In Measurements That Mislead, Jonah Lehrer tells us that many tests designed to measure talent turn out to be very bad at predicting long term success, such as SATs and the annual NFL combine. Economists at the University of Louisville say there’s no “consistent statistical relationship” between Combine results and NFL performance.
The reason is that those tests measure maximum performance, which is applied in short bursts. They don’t measure typical performance, which is measured over long periods of time. For example, supermarket cashiers were tested to find out who could scan a few dozen items the fastest. Those results were found to have a very weak correlation with job performance as measured by actual scanning results.
The problem is that some people who have the talent to do well over short bursts do not always have the work ethic or the motivation to sustain that performance advantage over time. These tests do not measure how hard and how consistently people apply their talents. (more…)
