How we should control time on task—or should we?

Typical experimental studies of learning aids will control in various ways for the time spent on the learning task. This paper reviews such methods and compares the possible confounding influences of indirect time effects in experiments with broad or restrictive time limits. Indirect time effects refer to changes in time spent on a text that are caused by learning aids. They are to be discerned from direct time effects that refer to the time the reading of the learning aids take. In particular, we give an overview of the adequacy of various time limits for answering different research questions. It is argued that experiments without time limits are potentially more likely to bring out performance effects. In the final part of this paper we give some illustrations of the different effects and methods, based on a series of six experiments into the influence of concrete analogies on learning. Those experiments show that indirect time effects are important and that performance results depend on the way one controls reading time.