The Folly of Crowds

Crowds sometimes make poorer decisions than do individuals.

Posted Aug 22, 2011

In 1999, the game show Who Wants to be a Millionaire debuted on American network television. The show proved wildly popular, and by 2000 each episode attracted up to 30 million viewers. Its basic format was and remains simple: contestants answer a series of trivia questions that become progressively more difficult and lucrative. Along the way, puzzled contestants are allowed to use several "lifelines" to break the tie when they aren't sure of the correct answer. Over the past 12 years, the show's producers have grown increasingly creative, replacing old lifelines with several new ones. In fact, the only original lifeline that still features is the cultishly popular "Ask the Audience" lifeline, in which the show's studio audience members enter the response they believe to be correct on individual keypads. Afterwards, contestants learn what proportion of the audience endorses each alternative, and armed with this information they either offer their own response or take the money they have already won and run.

The Ask the Audience lifeline may be popular for a number of reasons, but the reason that seems most compelling to me is the sense that it appeals to a sense of democracy. Many contemporary western societies are drawn to the idea that decisions improve as we incorporate more voices. We embrace democracy-the principle that everyone should have a voice when electing leaders-and we endorse trials by a jury of common peers. More recently, we've begun to rely on the masses for information about every imaginable concept (Wikipedia is now the 7th most visited site on the internet), relying on the wisdom of crowds rather than the wisdom of a few experts.

Crowds are often wise, and we're in turn wise for listening to multiple voices when making decisions. For example, crowds are famously skilled when their members guess how many jelly beans fill a jar, or how much a large animal weighs. If you average their responses, the errors-sometimes overestimates and sometimes underestimates-tend to cancel each other out and you're left with a surprisingly accurate figure. Sometimes, though, crowds are prone to lead us astray.

For starters, the audience on Who Wants to be a Millionaire isn't always right. As the questions become more difficult, the audience becomes less helpful. Trick questions and questions that inspire intuitively appealing but incorrect answers also cause problems. In one notorious case

, a contestant on the British version of the show followed 81% of the audience to her doom, losing £20,000 in the process. As the show's host remarked, "That's one of the highest percentages I can ever remember being wrong." The question asked, "Traditionally, what is the only occasion when alcohol is allowed into the chamber of the House of Commons?" The two remaining alternatives (after two others were removed with the help of another lifeline) were: "The budget speech" and "State opening." The audience overwhelmingly preferred the "state opening" alternative, presumably because they could imagine an alcohol-fueled celebration when parliament opened, but not on the dry occasion when the new budget was announced. One clever contestant came up with a cunning solution that may have helped in this case: he told the audience to choose response C if they weren't 100% certain, because he knew C was the wrong answer.

The same problems arise in complicated jury trials. There's something intuitively appealing about trying defendants before their peers, but expert evidence and the intricacies of the legal system are often difficult for non-experts to comprehend. When a very senior British lawyer was called to sit on a jury in 2004, the judge excluded him, noting that he might be able to understand legal matters that other jurors would fail to understand. Far from a weakness, the lawyer's deeper understanding of the law seems like a notable strength. If the trial process gets the better of laypeople, perhaps the wisdom of crowds should be replaced by the wisdom of experts.

I'm not suggesting that democracy should be swept aside, or that jury trials are always flawed, but rather that our institutions (and our intuitions) are sitting on shaky foundation when they blindly favor many over few. Decisions made by many are sometimes superior, but when the many lack expertise, are biased in one particular direction, or gravitate towards an appealing but incorrect alternative, the crowd's decisions turn out to be more foolish than the individual's decisions.