Limited accuracy of premorbid intelligence estimators: a demonstration of regression to the mean.

Regression-based premorbid intelligence estimators have been devised by Barona, Reynolds, and Chastain (1984), Barona and Chastain (1986), Hamsher (1984), Krull, Scott, and Sherer (1995; the Oklahoma Premorbid Intelligence Estimate: OPIE), and Vanderploeg, Schinka, and Axelrod (1996; BEST-3 approach), but little is known of their relative accuracy, particularly in outer ranges of intellectual ability (e.g., below-average, superior, etc.). Towards this end, the Wechsler Adult Intelligence Scale-Revised (WAIS-R) was administered to 150 neurologically normal adults, and estimated VIQ, PIQ, and FSIQ scores were computed according to each regression method. Results showed that methods based solely on demographic factors were most susceptible to meanward regression, rendering them poor estimators of IQ scores in outer ranges. Although the OPIE and BEST-3 performed somewhat better, their accuracy remained relatively weak. The findings suggest that regression-based estimates of premorbid IQ are very susceptible to error, particularly in outer ranges of intellectual function.