Dr. Harold Hislop:

One of the most important elements of those other data concerns students in each leaving certificate year cohort in each school. We tracked back what they got at junior certificate level. Very sound research has been published showing how junior cycle scores can be used, on a whole-class basis, to predict the level of achievement likely at leaving certificate level. It gives us a very good indicator, therefore, of what should be the expected level of achievement in that school.

There is also the historical national standard, which influences those data as well. The regression model or prediction that goes from junior cycle to leaving certificate, combined with the teachers' scores, can be compared and contrasted. That gives us a good indication of the extent to which the group of teachers teaching in individual subjects - this is done subject by subject and not school by school - conform to the pattern of results that we would have expected those students to achieve. There are also several protections built in, so that if there are exceptional students that stand out in the teachers' data, for example, parts of the software will pick up those students and ensure that they are adequately rewarded and that their scores are not moderated downwards by too great an extent.

The predictions that we have from the model, therefore, can give us a very close estimate and a good prediction of the class groupings and the scores which should or should not be increased. The details of how all that works, and the mathematical modelling behind it, will be published next Monday in conjunction with the publication of the leaving certificate results. A large technical report completed in April when the original proposals were drawn up will also be published. The details of the model, how it evolved, how it works and its mathematical underpinnings will also be published next Monday.