The lay perspective of memory retrieval is a passive one, not unlike checking out a book from a library. One retrieves the information then returns it, unchanged, to its holding place when finished. Research, however, suggests retrieval plays a much more active role in memory formation; namely that retrieving information at Time 1 facilitates retrieval of the same information at Time 2, a phenomenon dubbed the retrieval practice effect. This line of research is one that is currently receiving a considerable amount of attention from both theoretical researchers, who are trying to understand why retrieval practice is such an effective strategy for creating a durable memory trace, as well as applied researchers, who are interested in understanding how retrieval via testing can facilitate improved learning outcomes in an educational context.
My interest in this topic began a few years ago with a simple question: what is the underlying mechanism? While there was an intense amount of interest in the topic, on the whole, there was very little research aimed at identifying a mechanism that could explain the basic effect. In a recent article I laid out one of the first accounts helping to address this gap by drawing associations to other, better-understood memory phenomena. My studies showed that retrieving information at Time 1 strengthens certain kinds of associations at the cost of others. The likelihood of observing improved retrieval at Time 2 is critically dependent upon the degree to which those previously-enhanced associations are leveraged. If one structures the retrieval at Time 2 to be reliant upon different types of associations than those made at Time 1, and interesting phenomenon occurs: memory performance actually decreases. That is, retrieval at Time 1 actually disrupts retrieval of the same information at Time 2.
Though this was an important step in helping to develop a theoretical framework, the finding generates important applied questions as well. Namely, if there are particular laboratory conditions under which retrieval at Time 1 disrupts retrieval at Time 2, are there commensurate educational conditions? If so, this would imply that there are circumstances under which testing is an ineffective a (and possibly detrimental) learning tool. Such findings would argue for significant changes to our current educational policy.
With support from the James S. McDonnell foundation, I hope to further explore both of these questions. On the theoretical end of the spectrum, although I view this study as an important step in specifying a mechanism for the retrieval practice effect, it’s still just that—a step. There is still considerable work to be done to fully flesh out the theoretical framework underlying the effect. On the applied end, it’s critically important to determine whether these laboratory-born disruptions carry over to more ecologically valid educational contexts. Educators who pride themselves on evidence-based practice have long touted the benefits of testing, but I would argue there are potentially important limiting conditions that have yet to be fully explored.