Want to improve your memory for facts? Tom Stafford describes a counter intuitive idea for retaining information.
If I asked you to sit down and remember a phone number list or series of facts, how would you go about it? There is a chance that you doing it wrong.
One of the amazing things about the mind is that although we all have one, we do not have best insight into how to get the best from it. This is in part because of flaws in our capability to think about our thinking, which is known as metacognition. Studying this self-reflective thought process exposes that human species has mental blind spots.
One place where these blind spots are particularly big is learning. We are actually shockingly bad at having insight into how we learn perfect.
Researchers Henry Roediger and Jeffrey Karpicke set out to look at one aspect: how testing can combine our memory of facts. In the in the test they asked college students to learn pairs of English and Swahili words. So, for instance, they had to learn if they were given the Swahili Word “mashua” the best response was boat. They could have used the type of facts you might get on high-school quiz (who wrote the initial computer programs? Ada Lovelance), but the use of Swahili meant there was a small possibility their participants could see any background knowledge to support them learn. After the pairs had all been learnt, there would be an end test a week later.
Now if many of us were revising this list we might study the list, experiment ourselves and then repeat this cycle, dropping things we got right. This makes testing faster and permits us to focus our efforts on the things we have not yet learnt. It is an idea that seems to make best sense, but it is an idea that is disastrous if we truly want to learn correctly.
Roediger and Karpicke asked students to prepare for a test in various ways, and matched with victory – for instance, one group stopped testing themselves on their right answers, while another group kept testing themselves on all items without dropping what they were getting correct.
On the end exam differences between the groups were remarkable. While dropping things from study did not have much of an effect, the people who dropped things from testing performed bad: they could just remember about 35% of the word pairs, matched to 80% for people who kept testing things after they had learn them.
Finally, the expert had the best idea of asking their participants how well they would keep in mind what they had learnt. All groups guessed at about 51%. This was big overestimate for those who dropped items from experiment.
So it seems that we have metacognitive blind spot of which revision ideas will job best. Making this a condition where we need to be guided by the proof, and not our instant. But the proof has a moral for teachers as well: there is more to experiment than finding out what students know- tests can support us remember.