Wednesday, 18 May 2016

"My kid did fine, so the exam must be fine." Not.

I can see on the comments thread following my outburst about AQA on Facebook quite a few comments that read along the lines of 'my child thought it was fine'.

The problem with that as a critique of an exam or the exam system as a whole, is that it misses the fact that the whole point of the exam system is to engineer a sufficient number of fails. The fact that some kids come home and say, 'It was OK' is because in all probability they're not the 'fails'.

The questions arise then about:

a) why do we have a system of testing that engineers a sufficient number of fails? what is the use of value of such a system?

b) do the examiners know that the people failing fail for the same reasons? If not, then the test may well not be testing what it's supposed to be testing. So, if a 'subject' has 47 topics and you only examine on 11 of them (that's what our daughter calculated), then how do the examiners know if there is or is not some kind of bias on how they have chosen these topics?

c) what was the purpose in teaching 47 topics of the students are only going to be tested on 11 of them? Is it because the other 36 topics are of equal importance or is it really, as some of us suspect, simply slabs of learnable-off-by-heart stuff, with no real scientific principles involved? So really, it might just as well be the subway map of New York City, or the sequence of archbishops of Canterbury.

d) why is it necessary to have one of these kinds of tests at 16, if students now stay on till 18? If it's about choosing preferences for the next 2 years, there are ways of doing this other than plying students with the stress of 'summative' testing.

e) the test is also a race against time. Even summative tests don't have to be. Another model is to have hundreds of questions, a fixed but long length of time. The students are told that no one ever finishes the exam. Just pick the questions you can do. So in the fixed length of time, you show what you can do. Several of my finals papers for my first degree were in effect like that. This meant that at least we could show what we knew rather than examiners playing games with us to find out what we didn't know. After all, we always don't know loads! What's there to prove?!

Why don't they use that model? Because it is 
i) not punitive and authoritarian, and the exam system is based on puritan ethics about achievement being 'good' and failure being a sin (I jest not), 
ii) it may well not produce a nice bell curve and the whole system relies on a belief that we are all distributed on this bell curve when it comes to our abilities - the tests are devised to produce the bell curve not the other way round, that the bell curve so happens to describe what we are like! 
iii) plenty of students who would normally fail would do better and could be shown to have 'attained' something and even understood it and enjoyed doing it,
iv) teachers teaching to such a test would inevitably engage with much more 'formative' assessment, helping students to become knowledgeable about stuff that the students had chosen to study.

(I'm not saying this is ideal, by a long chalk, but in its own way , it's marginally better than the system in place.)