Why is America so divided?
Sorting, polarized leaders, primaries, media and rising dislike all have evidence behind them, and all have limits. Here is what each explanation can and cannot show.
America is divided for several overlapping reasons, and researchers have not settled on one. The leading explanations are people sorting into parties that match their views, more polarized politicians, changes in media, and growing dislike of the other side. Evidence for each is real but partial, and dislike has grown more clearly than disagreement on issues.
Key points
- Partisan dislike in the US has risen sharply since the late 1970s, mostly because feelings toward the other party grew colder.
- Political scientists still debate how far ordinary voters have moved apart on issues.
- In a study of 12 wealthy democracies, the US had the largest rise in partisan dislike, and six countries saw a decline.
- Every leading explanation has evidence against it as well as for it, and none of them has been shown to be the main cause.
In this post
What is political polarization?
Political polarization is the movement of political groups away from each other, either in what they believe or in how they feel about one another. Researchers treat those as two separate things. Ideological polarization is about issues. Affective polarization is about feelings: in the definition quoted by Iyengar and colleagues (2019), it is “the tendency of people identifying as Republicans or Democrats to view opposing partisans negatively and copartisans positively.”
| Ideological polarization | Affective polarization | |
|---|---|---|
| What it asks | How far apart are the parties on issues? | How much warmer do people feel toward their own party than the other one? |
| Common measure | Policy questions in surveys; roll-call votes for Congress | A 0 to 100 “feeling thermometer” for each party |
| One finding | Pew (2014): consistently liberal or conservative views rose from 10% to 21% of Americans over two decades | American National Election Studies: the gap rose from 22.64 degrees in 1978 to 40.87 in 2016 |
| State of the evidence | Clear for Congress, disputed for ordinary voters | The rise is well documented, its causes are not settled |
we argue that affective polarization is largely distinct from the ideological divide, and that extremity in issue opinions is not a necessary condition for affective polarization
Iyengar, Lelkes, Levendusky, Malhotra and Westwood, Annual Review of Political Science, 2019
Has America actually become more divided?
On feelings, yes, by a wide margin. On issues, the answer depends on whom you measure.
Iyengar and colleagues write that the thermometer gap grew because of rising dislike of the other party, not because people came to like their own party more, and that the gap dipped between 2012 and 2016. Boxell, Gentzkow and Shapiro, calculating it their own way, put it at 27.4 points in 1978 and 56.3 in 2020. The exact size depends on the method. The direction does not.
Pew’s surveys show the same movement in both parties. In 1994, 17% of Republicans and 16% of Democrats had a very unfavorable view of the other party. By 2014 it was 43% and 38%. In 2016, 47% of Republicans and 35% of Democrats said people in the other party were more immoral than other Americans. By 2022 it was 72% and 63%.
Issues are harder to call. Pew found the middle shrinking in its 2014 report, from about 49% of the public holding a roughly equal mix of liberal and conservative positions in 1994 to 39%, while noting that the majority did not hold uniformly conservative or liberal views. Political scientists still argue over how far ordinary voters have moved. Iyengar’s team says “We do not take a position on this ongoing debate.” They do report that over 40 years the thermometer gap rose by about the same amount among people with the most ideologically consistent views and those with the least. So dislike can grow without issue positions moving, which is why the two get measured apart. The same review says sorting may lead people to see partisans as more extreme than they really are, a subject we take up in our post on the perception gap.
Why is America so polarized?
Researchers point to several causes that may feed one another, and each has evidence against it as well as for it.
People sorted into parties that match their views
Party and ideology now line up much more closely. Pew found in 2014 that 92% of Republicans sat to the right of the median Democrat, compared with 64% twenty years earlier, and 94% of Democrats sat to the left of the median Republican, up from 70%. Iyengar’s review adds that race and religion have converged with party too, and cites Lilliana Mason’s work showing that people with aligned identities grew more hostile to the other party without necessarily changing their positions. The limit: the same review reports that sorting accounts for only about 5% of the variance in affective polarization, and that the causation may run both ways.
People live near people like them
Many Americans do live among their own side. Brown and Enos (2021) mapped more than 180 million registered voters and found that a large proportion live with virtually no exposure to voters of the other party, even within the same city or neighborhood. Whether people moved in order to get there is a different question. Martin and Webster (2020) estimated that the partisan bias in moving decisions is “on the order of five times too small to sustain the current geographic polarization of preferences,” and that movers appear to be choosing other neighborhood features that happen to go with party. They conclude that location must have some influence on political preference.
Political leaders grew further apart
A 2022 Pew analysis of roll-call votes found the parties in Congress farther apart, on average, than at any time in the previous 50 years, with about two dozen moderates left compared with more than 160 in 1971 and 1972. This may change how the public reasons. In two survey experiments, Druckman, Peterson and Slothuus (2013) found that polarization “intensifies the impact of party endorsements on opinions” and reduces the impact of substantive information. The open question is direction. Boxell and colleagues note that elite polarization may be influenced by affective polarization as well as the reverse.
Did one party move more? It depends on the measure. Pew’s roll-call analysis, which centers on economic and governmental issues, finds that both parties in Congress moved away from the center and that Republicans moved further. Its 2014 survey found the larger rise among politically engaged Democrats since 1994: consistent liberals went from 8% to 38%, while consistent conservatives among engaged Republicans went from 23% to 33%, after standing at 10% a decade earlier. These numbers describe positions. They do not say whose positions are right, and the rise in dislike shows up on both sides.
Party primaries
The argument is simple: if a small, committed electorate picks nominees, candidates will play to it. The evidence is weaker than the logic. Hirano, Snyder, Ansolabehere and Hansen (2010) tested it on congressional elections and wrote, “Many of our findings are null,” with little evidence that the introduction of primaries, primary turnout or the threat of a challenge is tied to polarized roll-call voting. Hill (2015) keeps the question open. He found that the primary electorate diverged from the general electorate in every House district in 2010 and 2012, which fits a pull away from the center, although closed primary rules did not seem to be what produced it.
Media changed
Partisan outlets are often blamed, but Iyengar’s review says “It is far from clear, however, that partisan news actually causes affective polarization,” partly because the most polarized people are the most motivated to watch. Prior (2013) reported “no firm evidence that partisan media are making ordinary Americans more partisan,” while adding that more media choice lowered the share of less interested, less partisan voters, which made elections more partisan. That review predates today’s feeds. Other work points the other way: one experiment in Iyengar’s review found that discussion groups became more polarized when they included people who had watched partisan media. The review also reports that people were 50% to 150% more affectively polarized by election day than a year earlier, so campaigns themselves may matter.
Is the internet to blame?
One widely cited test suggests it explains a small share, and that test has limits. Boxell, Gentzkow and Shapiro (2017) built an index from eight polarization measures in American National Election Studies data. From 1996 to 2016 it rose 0.47 points among people 65 and older and 0.23 among those 18 to 39, even though 2016 social media use was 30% in the older group and 88% in the younger one. The index grew about as quickly in the decade before 1996 as in the decade after.
This is a comparison of age groups rather than an experiment, and it stops at 2016. Iyengar’s review also cites a study finding a small positive link between internet access and affective polarization, and concludes that “it is too early to conclude that internet usage (and the availability of a wider array of information) plays a definite role in the growth of affective polarization.” The direct experiments are covered in our post on social media and polarization, and the idea of sealed-off audiences in our post on echo chambers.
Is political polarization in the United States unusual?
Among the 12 wealthy democracies in one four-decade comparison, the US had the largest rise in affective polarization. Boxell, Gentzkow and Shapiro harmonized 149 surveys, in work later published in the Review of Economics and Statistics. Polarization also rose, by less, in Switzerland, France, Denmark, Canada and New Zealand. It fell in Japan, Australia, Britain, Norway, Sweden and West Germany.
They then compared those trends with possible causes. The share of people getting news online was negatively or weakly associated with the trend. Private 24-hour news channels, the non-white share of the population, partisan sorting and elite polarization were positively associated, with the strongest links for the non-white share and elite polarization. Among the main factors they plotted, only elite polarization cleared the usual bar for statistical significance, with the non-white share just short of it (p = 0.052). The authors are plain about the limits:
With only twelve countries in our sample, it is difficult to draw firm conclusions about the association between trends in affective polarization and trends in explanatory factors
Boxell, Gentzkow and Shapiro, NBER Working Paper 26669, revised 2021
So which explanation is right?
None has won, and they are not mutually exclusive. Here is the short version.
| Explanation | Evidence for | Main limit |
|---|---|---|
| Partisan sorting | Party and ideology overlap far more than in 1994 (Pew 2014) | About 5% of the variance in affective polarization (reported in Iyengar 2019) |
| Geographic sorting | Many voters have almost no neighbors from the other party (Brown and Enos 2021) | Moving choices look too weakly partisan to explain it (Martin and Webster 2020) |
| Elite polarization | Congress is further apart; significant cross-country link (Pew 2022; Boxell and colleagues) | Cause and effect may run both ways |
| Primaries | Primary voters differ from general voters (Hill 2015) | Many null results (Hirano 2010) |
| Partisan media | Survey and experimental support, mechanism unclear (reviewed in Iyengar 2019) | Self-selection; “no firm evidence” (Prior 2013) |
| Internet | Small positive link in one study (reported in Iyengar 2019) | Largest rise among the least online (Boxell 2017) |
Notice what the table does not do. It does not name one cause, and it does not name one party. The dislike being measured runs in both directions.
I’m The People is a small project aimed at the personal end of this, and it is not open yet. When it opens, it would show you where the people who represent you stand next to where you stand, and put you in discussion with people who place themselves somewhere else, as described in how it works. We have no results and no claim that it would move any number above.
Questions people ask
Why is America so divided politically?
Researchers point to several overlapping causes: party and ideology lining up, more polarized leaders, media change and growing dislike of the other party. Each has supporting evidence and real limits, and none has been shown to be the main cause.
Why is American politics so polarized?
A 2022 Pew analysis of roll-call votes found the parties in Congress farther apart than at any time in the previous 50 years, with few moderates left. Researchers disagree about how much of that comes from voters, from leaders, or from each influencing the other.
What is the political polarization definition?
Political polarization is the movement of political groups away from each other, either in what they believe or in how they feel about one another. Researchers call the first ideological polarization and the second affective polarization.
What is affective polarization vs ideological polarization?
Ideological polarization measures how far apart people are on issues. Affective polarization measures how much more warmly people feel toward their own party than the other one, and Iyengar and colleagues (2019) argue that the two are largely distinct.
How has political polarization in the United States changed?
The American National Election Studies gap in feelings toward the two parties rose from 22.64 degrees in 1978 to 40.87 in 2016, mainly because of colder feelings toward the other party. Whether ordinary voters have moved as far apart on issues is still debated.
Why is America so polarized compared with other countries?
Nobody has established why. A study of 12 wealthy democracies found the largest rise in the US, and among the main factors tested only elite polarization was clearly linked to the trend, but the authors caution that 12 countries are too few for firm conclusions.
Sources
- Iyengar, Lelkes, Levendusky, Malhotra and Westwood (2019), The Origins and Consequences of Affective Polarization in the United States, Annual Review of Political Science
- Boxell, Gentzkow and Shapiro (2017), Greater Internet use is not associated with faster growth in political polarization among US demographic groups, PNAS
- Boxell, Gentzkow and Shapiro (2020, revised 2021), Cross-Country Trends in Affective Polarization, NBER Working Paper 26669; published in Review of Economics and Statistics, 2024
- Pew Research Center (2014), Political Polarization in the American Public
- Pew Research Center (2022), As Partisan Hostility Grows, Signs of Frustration With the Two-Party System
- Pew Research Center (2022), The polarization in today's Congress has roots that go back decades
- Brown and Enos (2021), The measurement of partisan sorting for 180 million voters, Nature Human Behaviour
- Martin and Webster (2020), Does Residential Sorting Explain Geographic Polarization?, Political Science Research and Methods
- Druckman, Peterson and Slothuus (2013), How Elite Partisan Polarization Affects Public Opinion Formation, American Political Science Review
- Hirano, Snyder, Ansolabehere and Hansen (2010), Primary Elections and Partisan Polarization in the U.S. Congress, Quarterly Journal of Political Science
- Hill (2015), Institution of Nomination and the Policy Ideology of Primary Electorates, Quarterly Journal of Political Science
- Prior (2013), Media and Political Polarization, Annual Review of Political Science (abstract at Princeton)
How this was written. Compiled with AI assistance and checked against the sources listed above. Polls and studies report what they found at the time, and they are not I’m The People’s findings. We have none yet. We are not a party, a candidate or a side. If you spot an error, tell us and we will fix it.