You are using an older browser version. Please use a supported version for the best MSN experience.

Today’s Chanakya gets the exit polls right

LiveMint logoLiveMint 16-05-2014 Vidhi Choudhary

New Delhi: Political research organization Today’s Chanakya came closest to correctly predicting the final tally in the 16th Lok Sabha elections. Hindi news channel News 24, which aired the findings of the exit poll conducted by it, has the bragging rights to the achievement.

The Bharatiya Janata Party (BJP)-led National Democratic Alliance (NDA) won 326 seats in the 543-member Lok Sabha, the BJP on its own winning 283. Today’s Chanakya had projected 340 seats for the NDA and 291 for the BJP.

“When we came up with such numbers, it sounded drastic to most people. Some media organisations and politicos ridiculed us, saying it was bizarre. But I was very sure these numbers were correct and the channel believed we should stick with it,” said Anuradha Prasad, editor-in-chief of News 24.

Most exit and opinion polls got their results horribly wrong in the 2004 and 2009 Lok Sabha elections. This time, however, all exit polls got at least the final outcome right in predicting that the BJP-led NDA would win the elections even if they were not bang on target on the number of seats.

The Congress party-led United Progressive Alliance (UPA) managed to win a meagre 60 seats with 47 for the Congress. Today’s Chanakya had projected 70 seats for the UPA and 57 seats for the Congress.

The total sample size used by the firm was 38,984 people spread over 22 states and one union territory. It used a three-step method involving face-to-face interviews, written interviews and mystery shopping—when anonymous researchers gauge voter sentiment—in calculating the election tally.

“The quality of the sample is more important than the sample size. Parameters like caste, community, religion, rural/urban, age etc go a long way in determining the same,” said V.K. Bajaj, chief executive of Today’s Chanakya.

“It was a combination of events that led to such polarized results. People had been unhappy with the Congress party, they had high expectations which were not met. There was a clear lack of communication from the government. At the same time BJP was doing things right, they had a fantastic campaign, their PM candidate (Narendra Modi) had an example of the Gujrat story to tell,” added Bajaj, explaining the voter sentiment.

It isn’t the first time that Today’s Chanakya has got its numbers right; it had predicted a strong performance by the Aam Aadmi Party (AAP) in last year’s Delhi Assembly elections, giving it 31 seats. The party won 28. Incidentally, it also predicted the return of the UPA government in 2009.

Polls conducted by news broadcasters like NDTV, Times Now, ABP News and India Today Group, among others, predicted that NDA would win between 249 and 289 seats and UPA between 97 and 148 seats.

Hansa Research Pvt. Ltd, which conducted the exit poll for English news channel NDTV, used voter lists for sampling purposes and tablets to conduct interviews across 333 constituencies.

“The main improvement came in the sampling, we got a national sample of 160,000. We are constantly improving methodology to get closer estimates. We were getting a huge voter share for the BJP, but not as large as what the number has come out today,” said Ashok Das, managing director, Hansa Research. Hansa predicted 279 seats for the NDA and 103 for UPA.

According to Shreyas Sardesai, researcher at the Centre for the Study of Developing Societies—Lokniti, which conducted the exit polls for Network 18’s English news channel CNN-IBN, the 2009 verdict was far more complex than this one.

“This time the trend was in one direction and it wasn’t hard to capture that. There was no ambiguity in the results. We got the trend right but the magnitude of victory was underestimated,” said Sardesai.

Nielsen, which conducted the exit polls for ABP news channel, added a secret ballot to the process this year. In the last general elections, the research firm moved to a post-poll survey since exit polls were restrictive in nature and the sampling demographic was not in control of the agency, said Umesh Jha, director at Nielsen India.

“The post-poll surveys have thus helped us pick voters across demographics,” added Jha.

The two-day exercise was carried across 280 constituencies while reaching out to 184,000 voters, roughly 250-275 people per constituency.

All feedback was taken on a fictitious ballot paper by Nielsen’s on-ground representatives, where voters were allowed to tick-mark the candidates they had voted for. Only people whose fingers had been marked with indelible ink to show they had cast their votes were picked for the survey.

Nielsen carried out this method from phase 1 to 8 of the 2014 polls.

More From LiveMint

image beaconimage beaconimage beacon