Why the Election Commission doubts opinion polls
While opinion polls offer inferences (that may affect voters, especially undecided ones), they are often neither scientific nor transparent about their sampling methods.ht view Updated: Apr 28, 2014 11:09 IST
When elections are upon us, opinion polls begin to proliferate. We are now in the midst of the electoral battle and so it is not surprising that we readers and viewers are confronted with opinion poll findings almost every other day. I find that those who conduct opinion polls fail to share vital scientific data which is mandatory for such sampling in many other democracies.
Why are opinion polls held? Are they an accurate assessment of public opinion, especially, in predicting electoral outcomes? The answer is that while they offer inferences (that may affect voters, especially undecided ones), they are often neither scientific nor transparent about their sampling methods.
This is why most developed countries restrict their telecast during the elections. Some countries ban them altogether, while others lay down stringent conditions for compliance.
From my tenure in the Election Commission, both as Election Commissioner and later as Chief Election Commissioner, I began to look out for the hidden fine print. First and foremost, vital data about sample size is most often not disclosed. Is it not critical to know whether the sample size considered is 1,000 or 5,000 or 20,000 or a significantly larger figure?
In a country, where we have now over 816 million names on the electoral register, even a sample size of 100,000 is relatively small. But if the sample size increases or becomes geographically widespread then the cost of sampling also rises. This is one reason why details are often deliberately not shared. Some opinion polls do play safe by offering the name of a sampling agency. A few others direct us to their websites if we want to glean these vital details. But very few share their data upfront.
Equally important is to know who have been asked questions and where. Is it an all-India survey or confined to some states or cities? Another factor to bear in mind regarding its methodology — whether the sampling has been door-to -door (which is expensive), or taken by phone-ins. It is instructive to know which age group or gender has been sampled, for women’s preferences may be different from those of men.
Again, is it just a random sample with people asked a question or two as they exit from a metro or bus station? I am not deprecating any of these methods, but seeking to assess how scientifically rooted the sample is, before making up my own mind whether to believe its findings or dismiss them with a pinch of salt.
Unfortunately, most often these details are not shared. Adding to the confusion is the fact that some of these polls are completely fraudulent, as a recent sting operation revealed. This sting operation made it amply clear that the cases it highlighted were those of ‘paid news’ or intended to illicitly subserve a particular party’s or candidate’s political interests. Such bogus opinion polls certainly do not serve the interests of ethical journalism and disturb the level playing field that the Election Commission strives to maintain.
There are other factors that also make opinion polls hazardous in our country. Apart from the sheer numbers of the electorate and the unscientific nature of our pollsters and samplers, there are other complex factors at play.
We are a multi-layered society where caste, religion, region and class play very significant roles. In urban areas, voters are less reticent about truthfully sharing their preferences, but that is not the case with communities who feel disadvantaged and vulnerable if they speak frankly, fearing reprisal of some kind.
Moreover, in a multi-cornered contest, where regional parties also engage, opinion polls need very careful sampling, indeed, if the predicted outcomes are not to appear similar to an astrologer’s guesswork or worse.
From my tenure in the Election Commission and beyond, I have found that most political players who have been consulted by the Commission from time to time have been wary of endorsing such polls. The Commission is after all an umpire in the poll process and the parties and candidates are the players. This is why it becomes essential to consult all recognised political parties regarding their views on the subject.
Before the start of the present poll process, the Election Commission sought the opinion of all recognised political parties. Of the 15 parties which responded, all but one opposed opinion polls. The Commission has since submitted a proposal to the government to amend the Representation of the People’s Act, 1951, to legally ban opinion polls.
The present legal status, however, is rather confusing. There is no ban on opinion polls in the print media. There is a ban on opinion polls on television during 48 hours before the poll. However, there a legal ban on exit polls, which cannot be conducted until the last phase of the polling is not completed.
This is why the Commission issued a warning on April 16 to the media, reminding them that they are banned from telecasting or printing opinion polls relating to the constituencies where polling has already taken place, for in such cases an opinion poll in effect gets converted into an exit poll, which is an electoral offence.
A final word. I am not against genuine opinion polls, but I am certainly against those polls that choose to hide all essential sampling data, or where ‘paid news’ or a channel or newspaper’s hidden agenda is masqueraded before the public as genuine.
Navin B Chawla is former Chief Election Commissioner
The views expressed by the author are personal