Cornell University, Ithaca, New York

A computer scientist wonders how much information is really good for us.

I am interested in understanding how groups of people or computer systems work together to solve complex problems. This is relevant in real-life situations that demand collective problem-solving, ranging from scientific research to military operations, so we hope to learn about the underlying mechanisms through experiment.

Stanley Milgram's famous 'six degrees of separation' studies form one such set of experiments. In these, participants were asked to help send a letter to a far-away stranger by forwarding it to a friend they thought might know the target. That this strategy often succeeded hints at how people lacking a global picture of the social network they inhabit can still jointly solve a difficult search problem.

One of the interesting questions here is how a group's ability to solve a problem is affected by the amount of information available. I expected that if people had a global view of the system, rather than just a local one, their effectiveness at solving the problem would increase.

A fascinating experiment (M. Kearns et al. Science 313, 824–827; 2006) shows that this isn't always so. The researchers posed a task in which they deliberately varied how much information was revealed to participants about what others in their group were doing.

For certain settings of the problem, giving participants a global view significantly slowed down progress. People faced with too much information in a time-pressured setting became 'overloaded', and this impaired the group's function.

As we consider designing tools to help people work together effectively, we should remember that increasing everyone's situational awareness might not always lead to improved performance.