Z Score Imbroglio

| by Prof. R.O. Thattil

( January 17, Colombo, Sri Lanka Guardian) I am writing this article to clarify certain matters raised at different forums, some in the form of allegations.

The Secretary of the UGC stated on TV (Rupavahini programme held on 14 January, 2012) that I had not been invited to the committee because I submitted a document that the Z score was incorrect in a court case a few years back. This is totally untrue. How could I go against the Z score that I advocated a decade ago? What really happened was I submitted a document stating that individual Z scores received for each subject should not be rounded off to the second decimal place before calculating the final average Z score. The computer can hold many decimal places! The rounding off error can go up to 0.005 which can decide the future of one or more students. Who rounds off intermediate calculations anyway when using a computer?

The second statement he made was that I was against the UGC since I had not been reappointed for a second term as Director of the Postgraduate Institute of Agriculture. What has this got to do with my speaking for justice? If something is wrong one has a right to point it out. The statements made by the Secretary and a few others try to distract the attention of the public from the main point; namely the error in pooling means and standard deviations from the old and new syllabuses of a subject in formulating the Z score.

The committee appointed by the President to look into the problems in releasing the A/L results, has stated that everything has now been corrected. However, they avoided the key question about the error in pooling. Perhaps, they were not mandated to do this.

A member of the committee that suggested pooling, stated that there is not much of a difference whether it was pooled or not. If so why choose the wrong method (pooling)?

Saying that there is not much of a difference is a feeble excuse. The so called small difference could affect a large number of students. I have shown in an article written in this Newspaper (13/01/2012) how pooling affects the Z scores. Some persons have reacted to this by saying it is only an example. True it is an example, but in theory too, when 2 means are pooled, the pooled mean will always lie between the 2 means! This is what leads to the discrepancies.

A Minister of the Government has stated that I was not the founder of the Z score but it was someone else in Sri Lanka! No one can claim to be the founder of the Z score which has been used for over a century. It may be Sir Ronald Fisher (the father of Modern Statistics) who used it first for scaling variables. However, I am the person who proposed the Z score to scale marks in 2000 for the A/L examinations, so that the average of the Z scores received for different subjects can be used in ranking.

I fervently hope that the main issue of pooling will not be clouded by arguing on irrelevant issues. Let justice be done!