For the Love of the Game—A Commentary by Ian Ayres ’86 and Barry Nalebuff
For the Love of the Game
By Ian Ayres '86 and Barry Nalebuff
We'd like you to take a short aptitude test. From among the possible answers listed for each question in the table (below), circle the one that is the correct code number for that word. Taken at face value this has to be one of the dumbest tests ever devised. If you can read, then you can find the answers. It turns out that you've just taken part of the Armed Services Vocational Aptitude Battery. These questions were originally designed to find people who would be well suited to clerical positions.
Now for the surprise. Harvard postdoc Carmit Segal has just completed a study that shows that a person's coding-test results predict how much money the person will earn 20 years later in life. A one-standard-deviation increase in coding speed translates into a 7% increase in future income.
So clerical skills predispose you for a chief executive slot? Not quite. As it turns out, the performance on this coding test was measuring something other than clerical skills, and it was that something that is an ingredient for success. That something is conscientiousness.
The Bureau of Labor Statistics surveyed 12,700 individuals between the ages of 14 and 22 and then followed them to see how well they were doing. The original battery of questions had several different parts, including a traditional Army intelligence test and the coding speed test. The participants were paid $50 for answering the battery of questions. As far as they were concerned, nothing turned on whether their answers were correct or not. And that made all the difference.
Segal's surprising result is that coding performance predicts income years later--even after controlling for IQ score. In fact, for participants who never earned a college degree, coding speed is just as important as IQ score in predicting income.
The key fact in explaining the predictive power is that the participants didn't have any real motivation to do well on the test. Thus, Segal argues, doing well was evidence of the person's intrinsic motivation. Many people work hard when paid for quality of performance. This test measured how hard people work when no one is watching.
Just about anyone can improve his coding speed by putting more effort into the test. Indeed, Segal has shown via experiments that coding speed increases when you provide incentives for performance. But there are some people whose speed would stay fairly constant and high regardless of whether you gave them a financial incentive.
Which type would you prefer to hire? In most jobs, pay and performance aren't perfectly linked. You'd like to hire people who will work hard even when they aren't being externally rewarded for their effort. In academia we seek faculty who will continue to be productive post-tenure. At Southwest Airlines (nyse: LUV - news - people ), founder Herbert Kelleher wanted to hire people who were working not just for a salary but also for the psychic reward of doing something extraordinary.
To measure intrinsic motivation, you want to see how well someone does on a test that doesn't matter. That creates a paradox. If you use that performance for hiring, the test matters. You can't ethically or legally give employees a test that says hiring isn't going to turn on a certain section and then base your selection on that portion of the test.
You can look for indirect signs. Google (nasdaq: GOOG - news - people ) has a 300-question survey that asks job applicants about their attitudes, behavior and personality. Some of the questions may pick up on intrinsic motivation. The survey, for example, asks applicants whether they have ever tutored or established a nonprofit organization.
Segal's study suggests that Google and others might also include something as banal as the coding speed task. Imagine that you tell applicants that only half the questions (without specifying which ones) in the coding speed section will be considered in hiring. Applicants who are motivated by extrinsic rewards won't put as much effort into this section as those who just can't help themselves from working. An ethical employer can honestly ignore half the answers while using the other half to indicate how hard the prospective employee will work when external stakes are low.
Postscript to our July 2002 column. We suggested babysitting at movie theaters. Muvico, a chain with 13 new megaplexes, has begun offering just this service. Vice President of Operations Charles DeWitt says the company got the idea from customer feedback. The kid-watching costs the price of an adult ticket plus $9 per child and is thus cheaper and more convenient than a home babysitter. As at Ikea furniture stores, parents are given pagers in case the child needs attention and a matching id bracelet so that only parents can pick up the right kids. Demand is especially high on weekends.
Ian Ayres and Barry Nalebuff are professors at Yale Law School and Yale School of Management, respectively.