Correcting Intuitive Predictions

A CORRECTION FOR INTUITIVE PREDICTIONS

Suppose Julie learned to read unusually early and you want to predict her college GPA. Kahneman uses a schematic account of the factors behind the two outcomes:

reading age = shared factors + factors specific to reading age = 100%

GPA = shared factors + factors specific to GPA = 100%

The shared factors involve genetically determined aptitude, the degree to which the family supports academic interests, and anything else that would cause the same people to be precocious readers as children and academically successful as young adults. Of course there are many factors that would affect one of these outcomes and not the other. Julie could have been pushed to read early by overly ambitious parents, she may have had an unhappy love affair that depressed her college grades, she could have had a skiing accident during adolescence that left her slightly impaired, and so on.

The shared-factors account is an intuition for why the measures are related, rather than an exact decomposition of their correlation. How strongly does early reading predict later GPA? Kahneman’s optimistic estimate is about 30%. He uses this estimate in a four-step correction for an overly extreme intuitive prediction:

1. Start with an estimate of average GPA.

2. Determine the GPA that matches your impression of the evidence.

3. Estimate the correlation between your evidence and GPA.

4. If the correlation is .30, move 30% of the distance from the average to the matching GPA.

Step 1 gets you the baseline, the GPA you would have predicted if you were told nothing about Julie beyond the fact that she is a graduating senior. In the absence of information, you would have predicted the average. Step 2 is your intuitive prediction, which matches your evaluation of the evidence. Step 3 estimates how predictive the evidence is. You use that estimate at step 4 to determine how far to move from the baseline toward your intuition.

This approach can help with quantitative predictions such as GPA, investment profit, or company growth, provided the baseline and the relationship between evidence and outcome are meaningful. The approach builds on your intuition, but it moderates it, regresses it toward the mean. When you have good reasons to trust the accuracy of your intuitive prediction—a strong correlation between the evidence and the prediction—the adjustment will be small.

Intuitive predictions need to be corrected because they are not regressive and therefore are biased. Suppose that I predict for each golfer in a tournament that his score on day 2 will be the same as his score on day 1. This prediction does not allow for regression to the mean: the golfers who fared well on day 1 will on average do less well on day 2, and those who did poorly will mostly improve. When they are eventually compared to actual outcomes, nonregressive predictions will be found to be biased. They are on average overly optimistic for those who did best on the first day and overly pessimistic for those who had a bad start. The predictions are as extreme as the evidence. Similarly, if you use childhood achievements to predict grades in college without regressing your predictions toward the mean, you will more often than not be disappointed by the academic outcomes of early readers and happily surprised by the grades of those who learned to read relatively late. The correction aims to reduce this tendency toward extreme predictions. Its accuracy depends on the baseline, the estimated relationship, and the way the intuitive estimate is scaled. It does not guarantee unbiasedness or smaller errors in every case, and unbiased predictions need not have equal probabilities of overestimating and underestimating an outcome.