Mehdi Baygi

Retrieval practice = retrieving information from memory, unassisted. Sounds simple, right? Well, there are two vital components that people often get wrong. #1: The information must have already been written to memory. The information cannot be new. It may be fuzzy if some time has passed since writing it to memory -- but it must have been written to memory in the first place. The goal of retrieval practice to take a memory trace that has decayed a bit, and refresh it in a way that slows future decay. You can't do this if the memory trace doesn't exist in the first place. #2: The information must be retrieved from memory, unassisted. The whole point of retrieval practice is to not only refresh the memory trace, but also slow future decay. If you are simply re-told the information instead of being made to recall it on your own, then you may refresh the memory trace, but you will not slow its decay (at least, not nearly as much). You will forget it again, as quickly (or nearly as quickly) as the first time. If you're doing retrieval practice properly, then you should be remembering the information for substantially longer each time. Granted, if retrieval failure occurs, then it is productive to introduce cues/reminders to support a re-attempt at retrieval -- but the cues/reminders should only be introduced after a failed attempt at retrieval. The goal is to strip away the assistance as soon as possible. If you rely on assistance by default, it will turn into a crutch and you will never move beyond it.