Tallying votes with VVPAT at one booth in every seat is more than needed, experts tell EC
Governance

Tallying votes with VVPAT at one booth in every seat is more than needed, experts tell EC

Experts from the Indian Statistical Institute have opined that the EC must consider the sample size, not the sample proportion, to verify votes.

   
An election official demonstrates the working of a Voter-Verified Paper Audit Trail | PTI

An election official demonstrates the working of a Voter-Verified Paper Audit Trail | PTI

New Delhi: The Election Commission’s current practice of tallying EVM votes and VVPAT slips at one booth in every Lok Sabha constituency is more than what’s required for statistical accuracy, experts from the Indian Statistical Institute (ISI) have said.

The Election Commission had roped in the experts in view of the opposition’s demands to increase the VVPAT slip count during the elections.

Sources in the ISI told ThePrint that they’ve recommended a number lower than the existing practice. The number cannot be divulged since the matter is sub-judice, but the experts say it would ensure “an extremely high confidence level” of 99.99366 per cent, sources said.


Also read: Statistical body says opposition demand for 50% VVPAT verification not ‘sound’


Sample size crucial, not sample proportion

The Supreme Court is hearing a joint petition filed by 21 opposition parties, which have demanded that 50 per cent of all votes across the country to be tallied with VVPAT slips. The apex court said Monday that it was in favour of increasing the verification.

However, the ISI is learned to have advised the EC that it is not important to have a fixed percentage to be tallied.

Sources in the ISI had earlier told ThePrint that its recommendation was guided by the principle that in any sampling scheme, the accuracy of the estimate depends upon the sample size and not sample proportion.

The ISI report, titled ‘Random Sampling For Testing of EVMs via VVPAT Slip Verification’ is divided into two parts. In the operational part of the report, an optimal sample size is suggested, which would have a two per cent margin of error and a 99.99 percent confidence level. It also proposed a procedure in case of a mismatch being found between VVPAT slip counts and EVM counts in the sampled EVMs.

In the second part of the report, the experts sought to clarify some misconceptions like demanding a sample proportion over sample size.


Also read: Supreme Court wants more random physical verification of VVPAT slips in elections


EC set to cite experts’ opinion in SC

When deputy commissioner Sudeep Jain told a bench comprising of Chief Justice of India Ranjan Gogoi and Justice Deepak Gupta that the current practice of tallying the votes in one booth per constituency is enough, CJI Gogoi asked the EC to submit an affidavit explaining why.

The EC is now expected to cite the ISI report in its affidavit, which is due to be filed Thursday.

The experts’ opinion can be considered a shot in the arm for the poll body, which has maintained that the existing count of VVPAT slips is sufficient to ensure accuracy.