Aug. 26, 2010 -- The old adage "two heads are better than one" really is true when both heads are equally competent, new research finds.
In a study of shared decision-making, researchers from the University College London (UCL) showed that two people of equal abilities solved problems better when they worked together.
But when a competent person was paired with someone who wasn't so competent, group and individual performance suffered.
The findings challenge the prevailing wisdom that groups rarely outperform their best individual members, neuroscientist and researcher Chris Frith tells WebMD.
The study appears in the Aug. 27 issue of Science.
"Our model showed that two equal heads are better than one for solving a problem," he says. "But if the abilities of two people are very different, they are probably better off working alone."
Joint Decisions Often Better
Bahador Bahrami, PhD, designed the study to test the "two heads" hypothesis during his time as a UCL research fellow.
In one experiment, pairs of volunteers were briefly shown weak images on separate computer screens and asked to recall when they saw an "oddball" image with a slightly higher contrast than the others. When their answers did not agree, the two volunteers discussed the matter and came up with a joint decision.
Some volunteers were better at the task than others, but the joint decisions proved to be more accurate more often than individual ones when two people were paired who were equally good or equally bad at identifying the oddball images.
Frith believes this may be because individuals tend to lose focus over time.
"Attention spans can wander in trials like this, but it is not likely to happen to two people at the same time," he says.
In another experiment designed to explore the impact of incompetence on shared decision- making, one person was shown clear images while that person's partner was shown images that were much more difficult to see.
In this case, when the volunteer with the good information conferred with a partner made incompetent by bad information, joint decisions tended to be worse than individual ones made by the better performing partner.
The findings suggest people can work together most effectively when they understand their individual competence level, Frith says.
"Joint decisions don't work when a member of the team is incompetent, but doesn't know it," he says.
Bahrami acknowledges that such self-insight is uncommon in the real world.
"If you ask a group of people whether they are above-average drivers, most of them will say yes," he says. "But this is logically wrong because everyone can't be above average."
He does believe the findings have real-world applications, especially in situations where competence is quantifiable.
"If you have a stock trader with a very good track record for picking stocks and one with a poor record, the company may earn less money if they are asked to work as a team," he says. "But two highly competent ones might make more money for the company by working together."