Self-explanation

The EEF talks a lot about fidelity when it comes to implementing research evidence. In essence this means that when we adopt evidence-informed practices we stay faithful to the core ideas that underpin them.

It is all too easy to warp and mutate research evidence to fit our own agendas, biases or predilections. However, to do so is dangerous. Hiding behind a shield of research evidence we enforce change based on a non-existent mandate. We use the language of the research but the end product is not the same. In fact, even the subtlest loss of fidelity can render the best intentions not only useless but potentially damaging.

So why bring this up now?

The reason is that from my reading around self-explanation, this is an area of research evidence vulnerable to misinterpretation.

It sounds straight forward, as many evidence-informed practices do. You get students to explain back to themselves something that you have already taught them. Easy.

However, that’s not quite it.

Work going back several decades by Michelene Chi and colleagues demonstrated that American college students who explained physics problems to themselves outperformed those who did not, and that younger students prompted to explain passages about the circulatory systems learned more about it than those who read the same passage twice.

The basic premise then is that it is advantageous to learning to give students some stimulus material and ask them to explain its meaning themselves. In fact a meta-analysis of 64 studies containing 6,000 participants found that it is better to ask a students to see if they can explain something to themselves, than for a teacher or a book to always explain it to them.

Furthermore, this article from Tania Lombrozo via the American Psychological Association highlights the following aspects of self-explanation that are useful for teachers to know:

Studies have shown learning is improved when students are asked to “explain why” a solution to a problem may be correct, as opposed to simply describing the problem or the process they are following.
Self-explanation can be supported with teacher prompts (although these are scaffolds for self-explanation rather than partial explanation from the teacher).
Self-explanation creates links between existing knowledge (or schema). These inferences allow us to create new understanding and therefore aid deep learning of the material.
Self-explanation does not necessarily require feedback to have a positive effect.
We often favour the most obvious explanations and as a result can misinterpret the correct explanation for a “messy” concept.

As a school that, while having a tight but loose approach to teaching, would align itself with direct instruction as the principle that best describes what most of our teachers do, this evidence doesn’t necessarily fit neatly with our model.

However, this is the essence of being evidence-informed. We must be prepared to have our conceptions challenged and to leave our minds open. What we must avoid doing is sending out a twisted version of how to implement self-explanation because it suits us to do so.

Another salient point is that some of the principles of self-explanation chime well some key aspects of our teaching and learning approach. In particular elaboration and metacognition. Good self-explanation requires the elaboration and Socratic questioning approach that we have prompted as a key mechanism for moving learning from surface to deep. Similarly, structured self-explanation would represent a useful way to articulate metacognitive regulation in which students actively monitor their own learning as it happens.

Ultimately then, the approach here should not be to dismiss self-explanation because it challenges us, and not to mutate it to fit our purposes. Instead we should add it to our layers of understanding in the messy and confusing world of learning, accepting its lessons, and attempting to learn them faithfully.

21 May, 2019 —