Mr. Pat Burke:

Where does one draw the line? There were a number of other issues. I take the point about the particular challenges we have in the area of languages and we are probably unusual in this country in having compulsory Irish, which may be a factor in the lower uptake of a multiplicity of languages than might otherwise be the case. That is not an argument against compulsory Irish but an observation.

One question was on errors in the papers and the subject of the bell curve, and there is a slight connection between the two. I will try to explain what our system is because one does not often hear it articulated. In some jurisdictions there is a bell curve, where the top X% get grade A, the next grade B, the next grade C, etc. They change the grade boundaries every year to ensure that happens and it might mean the boundaries for grade A are between 79% and 100% in a particular year. That is an open and transparent thing but it is not the system we have. We have a system that works on the assumption that, year by year, there is not a dramatic change in the cohort. There may be slight variations but it would almost be perverse if the cohort for 2015 was appreciably different from the cohort for 2016. One would have to ask why that would ever be the case as human beings do not alter in that way. There is an expectation that, on a year-by-year basis, there will be a broad symmetry between the results, though it will not be absolute and there will be some variations.

The marking commences with a marking scheme which is designed to reward fairly what the students do on their papers and to result in an outcome that is not materially different from the preceding year or two. That does not mean that, over a period of time, things do not change in incremental steps because they do. Sometimes there is a syllabus change and that brings about something more material. If it emerges during the marking that we are getting a very significantly different outcome from the preceding year alarm bells ring. It can be because there is something up with the paper, such as a question that is phrased badly. In such a case the scheme is openly and transparently changed to reflect that. The marking scheme is the piece of equipment our system uses to generate a fair outcome. Some jurisdictions change the grading bands but we change the scheme in the way I have just described. There are arguments for and against both systems. It should mean that if in a given year - and it happens every year - there is a question that seems to throw a sizeable proportion of kids or is confusing it is, quite legitimately and transparently, dealt with in the marking scheme.

There is not a bell curve but we do accept, as a legitimate proposition, that there should not be 25% getting grades A, B and C in one year while 50% get those grades in the next year. It would be bizarre if that happened and it would only happen if there was something really strange in the test instrument and it tested the second cohort of kids less onerously than the first. That is the best answer I can give to the question on the bell curve and it also partly answers the points about questions that throw kids.

The other question was on errors. The fundamental problem with exam papers is that not that many people see them and that is for good reasons. When I ran exams I never saw an exam paper and would not have wanted to see one until there was a problem with it after the event, in which case I would take it home. Before the event, very few people actually see exam papers. In some countries they pre-test exam papers but this is not that common and there would be all sorts of security risks. There will inevitably be some erratabut they should always be marginal. If there was a serious mistake one would have to question whether the paper was fit for purpose and consider applying the contingency paper in substitution for the actual paper.

The purpose of sending in the batches is for individual examiners. A supervisory regime applies.