Exit polls get it wrong, again

Many times bitten, but never shy – television pollsters again called an election wrong on Sunday. After they went off the mark in Uttar Pradesh and Punjab earlier this year, the exit polls this time failed to assess the strength and sweep of Narendra Modi’s victory in Gujarat.

None of the exit polls gave the BJP more than 110 seats, down 17 seats from 127 it won in the last polls in 2002. The BJP ended up winning 117, retaining power comfortably in what was being seen as a close election. Congress could manage only a minor improvement, winning 62.

NDTV, which surveyed 18,100 voters in 61 constituencies, predicted the BJP would win 90-110 seats, the Congress 70-95, and others between three and five. CNN-IBN surveyed over 9,000 voters in 60 constituencies and predicted 92-100 seats for the BJP and 77-85 for the Congress. The Star News-Nielsen exit poll showed BJP getting 103 while Zee News and C-Voter poll projected BJP winning 93-104 seats to Congress’s 75-87.

The only pollster who got close to reality was GVL Narasimha Rao, who predicted BJP would win at least 115 seats in his column in Mint, a sister publication of Hindustan Times, of December 10.

Analysts said market researchers conducting the polls were unable to catch the social trends and the influence of bread and butter issues on voting, which introduced margin for errors in the forecast.

For example, in Saurashtra, a boost in agriculture and industrial progress helped voters make up their minds in Modi’s favour, said Mahesh Rangarajan, political analyst and a professor at Delhi University.

“They have got the winner and the runners up correct. But translating voter mood swings into seats is a very tricky affair, as also it is very complicated. With a three per cent swing you may get 120 seats and may even get 100 seats,” he said.

Chandrabhan Prasad, a political analyst who tracks Dalit issues, said pollsters did not take into account social factors affecting voters. “The voter is becoming very cunning and gives out information the pollster is seeking. The pollsters are not capable enough to overcome this guile and cunning of the Indian voter and capture the real mood.”

In Gujarat, he said, there was a massive consolidation of non-Patels behind Modi as a result of his social engineering. This aspect was not factored in the exit polls.