This study examines the concept of overqualification as a phenomenon of nonlinearity in the upper deciles of the ability-performance relationship. In three job contexts (military, academic, and industrial), the ability-performance relationship is characterized by a strong linear component below the mean and a ceiling effect at various points above the mean. Consequently, overqualified individuals maintain above-average performance levels, rather than having low performance as was previously hypothesized. However, these individuals also fail to yield increasingly higher performance above the point of overqualification, as classical selection methods would have predicted. Overall, these findings raise possible concerns for traditional top-down ranking procedures, for the setting of high cut scores, and for possible overestimations in the validity and utility of upper-tail test scores. (Contains 4 figures and 2 tables.)