It once seemed obvious to me that feedback on writing was less useful when it was too context specific.
So I’d try to avoid writing a comment such as:
The example of Hitler’s appeal to the middle classes would be useful here.
Instead I’d write something that could transfer to other essays:
Give more specific examples to back up your points.
This seems in line with what Harry Fletcher-Wood wrote recently in an excellent blog on feedback. He writes the following about teacher responses designed to improve performance on the current task:
This can help students improve the current task, but its effects are limited: students are unlikely to be able to transfer what they learn about one task to another (Hattie and Timperley, 2007; Kluger and DeNisi, 1996; Shute, 2008): people struggle to recognise that they can use the solution to one problem to solve an analogous problem, unless they receive a hint to do so (Gick and Holyoak, 1980)…We may therefore consider giving more general feedback.
In fact Harry is saying something a bit more complex – more on that later – but just to say that I’ve realised it is mistaken to assume that very task specific feedback is wasted.
‘Iballisticsquid’ helped me reach this conclusion. He is a Youtuber, a genial if rather pasty-faced young gentleman, who has made his millions early by recording himself playing computer games and posting these recordings on Youtube. He and his colleagues such as ‘Dan TDM’ and ‘Stampycat’ have worldwide followings of primary aged fans, including my 9 year old son. Initially I found the whole concept bizarre. Why would anyone want to spend hours watching someone else play computer games when you can just play them yourself? True, the commentary as they play is lively, tells the viewer what the presenter is trying to do and is aimed squarely at the humour level of a nine-year-old boy, but still… The other day my son was gazing in rapt attention as ‘iballisticsquid’ played ‘Bedwars’ a game in which he defended a bed – on an island – while trying to obtain diamonds and emeralds. Suddenly I got it! Even I, seeing Mr iballistic model the game-play felt I could have a go. Manuals and instructions would have simply made me glaze over (as they always have) but seeing iballisticsquid play and I naturally inferred the game premise, appreciated the tools at my disposal and felt empowered. Iballistic Squid and co. are superb teachers. It is fascinating that often they make mistakes and lose the game they are playing. If anything these examples of ‘what not to do’ or ‘non-examples’ simply add to the success of their ‘teaching’ of game-playing. Children watch them play, naturally infer what is transferable to their own game-play and are thus empowered to have a go themselves.
It might seem utterly unconnected but something similar happened when I received copious and very specific feedback on my MEd writing. If I’m honest, at the outset of the MEd I did not even appreciate the nature of academic writing, let alone how my own efforts fell short. Sometimes feedback on my drafts just took the form of examples of how my sentences could be better phrased and feedback was nearly always pretty specific to the content. However, I learnt fast. I infered from the examples how I should write in similar contexts. Inference is what humans do naturally IF the examples are pitched correctly so that inferences can be made.
It strikes that my first suggestion for feedback :
‘The example of Hitler’s appeal to the middle classes would be useful here’.
… is more useful than the second because it provides much richer inference possibilities. This example allows the student to appreciate the nature of the sort of examples that are appropriate in this form of writing, the degree of specificity of those examples, the occasions when such examples need to be used. All this can be inferred because inference is what humans do naturally when given appropriate examples (and non-examples).
It is absolutely correct that people don’t transfer what they learn about from one task to another at all easily and this insight is one teachers must grasp. However, it is when people DO transfer an insight from one context to another that we can say that they have learnt something new. It is also the case that examples are incredibly powerful tools for learning. We make a mistake when we think of ‘inference’ as the skill to be taught when in fact it is what humans do automatically. We make a mistake when we try and teach through generalised principles. People learn through examples. (This is a central insight of Engelmann’s Direct Instruction method which I’ll discuss further in a future blog.)
What can we do to ensure our pupils DO infer? We can repeatedly model the (carefully chosen) specific and as ‘iballisticsquid’ instinctively appreciates, from that modelling our pupils will infer and transfer to new but similar contexts.
Finally do read Harry’s excellent post in which he explains how linking the specific with the more general in feedback can make transfer more likely.
A great teacher: iballisticsquid