we the people are at ideological extremes; it’s that we have fewer ideological points in common with one another. It also turns out that one of the biggest drivers of this mutual hatred is media coverage of polarization—the more we talk about how polarized we are, the more we tend to believe it. Political campaigns have also figured out that demonizing ‘the other side’ is a great way to activate voters, so don’t count on them to resolve things anytime soon.
So, what can we do? Well, step one is awareness, which you’ve already done by reading this article.
Step two (and this is the hard one) is to talk to people on that ‘other side.’ Before you come at me for suggesting you talk to someone who hates you, it’s also worth noting that most of us incorrectly perceive these opponents to be more extreme than they really are. Most of us are in the middle ideologically; the vast majority of us want a country that’s safer, more prosperous, and happier for ourselves and our families.
So maybe let’s talk about that—if we do it right, perhaps Merriam-Webster’s 2034 word of the year will again be “polarization”—in the sense that we prevented the polar ice caps from melting thanks to actually trying to see one another’s humanity.
Andrea Jones-Rooy is the founder of The Data Thinking Revolution, host of the podcast Behind the Data and holds a PhD in data science from the University of Michigan.
