The insightful Carl Hendrick has produced another interesting read looking at potential lethal mutations of retrieval practice.
Sparked by some new research that shows limited benefits when using retrieval practice with highly complex materials, he flags that we need to be careful not to inadvertently mutate what we know works. In this case, that retrieval practice is shown to be highly impactful when it’s done to retrieve information with ’defined edges’ – facts, definitions etc that are discreet. It’s not so impactful if you are seeking to use it as a tool to learn new things that are not part of your existing schema or for interpreting and/or analysing something highly complex.
Carl flags other ways retrieval practice can inadvertently go wrong. If you’d like some one liners without having to read the whole article:
- Only retrieve what you want to be remembered i.e. prioritise the key components of the curriculum in your retrieval practice activities.
- You can overload working memory with overly complex retrieval – in which case you lose its benefits – as with any cognitive overload.
He also reminds us that to get the most out of retrieval then space it. Repeated retrieval dramatically outperforms single retrieval, and the benefits compound with each successive attempt. Regular readers may remember we have written on how often to space before and crucially sought to flag practical solutions to this challenge.
Finally, we wanted to explore another ‘lethal mutation’ headline from Carl’s article, as we think it has the scope to be confusing. That is “You Can’t Retrieve From Memory What Never Made it Into Memory”. The point he’s making is correct – i.e. Retrieval practice works by strengthening memory traces that already exist, so don’t try and do ‘retrieval’ practice to embed brand new knowledge. As Carl rightly says, “retrieval practice belongs after initial learning, not during it. It is a tool for consolidation, not a substitute for instruction.” However, some of the wording implies that you shouldn’t be retrieving something unless it is already fully embedded. That isn’t the case. Retrieval, particularly spaced retrieval, by its nature, is something that is being done on information that yes, has been encoded, but due to the influence of the forgetting curve over time, is partially forgotten. The whole point of it being retrieved is to use the ‘desirable difficulty’ of remembering it to strengthen the memory trace. Point being, it’s absolutely fine, in fact it’s correct, to seek to retrieve information that isn’t fully encoded, but don’t seek to introduce something brand new via retrieval practice.
As we often say to participants on our Cognitive Coach Course (and this would of course be true for meta learners in development too), ”One of the key skills of the role is to use the right tool for the right job.”. Here’s some suggestions to sense check you aren’t ‘mutating!!
The Retrieval Practice Check: Is it the right tool for the job?
Before you start your next retrieval session, ask yourself these four questions to ensure the strategy remains evidence-based and effective:
- Are the “edges” of this knowledge defined?
- The Check: Is the information I’m asking them to retrieve a discrete fact, a key term, or a foundational concept?
- The “Mutation” Alert: If the material is a brand-new, highly complex theory they haven’t fully grasped yet, stop. You need instruction and modelling right now, not retrieval.
- Is this “Desirable Difficulty” or “Detrimental Difficulty”?
- The Check: Can the majority of the class (roughly 80%) retrieve this with some effort?
- The “Mutation” Alert: If the whole class is staring at blank pages, they haven’t “forgotten” the information – it was never encoded in the first place. Move back to guided practice.
- Am I retrieving “Prioritiy Knowledge”?
- The Check: Am I asking them to remember the “fluff” (the fun story I told to illustrate the point) or the “core” (the actual curriculum goal)?
- The “Mutation” Alert: Avoid “Trivia Retrieval.” Ensure the questions align directly with the high-leverage knowledge required for the next stage of learning.
- Is there a “Feedback Loop” ready?
- The Check: How will students know if they are right? Will I provide the correct answers immediately after the attempt
- The “Mutation” Alert: Retrieval without feedback is dangerous. Without the “correction” phase, students may inadvertently strengthen the “memory trace” of an error.