Complexity and the testing effect: The potential importance of securing understanding for the benefits of retrieval practice with complex learning materials

Abstract

Testing (i.e., active retrieval of information already encoded in long-term memory) is considered as one of the most effective ways to promote lasting learning. The testing effect is robust and applies to various groups of learners and learning materials. However, since the testing effect has been discovered in the early 20th century, it has been discussed whether it also applies to highly complex learning materials. The present article reiterates and reflects this debate. In particular, it discusses the conceptualization and operationalization of complexity, and the empirical evidence for the assumption of complexity as a boundary condition of the testing effect. An alternative explanation for the potential lack of a testing effect with complex learning matters is outlined, according to which the occurrence of the testing effect is conceptualized as largely independent of the complexity of the learning material. According to this conception, the prerequisite for the occurrence of the testing effect is that the learning material has been understood sufficiently well during initial learning (i.e., the phase preceding testing) so that learners have successfully encoded the information in long-term memory, the retrieval of which they are supposed to practice in a testing phase. For complex materials, this may require a longer initial learning phase than for simple materials. Finally, we make suggestions as to how this assumption can be tested and outline practical implications for an efficient implementation of retrieval practice even with complex learning materials.