The plain man's guide to alphabetical marking

The exam season is ending. Our Classics Part II results came out on Monday (of the six Newnham students, two got Firsts and four got  2.1s -- well done, ladies!). The final meeting for Part 1B happened yesterday, and the Part 1A and Prelims examiners meet today to agree their final grades.

The main job at these meetings is to go through all marks scored by each individual candidate and to assign a "class" to their performance: at the top are the "firsts", then the 2.1s, then the 2.2s, then, at the bottom but still passing, the thirds. (Actually in Part 1A, which is taken by most students at the end of their first year, we don't divide the second class into 2.1s and 2.2s -- it's just firsts, seconds and thirds, if you follow me.)

Each examiner gets a mark book, containing all the marks of all the candidates (known by number not by name -- this is all anonymous), ranked in order of achievement, and we go through the candidates one by one, looking carefully not only at the total number of marks they got, but also at their "profile". So for example a candidate might not have quite enough marks in numerical total - notionally 60% -  to get a 2.1; but of their six papers, 4 might be at 2.1 level with the total brought down by a couple of 2.2 marks . . . That would still get a 2.1. It all takes a long time. Yesterday we took all morning to class about 90 candidates.

To the outsider, the mark books would look like incomprehensible gibberish. Not only does each candidate have three marks for each paper -- and independent one from each of two examiners, but also an" agreed mark" assigned after they have met to discuss the different assessments and possibly revisit the paper. But in addition, each independent mark is given in two forms: in numbers and letters (a series of Greek letters from alpha to delta, with a variety of modifiers, pluses, minuses and question marks. For example, "68: b + +/ a - -" or "63: b+?+" or "51: b - - -" or (lets go up) "80: a ?+"

What's the point.

Well -- it's to try to nuance the simple number. Take, for example, a mark at the very top of the 2.1s -- say 69. Now a 69 could be given to a paper that was consistently very good, right across the board, but didnt ever show that kind of extra flair that you would expect in a first. Or it could be given to a paper that was much more mixed -- with definite bits of first class work in it, but pulled down by less good work. If yoou want to indicate that it's a 69 of the former type, you would give it b+++; if you think it is of the latter, you would give it ba or perhaps if it was even more dodgily mixed b++/a- - -.

The letter marks also allow the examiners quickly to see what kind of performance makes up a translation paper. Imagine now that a candidate has done 3 translation passages and each of them was roughly in the middle of the 2.2s -- at 55. The latter mark would be a simple b. But suppose that the 55 was the total that emerged from a very mixed performance in each of the three passages -- one  first class passage, almost completely correct, plus a complete mess of a translation which on its own would utterly fail, and a very low 2.1 kind of performance. It may average at 55, but not a single passage was at the middle of te 2.2 grade. You indicate that by an alphabetical mark like: b/a/d

There are other refinements too, of course -- like the letter "u" to indicate that the candidate hadnt finished. In that case you give it a numerical mark, taking off the marks for what hasnt been finished. 50u perhaps for an other wise good paper but only got 5% for the last question, where he or she had written only 4 lines. But you then give it a letter mark which reflect the quality of the work that was finished, "ab" perhaps.

All this means that you get a lot more information about the candidate. And if you are discussing a candidate at the borderline of the firsts and 2.1s, you can look and see how much alpha quality work the different examiners had detected. It is of course a considerable art form in its own right. When I first examined here, 25 years ago, we had a wonderful document explaining all this, "A plain man's guide to alphabetical marking". This was later changed, for politically correct reasons, to "A plain person's guide to alphabetical marking", and now we have the "Guide to combined alphabetical and numerical marking", in which the mysteries of the system are elucidated in loving detail.

If you've got this far, you will by now realise what an immense amount of time and care we devote to this process. You might also wonder why we dont just go over to a simple transcript system (there ARE good reasons, I can assure you -- but that's for another time).

(And thanks to JIW who first suggested that I share the pleasures of alphabetical marking on the blog -- and who points out that the basic rules -- although not all the refinements - are available here.)