Coding speed score predicts adult earnings as reliably as cognitive test scores

Interesting information from Character Gaps and Social Mobility by Richard V. Reeves and Kimberly Howard. We already know that Cognitive ability (IQ) and the behavioral attribute of Self-Control are, in tandem, powerful predictors of future life outcomes in terms of health, education attainment, wealth accumulation and income. Reeves and Howard add one more behavioral attribute - Persistence.

Measuring character strengths is an empirical challenge. But data are available. One important strength is the ability to persist with a task even if—especially if—it lacks any obvious reward. Professor Carmit Segal (another of our advisers) shows how performance on a coding speed test, when taken with no incentive to do well, predicts adult earnings for male participants over 20 years later, controlling for cognitive skills.

The test first provides a long list of words and associated numbers:

game = 2715
chin = 3231
house = 4232

Then the words are listed again, with a multiple choice of five possible answers:

house = a) 4232 b) 2715 c) 3231 d) 4563 e) 2864.

The test is absurdly easy, but spine-crackingly dull.

Under certain circumstances, the test measures effort, not brains. Those who do well on it do better in life, regardless of scholastic abilities. Indeed, for those without a college degree, the coding speed score predicts adult earnings as reliably as cognitive test scores. An inclination to try hard, despite no obvious extrinsic motivation, turns out to be important in the labor market.

Wide Class Gaps in Motivation

If the coding speed test is capturing an important character strength—motivation, grit, or persistence—an important question is whether there are gaps by social or economic background. It seems so. Adolescents with more affluent backgrounds demonstrate higher levels of motivation, as measured by the coding speed test:

The findings are consistent with the research in The Millionaire Next Door by Thomas J. Stanley. IQ is important and predictive, but behavioral attributes are as important and perhaps are more determinative. And that is great. Not much can be done to materially change IQ; some, but not much. But there is a lot that can be done to change behaviors. It is not easy but at least it is tractable.