The three-member expert panel, in its report submitted to the Election Commission (EC) on Friday, has recommended an absolute number of polling booths that should be sampled to test the accuracy of EVM results, instead of a fixed percentage or proportion of all polling stations in an Assembly segment or constituency.
The report comes ahead of the Supreme Court hearing next week on a joint petition filed by 23 Opposition parties demanding random verification of EVM count in at least 50 per cent of polling stations every Assembly segment or constituency. In the last hearing, the court had issued a notice to the poll panel, seeking its response to the petition.
The report, The Indian Express has learnt, clarifies that the Commission should depend on sample size (number of polling stations) and not sampling proportion (percentage of polling stations) to test the accuracy of EVM results. “Once the population size (total number of polling stations or EVMs) is large, it doesn’t have a role in determining the sample size. All petitions currently in court seek VVPAT slip tallying in a fixed percentage of polling booths. That is not the correct approach,” said a source.
EC move to get experts bid to build confidence
The decision to compulsorily tally VVPAT slips with EVM count in one polling station per seat was taken by the EC to build confidence among political parties about the accuracy of electronic voting machines. However, the sample size was decided internally without consulting any experts on the matter. In less than a year, after the decision was implemented, parties started questioning the logic behind the Commission’s sample size. In yet another attempt to build confidence, the Commission roped in the Indian Statistical Institute to “arrive at mathematically sound, statistically robust and practically cogent solutions to the issue of a number/percentage of VVPAT slip counts to done during the elections”.
The expert committee that submitted the report to the EC Friday has three members — Abhay G Bhatt, head of ISI Centre in Delhi, Rajeeva L Karandikar, Director of Chennai Mathematical Institute and Onkar Prosad Ghosh, Deputy Director General (Social Statistics Division), Central Statistics Office, Ministry of Statistics and Programme Implementation.
When contacted by The Indian Express Friday, Bhatt and Karandikar refused to comment on the contents of the report on the ground that the matter is currently sub-judice. EC officials too were tight-lipped about the recommendations. However, in a presentation made in August last year to the then Chief Election Commissioner O P Rawat, Karandikar and Bhatt had suggested that mandatory counting of VVPAT slips in sample size of 479 randomly-selected polling stations across the country should be enough for EC to satisfy itself, with 99 per cent confidence, of the integrity and flawlessness of the elections. Based on this presentation, EC’s practice of tallying VVPAT slips with EVM count in one polling station in each seat is satisfactory. However, it’s not clear whether the report submitted Friday is any different from the presentation made to Rawat last year.
The three-member expert committee was asked for its opinion on the Commission’s current practice of mandatory tallying on VVPAT slips in wake of demands by the Opposition parties to increase the sample size.
“ISI is one of the most prominent and reputed national institutions devoted to research, teaching and application of statistics and sampling methodology in the country. Keeping in view their domain expertise and subject specialization, the Commission decided to engage the Institute in arriving at mathematically sound, statistically robust and practically cogent solutions to the issue of a number/percentage of VVPAT slip counts to done during the elections,” according to the EC statement.
The report will be examined by EC next week and a decision on a change in the number of polling stations for mandatory counting of VVPAT slips, if needed, will be taken soon, said sources.