Working memory span tasks are popular measures, in part, because performance on these tasks predicts performance on other measures of cognitive ability. The traditional method of span-task administration is the experimenter-paced version, whose reliability and validity have been repeatedly demonstrated. However, computer-paced span tasks are becoming increasingly more popular. Despite their popularity, no study had systematically compared experimenter-paced and computer-paced versions of the reading span and operation span tasks. Such a comparison is important because research labs in many universities across many countries administer these span tasks with a variety of methods. The purpose of the present study was to evaluate the reliability and validity of computer-paced span tasks and to compare these estimates to those of experimenter-paced span tasks. Results indicated that experimenter-paced and computer-paced span tasks share some overlap, but also measure additional and distinct processes. Computer-paced span tasks were highly reliable measures as well as valid indicators of fluid intelligence (Gf). Thus, computer-paced span tasks may be the optimal type of administration given their methodological advantages over experimenter-paced span tasks. (Contains 1 figure and 5 tables.)