Opinions and anxieties
- CBI sought part RTI exemption, Govt gave it full
- Screen Awards: Milkha, Ram-Leela and Madras Cafe dominate
- DGCA seeks fresh public objections after clearing AirAsia for take-off
- Delhi: 51-year-old Danish national alleges gangrape, 15 detained for questioning
- I wonder if I will be able to ever reunite with my husband, my kids. I miss them: Devyani
There are many reasons why election forecasts made on the basis of opinion polls should be treated with caution. Conducting opinion polls on voting intention in India is incredibly difficult, due to the social complexity, the variation in political context and the sheer logistical practicalities of conducting a survey that is representative of the public. Yet the idea that opinion polls in the run-up to elections should be banned is a real threat to democratic freedom.
Politicians, especially those in government, like to control access to information. Opinion polls provide insights into public attitudes and opinions that often challenge the assertions of party politicians. A robust and transparent process for conducting public opinion surveys is essential for an informed electorate and democratic accountability.
Advocates of a poll ban have focused on the integrity of the organisations that conduct election surveys, suggesting the process is corrupt and outcomes are manipulated. If opinion polls are merely uninformed constructs with no reflection of political reality then surely the right response is to ignore them, not ban them? Another argument suggests that information from opinion polls confuses voters, or in the convoluted language of the Election Commission, "would be a deleterious intrusion into the mind of the voter". This seems to be a rather condescending view of the Indian voter, and at odds with an electorate that appears to be responsive and relatively sophisticated in holding candidates and parties to democratic account.
The EC has also argued that an opinion poll ban is justified as it operates in some of the "advanced democracies of the world", and notes the practice in countries such as Canada, Italy, Turkey and Argentina. Indeed, surveys by the World Association for Public Opinion Research show many countries do prohibit opinion polling, and the number has increased since the 1990s. Yet most countries manage to hold elections with no prohibition of pre-election polling, and this includes most of the consolidated democracies across the world.
A ban has the bizarre effect of allowing any election prediction unless it is based on robust survey practices. This is the situation in Greece and Italy, where an embargo is placed on the publication of forecasts, starting two weeks before election day. After this point, speculation, rumour and insider tips are privileged. Politicians are allowed to make assertions about the state of public opinion and how well their party is doing, astrologers are allowed to predict the outcome, and internet bloggers present their forecasts unchallenged. Yet surveys based on robust methodological practice are suppressed.
There is a recognition that any ban comes at a price, in terms of freedom of information. In Canada, where the publication of opinion surveys is only prohibited on the weekend before voting takes place (always on a Monday), an attempt was made to extend the period of prohibition. This was struck down by the supreme court, which held in a 1998 ruling that "The doubtful benefits of the ban are outweighed by its deleterious effects ... The ban interferes with the rights of voters who want access to the most timely political information available, and with the rights of the media and pollsters who want to provide it".
In the United Kingdom, there is no ban on the publication of opinion polls. Exit polls are carried out, with outcomes released as soon as the polls close. Rather than controlling opinion polls through legislation, there is a process of self-regulation through the British Polling Council's code of practice. This requires companies which undertake public opinion surveys to disclose important details about how the survey was conducted. This includes transparency about who commissioned the survey, the dates when interviewing took place, the basis of the sampling and surveying methods, and the unadjusted results of the survey. Politicians will pick surveys favourable to their party, but everyone can see the basis on which forecasts are made and compare them with other surveys. Pollsters have a reputational interest in conducting their work as accurately as possible.
Opinion polling in the United States is protected by the First Amendment commitment to freedom of speech. Here, the latest innovations in election forecasting take into account the strengths and weaknesses of opinion poll surveys. Analysts such as Nate Silver, whose FiveThirtyEight blog became renowned for its successful prediction of the 2012 presidential election, controlled for the fact that polling organisations used different methods and came up with varying pictures of public opinion. By taking all this information into account, and combining it with an understanding of what made voters likely to turn out and which party they were likely to support, opinion polls could be used to make predictions which were not perfect, but which were much better than the uninformed speculation of politicians, spin doctors, and pundits.
Another innovation in the United States is the use of predictive markets, which are essentially based on betting on the outcomes of elections. Because participants have a financial stake in the outcome, they are expected to respond to changes in the election campaign in ways which, on balance, are likely to reflect the real odds of a candidate winning or a party gaining a particular share of the vote. However, for a market to operate efficiently, there needs to be accurate information, and without opinion polls, predictive markets become less reliable.
Elections in India are exceptional in their size and complexity. Any prediction based on opinion polls, no matter how carefully conducted, has to be interpreted with a great deal of scepticism. Yet if such polls are carried out in a manner that is methodologically robust, they could provide insight into the state of public attitudes, which should be made available through the press and television. It is important that surveys are conducted in a way that is transparent, and it is in the interests of the polling companies that they develop a reputation for reliability and accuracy. Opinion polls play a part in the process of democratic accountability and the process of representation. Censorship should be resisted.
The writer teaches at the University of Sheffield, UK, and is author of 'Standing at the Margins: Representation and Electoral Reservation in India'