Since Ebbinghaus first came up with his theory of the ‘forgetting curve’ in 1885, we have understood that memory degrades and the approximate rate at which it degrades, and that repetition in the learning process (eg active recall) aids knowledge retention. Ebbinghaus’s forgetting curve model even suggests the frequency of the repetitions to enhance retention.
Of course, what Ebbinghaus was actually describing is the degradation of short-term memory, and one process that enables information to be slowly embedded in the long-term memory.
Ebbinghaus recognised that the timing, or spacing, of the learning inputs was key to knowledge retention. Some learning methodologies, such as active recall, utilise the principles laid down by Ebbinghaus. The solution we’ve developed, Download™, employs neuroscience-derived spaced learning to embed information straight into the long-term memory.
However, there are two aspects to the training process – and both are time-critical. The first is the learning input; the second is testing to ascertain the effectiveness of the learning input. While some learning methodologies address the timing/spacing aspect of learning inputs to enhance learning outcomes, the time-critical aspect of testing is often ignored.
We know that short-term memory degrades and if we want to test knowledge retention we need to test the information that’s stored in the long-term memory. If we believe Ebbinghaus, and more recent studies have verified his findings, then we know that up to 79 per cent of learning is lost within 31 days. So why, more often than not, do we test learners straight after a learning input? This is only a test of short-term recall – not long-term memory knowledge retention.
We provide employees with training so their skill levels and performance improve. But training can only be applied if it can be remembered – and it can only be remembered if it is in the long-term memory. To test the short-term memory will not help evaluate the training method or the learner.
Memory is formed through a chemical process and that process takes time. You don’t go to the gym and walk out expecting to have instantly lost weight or become stronger. The same is true of long-term memory. Test someone an hour after completing a Download learning module and their recall would be limited. Test them five days later and they should have good recall; the chemical processes would be complete and the memories formed. Test them again after 31 days and you should see the same level of recall because the information is in the long-term memory. It’s neither fair, nor accurate, to test a learner’s knowledge until at least five days after a learning input.
This discipline should be applied to testing irrespective of the learning methodology used. Testing should be designed to reflect Ebbinghaus’s forgetting curve and be conducted after short-term memory has degraded to a point at which the testing is a true reflection of what the learner has actually retained in the long-term memory.
Ebbinghaus suggested that 67 per cent of knowledge is lost within 24 hours and, while other studies have revealed a slightly less dramatic degradation of memory, the most optimistic research appears to suggest that approximately half of knowledge is lost within four weeks (Ebbinghaus puts this proportion at 79 per cent).
It may be impractical to test learners a month after the learning input – some roles require certification in order to work, and such delays would not be acceptable. But surely a minimum period of a week would be practical and provide a much more robust test of both the learner and the training delivery methodology. If we, as trainers and learning providers, are serious about delivering strong learning outcomes, we must put in place measures that help us more accurately assess and refine how we train, and the true ROI that that training provides.
Adam Beard is a managing partner at Download Learning