Consider this little gem that periodically does the rounds in education circles:

(This is just one of the very many variations on the Learning Pyramid doing the rounds on the internet. For more examples visit The Corrupted Cone of Experience.)

Seductive, isn’t it? The false sense of security comes from the fact that it bears out and validates our experience as teachers: we get to know our subjects so much better because we teach them, so it follows that the best way to retain new information is to teach it to someone else. And look: there are some numbers so it must be true!

But, hang on, all the percentages are multiples of 5. I’m no statistician, but what are the odds? It looks a little too good to be true. And guess what? It is too good to be true. The National Training Laboratories based in Bethel Maine, who are cited as having conducted the research on which the learning Pyramid is based, apparently have no real idea of its provenance despite having claimed to have worked it all out sometime in the 1960s. When asked to provide this research they said, “NTL believes it to be accurate but says that it can no longer trace the original research that supports the numbers.” (Magennis & Farrell, 2005)

But it’s not just the percentages that are dodgy, it’s the fact that the whole thing has simply been pinched from a chap called Edgar Dale. Here is Dale’s actual Cone of Learning from his book Audio-visual Methods in Teaching. (p. 39)

Dale’s actual Cone of Experience

Spot the difference: no numbers! Dale also warned readers that the cone should not be taken literally – it portrays more concrete learning experiences at the bottom of the cone and more abstract experiences at the top of the cone. “The cone shape was meant to convey the gradual loss of sensory information” as students moved from lower to higher levels. “The root of all the perversions of the Cone is the assumption that the Cone is meant to be a prescriptive guide. Dale definitely intended the Cone to be descriptive – a classification system, not a road map for lesson planning.” (Subramony et al 2014)

The problem with any representation of the learning process is the dubious idea that a one-size-fits-all, magic bullet approach to learning will work in every context for every student. Unsurprisingly, reality is a little more complex and such percentages are an attempt at simplification. Making lessons multimodal may well result in an increase in learning but not always and certainly not as a de facto cause and effect. Sometimes it could be much more useful for students to get on with practising something on their own than to have a discussion with their peers. It might be true that teaching others could be an effective way to learn, but what’s the evidence beyond our intuition? This kind of belief is an emotional one and is pretty much like everything else that gets believed: a classic case of confirmation bias. But that doesn’t excuse people making stuff up to support what they believe.