People can use a variety of different strategies to perform tasks and these strategies all have two characteristics in common. First, they can be evaluated in comparison with either an absolute or a relative standard. Second, they can be used at varying levels of consistency. In the present article, the authors develop a general theory of task performance called "potential performance theory" (PPT) that distinguishes between observed scores and true scores that are corrected for inconsistency (i.e., potential scores). In addition, they argue that any kind of improvement to task performance, whatever it may be, works by influencing either task strategies, which comprise all nonrandom components that are relevant to the task, or the consistency with which strategies are used. In the current study, PPT is used to demonstrate how task strategies and the consistencies with which they are used impact actual performance in the domain of morality. These conclusions are extended to other domains of task performance. (Contains 5 figures, 2 tables and 2 footnotes.)