What we might have to do to check learning has happened. Part II

Recently I tweeted

Learning: When tested can reproduce; when tested later can reproduce; when tested in a new context (same domain) can reproduce.

I am trying to sort out the issue we have where performance is not the kind of evidence of learning we previously may have thought is was.

Part 1 is a blog which tries to identify the issue. This will try to explain a sort of way to move forward. I wish I could be less cagey about this. I am still mulling it all over.

One of the difficulties is that even if we could say that a child had learned something and we tested them today how would we know they had retained and could access and use that learning on a future day? The simple answer is that we cannot. All we can be sure of is that, subject to the vagaries of the testing process, they seemed to be able to show they had learning when they were tested.

So what we need to do is have a process that provides the opportunity to learn, ie teach them and then provide follow ups that challenge the learner and provide them with the opportunity or, better, the need and desire to think even more about the learning.

The more they think, the better the learning.

(Someone quotable must have said something like that, surely!)

So we accept that we can only really say that they know it now. If they know it now and we have done some securing of learning things then they are likely to have learned well enough to be able to find evidence of that learning in the future. Hopefully when they do the exam the learning was aimed at. (Yes, I know about the lifelong learning stuff!)

Part 3, the next blog, will try to identify a structure that will make the retention of learning more likely.

I have finally sorted how we know learning has happened

Hmnnn… It has annoyed me ever since I read the work of folk, such as David Didau @learningspy, have that suggested we could not see learning and that performance had a somewhat unexpected¬†and contrary connection to learning.

As with most teachers, I expect, I had always assumed that if I followed a model something like:

  • teach it
  • check a few kids using questions
  • show a few examples and then
  • let then practice
  • check that they were doing it right

that this process would result in learning and would be evidence that learning had taken place.

Then along come two pieces of evidence. The first is that performance now does not of itself give us a very secure view that the thing being learned will last over time. A quote that exemplifies the issue is

“Teaching takes place in time and learning takes place over time.” Professor John Mason

There is a ‘now’ of teaching and a ‘later’ for learning to have happened, or not!

The second is that a good performance now may well mean less effective learning will take place over time. So if we try to secure great performance while we are teaching a child that may well limit the security of the learning that should result from our teaching. I will probably explore why this might be in a later blog.

So we seem to be left with no way to monitor how effective our teaching is and has been because we first can’t rely on a learner’s performance to tell us and also if we get a good performance that may will indicate insecure learning later.

The ‘solution’ in the next blog.

Why it might be a good idea to learn formulae

This post was prompted by a slightly frustrating set of twitter tweets with some science teachers who were complaining that the new science GCSE specifications required students
to learn 20 formulae. Previously the formulae had been given on a sheet, presumably, available in the examination. Their views seemed to be:

It is pointless to learn something one can easily look up, or be given

That learning stuff fills up your brain with useless information

That to learn a formula would detract from understanding. Children would stress about learning the formula and not then focus on using and understanding.

It takes away precious time from the specification which is already very full, overfull possibly.

The students do not need to know the formula, they only need to be able to recognise it.

My own view is that it is good for the students as learners to commit to memory, rote learn, learn by heart – I don’t mind which phrase you use, but do not have a prejudiced response to the sometimes emotive idea of ‘simply’ learning.

It is utterly spurious to think that if we get children to learn by heart that we somehow deny them access to understanding. My contention would be that they are better served by knowing the equations than having to look them up each time.

I think that the teachers who undoubtedly do know and understand the equations are making the error that experts often make. They place too little credit on their ability to use the equations as a result of their knowing the equations.

The issue is one of the cognitive load we place on students by denying them the opportunity to learn the equations. When solving a problem there are a number of things we need our working memory to do. One of them is to keep the equation in mind. This is supported by writing it down but it is better supported by already knowing the equation to be used. The problem is likely to be one that requires the equation to be use, manipulated in some way. Why add to the cognitive load by requiring the, at best, poorly known equation to add to this load?