The last ten years have been tough for American democracy, with many scholars focusing on events that seem to signal its decline. But a new study from UC Berkeley suggests that U.S. political scientists often lean too pessimistic, which makes their predictions less reliable.

The research, led by Bright Line Watch, a group of experts studying democracy, proposes a fix: collect predictions from many scholars, find the middle ground, and adjust for pessimism. The result? Far more accurate forecasts.

Unlikely, yet memorable

The study reveals that some dramatic scenarios—like Donald Trump pardoning himself or his allies investigating President Joe Biden—are unlikely, even if still possible. By averaging predictions and adjusting for bias, the odds of these events drop below 50%.

The researchers analyzed surveys from political scientists taken during the 2020, 2022, and 2024 election cycles. The initial responses painted a grim picture, predicting a 44% chance of negative events. In reality, fewer than 10% occurred. This gap points to a pattern: experts often overestimate threats.

Why are scholars so pessimistic? One reason is their focus. Specialists in authoritarianism or democratic backsliding are drawn to these issues because they see them as urgent. This focus can make them more likely to expect the worst. By contrast, experts in American politics tend to be more accurate and less pessimistic.

Attracting clicks

There’s also a media factor. Dire predictions can attract attention, putting pessimistic scholars in the spotlight, even when their forecasts miss the mark.

Bright Line Watch suggests a better approach: trust the collective wisdom of experts while filtering out excessive alarmism. This doesn’t mean ignoring scholars. When adjusted for bias, their insights remain essential.

By focusing on consensus and dialing back unwarranted fear, the study shows we can better understand democracy’s challenges—and build a clearer path forward.