IIT math: one query, 4 answers

[ganpat1956]

New Delhi, Aug. 7: The Indian Institutes of Technology still cannot explain the method they followed in setting the admission criteria in 2006 — a whole year after the process.

They have so far given four answers, some contradictory and some impossible to verify.

The Telegraph had on Monday reported an allegation by some candidates’ parents that the IITs had flouted their stated procedure — divulged under the Right to Information Act (RTI) — for setting the cut-off marks for physics, chemistry and math.

That procedure was one of two contradictory explanations the IITs have given the parents. They have now given a third explanation to Calcutta High Court, where one parent has challenged the 2006 admissions.

An IIT administrator involved with the 2006 Joint Entrance Examination (JEE) took position No. 4 when contacted by The Telegraph. He said “some fixed process had to exist” but had no idea what it was.

Replying to the parents’ RTI application last December, five months after the exam was over, the IITs had said there was “no fixed procedure” to determine cut-off marks. That reply was issued by D. Gunasekaran, registrar of IIT Kharagpur, the institute that oversaw the implementation of JEE 2006.

The second answer came five months later after the Central Information Commission (CIC) intervened. The parents were given a definite formula, explained in this newspaper on Monday.

Calculations based on that formula — and checked by this newspaper — show the cut-offs for physics, chemistry and math should have been 22, 26 and 24. But the cut-offs the IITs had actually used were 48, 55 and 37. They had also set an aggregate cut-off of 154.

The explanation to the high court tries to address this problem by offering a slightly amended version: formula II.

According to this, the marks of students who scored zero or less in any subject — the JEE awards negative marks for wrong answers — were not considered while determining the subject cut-offs. This would raise the cut-offs.

But one cannot verify if formula II exactly explains the gap between the official cut-offs and the parents’ cut-offs unless the IITs reveal the marks scored in each subject by all two lakh candidates.

The institutes had flatly refused to do so when the parents asked for it under RTI, later releasing only the marks of the top 32,000 under CIC pressure.

Several independent statisticians told this newspaper that neither formula I (the one provided under CIC pressure) nor formula II “seems feasible”. Both methods could —and probably would — allow the majority of candidates who sat the exam to qualify.

This is because either formula would let in “nearly 70 per cent” of the candidates considered while calculating the subject cut-offs, said Anish Sarkar, who teaches at the Indian Statistical Institute, Delhi.

Ravindra Bapat, who heads ISI Delhi, and his Chennai-based colleague B.L.S. Prakash Rao gave a slightly different figure: “definitely over 50 per cent”.

Since the first formula considers all two lakh who sat the exam (as explained in Monday’s report), this means up to 1.40 lakh students could make the subject cut-offs. The second formula only leaves out those with negative scores, and unless their number runs into several tens of thousands, even this amended procedure would not help.

Since the IITs cannot have known in advance how many students would end up with negative marks, why would they choose this method prior to the exam, the experts asked.

With such huge numbers clearing the subject cut-offs, it would be the aggregate cut-off — based on the around 6,000 seats available — that alone would make the difference.

Why should the IITs then set subject cut-offs at all, saddling themselves with a useless and cumbersome intermediate process, the statisticians asked.

A selection process that initially weeds out less than 50 per cent seems incongruent with the objective of choosing 6,000 students, which is just 3 per cent, they said.

Shishir Dube, who headed the Joint Admission Board that decided the policies for JEE 2006, initially said the cut-offs were set by another body, the Joint Implementation Committee.

When told that all policy matters are decided by the board, the former IIT Kharagpur director agreed that a definite procedure “must” exist.

“But that must have been set before my time (as board chief). We didn’t decide any procedure,” said Dube, now a faculty member at IIT Delhi.

Gunasekaran declined comment.

The Telegraph - Calcutta : Nation