Exit Polls and Exact Polls
Minutes after the buttons on the Electronic Voting Machines were pressed for the last time on May 19 in the Indian national elections, the people were flooded with a flurry of exit polls broadcast over TV channels. This was accompanied by high-decibel debates in TV studios analysing the exit poll forecasts in spite of the recognition that they are not exact official results as vote-count takes place on May 23.
After a 39-day voting process, suspense has been building up about the possible list of victors and the vanquished. Naturally, Indians were glued to their TV sets for the exit polls, trying to get a sense of what kind of government may be formed. But the question is: did they get an accurate picture?
The reactions of the parties on both sides of the political divide to the exit polls have been on predictable lines: the opposition dismissing them with the lone discordant voice being that of National Conference leader Omar Abdullah, who opined that not all exit polls are wrong, while senior Bharatiya Janata Party leader Arun Jaitley said the actual results will be "in consonance" with the exit poll results, almost all of which predict that the BJP-led National Democratic Alliance (NDA) will get a comfortable majority for a second successive term.
How reliable are the exit polls? To what degree are these polls credible? Should one go by the brand name of the agencies the TV channels hire to conduct the exit polls or of the channels that put them out? Poll experts tend to believe that a lot of the reliability of the exit polls depends on the sample size of voters who are interviewed about their choice of candidates soon after coming out of the polling booths. However, what is the guarantee that the voters have spoken the truth to the survey agency as to who they have voted for, even if their identities are not disclosed? It is quite possible that the selection of voters interviewed by exit poll agencies depends on the latter's choices as per their biases towards the candidates and parties they would like to win. Should one reject outright the exit polls or take them with a pinch of salt? These questions are directed against the very basis of exit polls and raise a fresh debate about their desirability.
Sanjay Kumar, Director of Centre for the Study of Developing Societies (CSDS), a New Delhi-based think-tank, writes in an article in The Indian Express: "The science of surveys, which includes exit polls, works on the assumption that the data have been collected after interviewing a large number of voters using a structured questionnaire." Kumar points to some basic requirements of surveys: (1) a big sample size of the electorate and (2) a structured questionnaire.
Indian exit polls are known to often be way off the mark—failing to predict the final result with a fair degree of accuracy. However, there have also been occasions when they have succeeded in pointing towards a broad direction as to which way the wind is blowing as revealed later by the official results.
Indian Vice President M Venkaiah Naidu, a senior leader of the BJP, was quoted as saying just a few days ago that most of the exit polls since 1990 have turned out to be wrong. In 2004, all exit polls had forecast a majority for the BJP-led combine. But the actual results led to a hung parliament with the Congress emerging as the single largest party and forming an alliance government under Manmohan Singh as prime minister. In the national poll in 2014, most of the exit polls had predicted the BJP to win but none of them projected the BJP to get a majority on its own. In fact, five years ago, there was such a strong wave in favour of the BJP that one did not even need the exit polls to predict the election outcome.
According to Kumar, the methodology of conducting exit polls in India has evolved since it began in 1957, with the sample size of voters having expanded from the thousands used by well-known psephologists like Prannoy Roy and Yogendra Yadav through the 1980s and 1990s to lakhs at present. But Kumar also says the CSDS exit polls were not off the mark in the 1998 and 1999 parliamentary elections. At the same time, he admits there were occasions when CSDS exit polls were inaccurate (for instance, wrongly predicting the winners in Chhattisgarh assembly polls late last year when the Congress Party trumped the Bharatiya Janata Party). In the case of Uttar Pradesh assembly polls in 2017, the actual results proved all exit polls wrong as far as BJP's stunning victory was concerned.
From these examples, Kumar comes to the conclusion that "there is no thumb rule for how to get the (exit poll) prediction correct." If that is the case, should we give a thumbs up or thumbs down to the trend of exit polls?
Pallab Bhattacharya is a special correspondent for The Daily Star.