Polls that mislead

Transparency needed in opinion polls, not an outright ban

Once again, pre-election opinion polls are under a critical scanner. Reports have emerged that some of them may be "fixed" - that some pollsters, not all, may be willing to alter the parameters of their polls in order to provide different outcomes. Why would this be a valid consideration in the first place? Well, because every political party in the run-up to a closely contested election tries to convince voters that it is close to winning. In democracies like India's, where a voting bloc's support of a winning party is a crucial step towards ensuring a steady flow of patronage and public goods to the constituents of that bloc, an air of winnability is something many people would pay for. And it appears some people are doing just that. This should not in the least be surprising. However, calls for an outright ban on opinion polls should be resisted. It is never a good idea to ban some form of speech just because it is feared it might influence voters. After all, all forms of political speech are meant to influence voters in a democracy. What makes opinion polls different, it is argued, is that they claim to represent hard data - facts, not speculation. But this, too, is a weak argument for a ban. Try to ban the act of misrepresenting facts to influence voters, and you might end up banning most political campaigns and several notable state and central government advertisements.

The problem is, of course, that predicting election outcomes in India is a tougher job than anywhere else in the world. Some of its largest states have a four-way winner-takes-all system. A swing of a few percentage points, or less, could lead to wild swings in seat counts. Finally, sampling all voters is a continual problem. For example, in Uttar Pradesh, the many Dalit voters of the Bahujan Samaj Party are continually undercounted by pollsters. And, overall, the poorer and more rural voters of the Congress party are less represented in polls than are the richer and more urban voters of its competitors. That's why general elections often have had opinion polls that underestimated the Congress' eventual strength in the Lok Sabha. Like champion athletes who can manipulate every nuance of the sport they're playing, pollsters can fiddle with sampling, with margins of error and with conversions from votes to seats to wind up with the results they - or others - want.

Therefore, the only real answer is transparency, not a ban. Pollsters need to be regulated to enforce openness about their methodology and assumptions, which will allow informed criticism of their numbers. This is in the interest of the polling industry, too. After all, in the absence of such transparency, the market will get to work - and pollsters who can't be trusted will be shunned by media outlets. And if it is difficult to tell an untrustworthy pollster from a trustworthy one, then viewers will begin to ignore all opinion polls - or approach them with the healthy scepticism of a cricket fan watching the Pakistani national team undergoing a batting collapse. A ban doesn't make sense. But open and transparent opinion polls would be an asset to the political process.