Scientists Find How to Make Crowds Smarter Than Experts

Scientists at the University of Pennsylvania just cracked the code on why some groups solve problems brilliantly while others fumble, even with experts in the room.

The answer flips everything we thought we knew about teamwork on its head. It's not about finding the smartest person and following their lead.

In a study published in Proceedings of the National Academy of Sciences, researchers discovered that groups perform best when individuals are rewarded for moving the team closer to the right answer, even if their own guess is wrong. They call these helpful contributors "reformers."

The team tested three different reward systems using mathematical models. First, they rewarded traditional experts who got answers right. Second, they rewarded niche experts who focused on overlooked details. Third, they rewarded reformers who improved the group's collective prediction, regardless of their personal accuracy.

The expert approach failed spectacularly. Everyone just copied the top performer until the whole group watched the same thing and ignored everything else. Diversity of thought vanished.

Niche experts worked better but crumbled when problems shifted unexpectedly. The group became fragile, like a house built on narrow supports.

But rewarding reformers changed everything. Groups stayed diverse, recovered quickly from surprises, and kept working even when individual members made mistakes or felt overconfident.

"Reformers don't need to be accurate on their own, but they should be rewarded for improving the collective accuracy of the group," explains first author Guocheng Wang.

Why This Inspires

This research matters because the world's biggest challenges, from climate change to disease outbreaks, are too complex for any single genius to solve alone. We need groups that can pool their diverse perspectives without losing what makes them smart together.

Lead researcher Joshua Plotkin points out that stock markets already work this way. Traders profit not from being right but from moving prices in better directions. Scientific collaborations, however, still rely heavily on individual expertise.

The findings suggest we've been running meetings, assembling teams, and organizing projects all wrong. Instead of asking "Who's the expert?" we should ask "Whose input makes us collectively smarter?"

Schools could reward students for helping their study groups understand material better, not just for acing tests individually. Companies could restructure bonuses around contributions that improve team decisions. Online communities could promote members who elevate group discussions, not just those with the most upvotes.

The research gives us a roadmap for designing systems where everyone's unique perspective strengthens the whole, and where helping others succeed becomes the fastest path to personal recognition.