If you want a job at Google or McKinsey, you’ll have to go through a rigorous process. This process often includes as many as 10 interviews, and requires one to provide six or more references.
The intention is good, for we all know that the more feedback one can gain on a candidate, the better. And this truth was discovered a long time ago. In 1906, for instance, Englishman Francis Galton, a cousin of Charles Darwin, stumbled upon an intriguing contest while attending a livestock fair. An ox was hanging on display, and the visitors at the fair were invited to guess the animal’s weight after it was slaughtered and dressed. Nearly 800 participated, but not one person hit the exact mark: 1,198 pounds. Galton’s insight was to examine the mean of these guesses from independent people in the crowd. Astonishingly, the mean of those 800 guesses was 1,197 pounds — accurate by a fraction of one percent.
Today, this phenomenon of having more accuracy collectively than any single individual is called collective intelligence, and the field is booming, as research has shown over and over again that estimations coming from many people, in the right circumstances, lead to results closer to the truth. This is because the extremes are essentially cancelled out. That is why it is often referred to as a statistical phenomenon.
Because of collective intelligence, organizations that try to stay on the cutting edge of information and technology think they should interview more individuals, because more is better.
This isn’t always true.
An article called “How social influence can undermine the wisdom of the crowd effect” discusses how we as social beings are subject to influences that will often make people revise their estimations and affect the accuracy of the outcome. In their research, the authors show how the “social influence effect” — such as learning that an interviewer, or your boss, favors a candidate — diminishes the diversity of the group without improving its accuracy. The experiments conducted by the authors tested objective, verifiable data and controlled for social conformity biases.
The group subject to social influence became actually less reliable in guiding the decision-makers. As the group is influenced to head in one direction, and the range of options are reduced, the supreme perverse effect is the “confidence effect,” or overconfidence, as everyone reaches an agreement although the decision may be wrong.
This can translate into a hiring manager who relies on 10 very keen interviewers being supremely confident in their decision and exhibiting a high level of certainty while the accuracy, in reality, isn’t improved a bit.
Ultimately, this very fascinating research shows that even though we tend to believe that debating a decision will lead to a better choice, this assumption fails to take into consideration the fact that no one is immune to social influence.
In an interview debrief that organizations tend to perform by phone or in person, the most vocal person usually impacts the outcome and the decisions of all the others. All those involved will eventually merge toward a common assessment and the confidence of the group will be raised. The group will then make a decision feeling very certain they have made a good choice, but a single answer is not the stamp of accuracy.
To prevent this, ask for feedback on the rating of the candidates before any discussion starts. This can be done online before the meeting or even at the meeting before it starts. If you are leading the hiring committee and are leaning toward a particular judgment, write it down rather than consciously or unconsciously leading the group in the direction you wish to take.
The next time you debate the hiring or promotion of an individual, think about Francis Galton and the ox and ask yourself if the decision has been truly driven by diverse feedback or the views of one strong minded, vocal individual.