Question: I was just preparing a lesson, and deciding which exemplar of GCSE work to share with the students *firstMATOMO_URL (The goal in this case was to familiarise students with the assessment levels - eventually they will see exemplars at all levels.) I instinctively went for a top-level model (to show them excellence), and then thought perhaps I should use a middle-level example instead (so there is something to compare with up and down on the assessment criteria, and they aren't completely intimidated). Is there any evidence that we should choose a particular level of model when choosing example responses for students to analyse and learn from? Does this depend on the type of task? For example, a written piece as opposed to an oral activity (such as a presentation)?

anon answered on 24 Apr 2018:

Really interesting question – unfortunately, I don’t know of any research looking specifically at this issue. I spotted some interesting research recently by Robert Nash, Naomi Winstone and team who were looking at whether students better remember comments on their work that were evaluative (what was good / bad on this piece of work) or directive (how they could improve next time). Contrary to expectation, the evaluative comments were better-remembered. – http://psycnet.apa.org/fulltext/2018-09034-001.pdf

Although your question isn’t exactly the same as the research, it could perhaps be applied here. If you started off with the mid-level example, you could then annotate that to show what was good / bad about the answer (based on the higher-level example) and it’s possible they might then remember this evaluative feedback and hopefully might help them with their own writing. In this way, they are presented with a mid-level example, but also know how to improve on that which gives them some information about the higher-level example.

There’s also research showing the benefit of worked examples – for example, https://www.tandfonline.com/doi/pdf/10.1080/2331186X.2017.1297213?needAccess=true – particularly when the content is more complex. This benefit reduces with expertise of the individual, when the effect of practice becomes more important than the worked example. Much of the research is related to problem-solving situations, so again does not directly relate to your question, and I haven’t as yet found research looking at level of difficulty of the worked example.