The government pays for billions of dollars of scientific research every year, using taxpayer funds. Most people don’t know anything about how the government decides how to spend the money. Unlike just about every other government program’s budget, spending decisions are not made based on political priorities or in backroom negotiations with powerful senators and congressmen. The key to the system is called “peer review,” which means that each researcher’s grant proposal is examined by other researchers who are experts in the same area. At the National Science Foundation, several times each year they fly researchers to Washington, DC from all over the country, put them in a room that’s wired with networked computers, and ask them to collaborate to select the best research proposals.
I’ve participated in several of these review panels, and I’m fascinated by the collaborative dynamics of these groups. In them, you find both the pitfalls of groupthink as well as the power of group genius. And with just a little redesign, I think these groups could make even better decisions about how to spend the money.
Basically, eight senior scholars read ten 80-page proposals back at home, and then fly to NSF headquarters in Arlington, Virginia to sit in a room with each other for two days. We discuss each proposal for about 30 minutes, and then enter our decision into a special computer system. Some aspects of this process are consistent with group research; for example, it’s a good idea for everyone to think about the proposals alone, before coming together. But with just a few more of the right structures, these groups could avoid the many problems associated with group decision making. For example, researchers have long known that if a group starts with a mild consensus, then talks for a while, each person will end up with even stronger opinions than when they started. In other words, if each scientist likes a proposal a little bit, and they spend 30 minutes talking about it, they’re likely to end up liking the proposal even more. Researchers also know that if everyone shares the same opinion, except for one person, that loner will typically moderate their stance to “get along” with the rest of the group.
Every group faces these problems; but because so few people are aware of the science of group genius, managers don’t put in place the simple structures that improve group decision making.