Time as a variable in learning on the World-Wide Web.

Research in education, psychology, and neuroscience motivates a hypothesis that learning takes time. Support for the hypothesis was found in four replications of an upper level undergraduate course in which the material and activities for 50% of the topics were delivered over the World-Wide Web. Computer records were correlated with three types of test questions: multiple-choice, short-answer, and open-ended essay questions. Positive and significant correlations with time were observed for 33% of the correlations involving multiple-choice questions, 13% of those involving short-answer questions, and 60% of those involving open-ended essay questions. An estimate of the common underlying correlation, rho, equal to .35 for the four replications was significant. The data also revealed generally low overall study times and a maladaptive pattern of "cramming" before tests instead of distributing practice. In the Discussion section, we argue that computer study times can be used as a predictor of subsequent test performance, which is a measure of student learning.