Opinion polls should be regulated, not banned. Ideally, it should be self-regulation by pollsters and media organisations.

The debate around the latest proposal to ban opinion polls is an opportunity in disguise. Beneath the familiar acrimony of partisan debates, a much-needed middle ground has emerged quietly. All we need is a group of stakeholders — pollsters, researchers, media heads and political leaders — to come together to turn this possibility into a reality. The time is now.

On the face of it, the debate was anything but rational. This was a familiar parade of tired and untested arguments for and against opinion polls, depending on who hoped to benefit from them in the next few weeks or months. Yet this political posturing and opportunism have had a positive outcome. They have ensured that a flat-footed ban on pre-election opinion polls is unlikely to happen soon. There is thus a window of opportunity for a saner policy intervention.

Systematic collection of views

This new intervention has to begin by recognising three basics. First of all, we must recognise that systematic collection of public opinion is a must in modern democracies. Since elections are not a private act, citizens wish to, and need to, know how others are making up their mind. Survey based tracking of the mood of the electorate performs that crucial role. In an unequal country like India, where a tiny but voluble elite is used to passing off its voice as public interest, scientific sample surveys of public opinions are one of the few ways in which the voice of the poor and the disadvantaged gets registered. All things considered, this is a better method to monitor the popular mood than anything else that exists. This creates a widespread need for this information among politicians, the media and people at large. That is why opinion polls and their use to track the electoral race are here to stay.

Second, it is precisely because opinion polls are the least inaccurate way of assessing the electoral race, making this information public affects the race itself. Of course, you do not win elections by leading in opinion polls; nor does a negative election forecast seal your electoral fate. But it does influence the race in small and, possibly, crucial ways. We do not have conclusive research on this point but the available evidence suggests that there is a small degree of ‘bandwagon effect’ of opinion polls. A party that is seen to be leading in the polls gets some additional support from fence-sitters. This small difference could be decisive in a close contest. More than the voters, opinion poll-based forecasts do affect the morale of party workers and supporters. This makes a big difference during the campaign. The shrill denunciation of opinion polls by political parties is often an illegitimate expression of this legitimate anxiety.

Finally, opinion polls in India have not lived up to the highest standards of professional, rigorous and non-partisan polling. The problem is not that opinion poll-based forecast has been inaccurate. On balance, the record of Indian polls has been quite impressive. While exit polls and post-poll survey based projections have done better than pre-election polls, all forms of polls have proven to be a better guide to electoral prospects than any drawing room or news room gossip.

The real problem with Indian opinion polls, barring some honourable exceptions, lies with their non-transparency and non-professionalism. Unfortunately, there is very little understanding among the common people or even mediapersons of what the polls can and cannot deliver. Pollsters make matters worse by making excessive claims, nothing short of black magic. A general unwillingness on the part of polling agencies and the media to share even basic methodological details about their polls compounds the problem. Most polls get away by announcing the most perfunctory methodological information and making vague claims about the representative nature of the survey. There is thus no way of telling a rogue poll from a professional effort. In the last few years, the proportion of rogue polls has increased. At least some of these, thankfully still a minority, appear to have been cooked inside a room or a studio.

Appropriate intervention

Thus, whatever the motives of the Congress and the BSP, there is some point in demanding an intervention in the business of opinion polls. The real question is: what would be the most appropriate and efficacious intervention? Unfortunately, most of the reformers have little patience and understanding to address this question. A call for a complete ban on pre-election polls, or a ban beginning the day of notification that amounts to the same thing, reflects the knee-jerk response that has come to dominate much of our policymaking. Unfortunately, many well-meaning democratic reform activists and the Election Commission itself have lent their weight to this ill-considered proposition. Banning pre-election opinion polls is a remedy worse than the disease it seeks to cure. There already exists a ban on publishing the findings of polls beginning 48 hours before polling and till the last voter has cast her vote. This is a reasonable restriction, enough to safeguard against manipulations. A full ban for the entire duration of campaign may not stand judicial scrutiny. It is hard to see how such a ban could be presented as a “reasonable restriction” of freedom of expression guaranteed under the Constitution. Besides, it would be very hard to implement. It could either exist only on paper, like the ban on smoking in public places. Or, worse, it could drive all the credible and law-abiding agencies out and leave the field open for rogue polls of fly-by-night operators. In all likelihood, it would open a black market of information where confidential polls and rumours will replace transparent and accountable polling.

Besides, a ban is only a measure of last resort, when all other methods have been tried and found wanting. The amazing thing about the Indian debate on opinion polls is that there has been little effort to explore alternatives to a ban, alternatives that have been successfully used all over the world. The present juncture offers us an opportunity to explore this mature and intelligent response, rather than the ill-informed and knee-jerk demand for a ban. What we need is a regulatory framework for election-related opinion poll — comprising a code of conduct, mandatory disclosures and independent inquiry — to be enforced by an independent agency. Such a framework exists in most of the older democracies and has worked reasonably well.


Here is a suggestion, then, about regulating rather than banning opinion polls. Every election-related poll, or any opinion poll for that matter, must be required to make the following disclosures: the ownership and track record of the organisation carrying out the survey, details of the sponsor; sampling frame, sample size and the exact technique used to draw the sample; the social profile of the achieved sample; where, when and how were the interviews conducted; the exact wording of the question and sequence of questions asked; raw vote shares reported in the survey and how they were converted into vote estimates and seats forecast.

Besides this proactive disclosure, the polling organisation should be required to supply some additional information on demand. This second-order disclosure could include providing basic tables for some key variables. Finally, in case of dispute or challenge, the polling organisation should be required to open its unit level data (raw data file) for in-camera examination by a committee of experts. There could be a provision for strictures and sanctions against those who violate these norms.

Who would formulate these regulations and implement them? Ideally, it should be self-regulation by pollsters and media organisations. An organisation like the News Broadcasters Association or the Press Council of India can take the initiative in this respect. Failing this, the Election Commission could take up this responsibility. What matters is the existence of a regime of mandatory disclosures rather than an agency in charge of implementing it. Once in place, such a mechanism would help the public tell the difference between a genuine and rogue poll and incentivise transparent practices. That would be a significant step forward in democratic public culture. After all, public opinion polling is too valuable and consequential to be left to politicians, or pollsters.

(Yogendra Yadav is a psephologist and a member of the Aam Aadmi Party)

More In: Lead | Opinion