Dr. Harold Hislop:

The three errors that were found were all connected with how the data was taken from the datasets dealing with the junior certificate results and then using it within the mathematical modelling. As Mr. Tattan has mentioned on the degree of changes that we were going through, we were trying to work out the best way to use the junior cycle data under what was an intensely short time period. We had discussions at the national standardisation group about the relative benefits of using all of the subjects, of using Irish, English, Maths and one subject or Irish, English, Maths and two subjects, and eventually came down in favour of the Irish, English, Maths and two subjects. Those discussions were really not possible until we had run the model a number of times and seen the effects. It would have been preferable if we had been able to run complete checks on the code but that proved very difficult, in fact impossible.

We were conscious that we had Polymetrika working on the modelling and we attempted to get standby or contingency cover for Polymetrika, including a contingency that would have carried out independent checking of the code for us. At the time it was impossible to get staff after we made approaches. We made approaches in the University of Oxford and we also made other approaches. Each of the agencies in the UK and Ireland, at the time, that would have provided that expertise to us, such as Ofqual, the Scottish Qualifications Authority or the Welsh equivalent was taken up with its own calculated grade process and was also hoovering up the international expertise to do so.