Many Americans have been shocked by the frequency with which people who claim to love our democracy have supported blatantly undemocratic efforts to limit people's ability to vote or to selectively discard votes already cast. Unfortunately, this sort of democratic backsliding is far from a US-specific problem. Despite widespread support for democracy in countries like Venezuela and Hungary, people have turned out in large numbers to vote for autocrats.

A new study performed in the US suggests at least one explanation for the problem: People across the political spectrum appear to believe their political opponents are likely to take anti-democratic action if given the opportunity. And the strength of this belief correlates with a slightly increased willingness to take those actions first.

Nobody says they like this stuff

The finding, from a University of California, Berkeley-Massachusetts Institute of Technology collaboration, is based on demographically representative survey populations, which were asked about several potential anti-democratic actions. For example, those surveyed were asked if they agreed with reducing the number of voting facilities in towns that support the opposing party. Similar questions got at things like banning rallies, limiting freedom of expression, ignoring court rulings, or resorting to violence. After being asked for their own opinions, people were then asked whether they thought their political opponents supported these anti-democratic approaches.

The good news is that, based on these surveys, nobody—neither Democrats nor Republicans—personally supported these ideas. The bad news is that everybody believed that their opponents had much higher levels of support for these policies than they actually did. (This is not an effort to both-sides the US's issues with support for democracy; we'll come back to that below.)

One trend that was apparent in the data: The more you were willing to believe that your opponents were likely to support the subversion of democracy, the more you supported taking those steps yourself. To test whether this might be causal, the researchers did a variation on the initial survey. This time, people were asked about whether they thought their political opponents supported the anti-democratic actions and then were told the actual low degree of support the actions had. Only after being corrected about their opponents were they asked for their support for taking these actions.

Doing things in this order reduced support for these attacks on the democratic process. The effect was small because support was pretty low to start with, but it was consistent across questions. This indicates that at least some support for subverting democracy comes from people fearing that their political opponents want to subvert it themselves.

The results were also replicated as part of the Strengthening Democracy Challenge, where it was tested in a study with over 32,000 participants as a means of reducing anti-democratic attitudes. In that work, it was the most effective intervention of everything tested.