I am knackered -- for a myriad of reasons, but partly because I have just finished being chair of our Part 1a exams. OK I am sure that taking exams is more exhausting than marking them; but marking takes it out of you too, and you don't get all the sympathy/ (The picture shows the examiners consuming a well-deserved lunch after their labours.)
It's tiring not only because of all those scripts to mark (pushing 200 in my case) but because of the absolute obligation to make the process as fair as it could possibly be. By fair I mean this: would it be possible to face each and every candidate with their results and explain to them why they had got a 2.1 or a 2.2 or whatever -- and to feel comfortable with that explanation? To put it another way, "classing" (explained here for those unfamiliar with this system) candidates cannot possibly be fair if the classers themselves don't believe in the results.
That is not to say that the whole process is without its rough edges. There are candidates who do badly on the day because of all kinds of unpredictable and unfair reasons (row with girlfriend, toothache etc). And it is not to say that there might not be all kinds of other ways of assessing university achievement (some of which we already use, side by side with sit-down exams... dissertations, portfolios of essays etc). but -- the bottom line at the final stage -- the examiners' job is to assess and grade the results in front of them fairly.
In my experience, we do this as well as is humanly possible within the constraints of the system -- and it is time-consuming.
How do we do it?
I blogged last year about using both alphabetical and numerical marks. But the nuances in our system don't stop there.
The key is to have principles and criteria, not rules. It's a key, in my view, that might have an application for other areas of life ( isn't part of the problem about MPs expenses that it has got fixated on universal applicable rules, rather the principles of what is reasonable . . .).
In examining, rules can end you up in trouble. Let's suppose you have a rule which says that anyone with an average of over 70% gets a first and that anyone with an average of over 68.75% might be "considered for one. Sounds about right. But then take the case of a candidate who has taken 5 papers -- in four of them they have got firsts (three 70s and one 71 -- agreed by each of two examiners, blind marking); in one they have got a 2.1 (62). No one really doubts that that candidate looks like a first, but their average is only 68.6% -- so not in the zone of consideration. Is that fair? No.
In my experience of other systems of marking, what often happens in those circumstances is that someone looks at the papers again and "finds" another mark somewhere and the average goes up just enough to tip over the boundary -- but that is a desperate strategy, if you ask me.
So what we do is look at a whole combination of factors: raw average score, spread and preponderance of marks, how many first class marks were given by individual examiners before the final markfor each paper was agreed, and so on.
At our final meeting we have all the marks for all the candidates
laid out in a list, in rank order -- starting from the highest
numerical total, down to the lowest. We then go through them one by
one, discussing each individually. Some -- those, for example, with all
first class marks and a stratospheric average -- take only a few seconds; the nearer a boundary you get, the longer the discussion. On Wednesday, it took us about an hour and a half to classify the first twenty or so candidates, from the firsts into the top of the seconds.
You may not like this system, but you cant say that it isn't conscientiously applied.
Anyway after all this... it was with a lighter step that I went with the husband to the weekend in New York, postponed from when the ash-cloud struck; on which more soon.