I went to a very interesting seminar this morning. Phil Race, from the UK, was presenting about making assessments better in tertiary teaching. There was a lot in his talk (you can download it and other information from www.phil-race.co.uk ) – I’ll just summarise some of the points that are most interesting to me.
1. Assessment started going downhill when, in 1791, the University of Cambridge introduced the first written exam. (Before that, it was purely oral). Not sure that this is ever likely to change – but I can certainly say that in my experience students seem to appreciate feedback a lot more when it is given in person.
2. Don’t put a mark or grade on a student’s assignment when you return it to them. The student will become focused on the grade, to the point of ignoring all your written feedback.
3. Instead, let them work out what their grade should be, based on the feedback you give and how their work compares to that of their peers. I tried this out very briefly this afternoon in a lab class. I normally mark student lab reports by spending a few minutes the following week with the student and going through their report together (see point 1). Today I asked my poor unsuspecting students what mark they reckoned they should get. All but one was spot-on – their assessment was the same as mine. The other one was harsh on himself – I thought his work was of better quality than he did, and I was able to explain why.
4. Never ask a student ‘Do you understand?’ This is likely to trigger the following train of thought:
What is it he wants me to understand? What if I don’t understand it? Will he think I’m stupid? Will my friends think I’m stupid? Will he ask me more awkward questions? How much do I have to understand? Is it a hint that this will be in the exam? etc. etc.
So the student answers …. Hmmm… I’m not sure…which gets no-one anywhere.
And 5. There is so much literature about what works and doesn’t work with assessment that there shouldn’t be any excuse for carrying on with the same methods that we know aren’t much good. Just go and do what works. As the Oracle of Delphi is supposed to have said "You know what the problem is… you know what the solution is…. now go and do it"