Opinion polls face threat of being voted out
Even if opinion polls were to be curbed, the world’s largest democracy would still be an exciting place for elections and psephologists alike. Only a little less democratic. Zia Haq reports. VOTE: Do you think opinion polls should be banned?india Updated: Nov 06, 2013 17:19 IST
Even if opinion polls were to be curbed, the world’s largest democracy would still be an exciting place for elections and psephologists alike. Only a little less democratic. So, pollsters have balked at the Election Commission’s idea of restricting pre-poll surveys. But that’s not the only debate, they say.
Some leading Indian pollsters are worried about losing their precious credibility because of a few rotten apples. For one, there has been a surfeit of poll-predicting organisations. How does the public ensure that their findings are not politically manipulated and methodologies scientifically sound?
Since surveys on the eve of an election may influence voters one way or the other, the country’s poll regulator has proposed banning them from the time elections are notified. The Congress, hemmed in by gloomy predictions from some recent surveys, has endorsed the proposal, while the BJP has fiercely opposed it.
But, experts feel political parties and the poll regulator itself have sidestepped a more vital issue: regulation, which would involve raising professional standards, so that opinion polls are more transparent.
“We don’t need to reinvent the wheel, but follow certain global standards,” said Yogendra Yadav, a pioneering psephologist. The World Association for Public Opinion Research, for instance, has a widely followed model code of ethics.
In Britain, pollsters formed the British Polling Council in 2004 as a self-regulatory body after criticism of unscientific “cowboy” polls. Disclosures on sampling methods, the procedures used, full wording of their questions and answers and sources of funding are now mandatory. Such disclosures preclude the need for ill-thought ban on surveys, Yadav argues.
“We would be open to some kind of regulation,” says psephologist Sanjay Kumar, the co-director of the Delhi-based Lokniti Centre, who has led more than one round of “National Election Study”, India’s best-known and the world’s largest academic survey of voting behaviour.
Most Indian surveys fall in the category of the so-called “swing surveys”. They are aimed at predicting how much “vote share” a party is likely to gain or lose since the last elections. The trickiest part is when vote shares are used to project seats, according to Praveen Rai, who co-authored the book, Measuring Voting Behaviour in India, along with Kumar. India’s first-past-the-post-system, where the candidate with maximum votes wins, is itself ill-suited for seat predictions.
A new rigorous approach in Indian psephology began in the 80s when two economists, Ashok Lahiri and Prannoy Roy, teamed up with British psephologist David Bulter of Nuffield College, Oxford, to focus on the concepts of “swing” and “index of opposition unity”. Both of these gauges predicted election outcomes with the Congress party as a reference point. Now, political choices have become more diverse.
“People should be required to write down their methodology clearly. We are being clubbed with the bad apples,” said Rajiv Karandikar of the Chennai Mathematical Institute, one of the few to have developed a near-proprietary model to convert “vote shares” into seats.
Opinion polls enrich elections, a national carnival in India. It’s a good idea to make the rules of the game clearer.