In this final phase, 13 seats each in Punjab and Uttar Pradesh while 9 seats in West Bengal will go to polls. Now in this last phase, is any exit poll needed to say that PM Narendra Modi will be elected from Varanasi constituency?
But still, once the official polling time ends on June 1, many news channels will start showing the exit polls. Predictions will start and so on. Exit polls provide an estimate of how people voted in an election.
It is based on interactions with voters after they exit the polling booths, as well as other statistics related to voter data. In fact, Indians place as much importance on exit polls as they do on actual results. The reason exit polls are released on the last day of polling is because the Election Commission of India has mandated such exit polling agencies to wait until all phases of polling are over to avoid influencing voters who want to or are yet to vote.
Most of the time, the accuracy of exit polls is tested by an individual’s opinion about political parties. This is likely to be repeated when exit polls for the ongoing Lok Sabha elections are released on Saturday, June 1 evening. Importantly, many exit polls have also proven unreliable in recent years, producing conflicting results. Last year, many polls predicted wrong winners in Chhattisgarh and Madhya Pradesh assembly elections, while some were far from the reality in Rajasthan. An agency may have predicted Madhya Pradesh results correctly, but erred in Rajasthan. There were states where all exit polls were right (Telangana) and others where all were wrong (Chhattisgarh).
In such a situation, how to read the exit polls coming out on Saturday evening? How to determine their accuracy? Today, few people judge the accuracy of an exit poll by the agency that conducted the survey or the television channel that commissioned it. Others look at sample size. A general perception is that the larger the sample size, the more reliable the exit poll. In reality, these should not be indicators to check the accuracy of exit polls, which depend on many different factors. Exit polls are based on structured questionnaires that are used to collect data after interacting with a large number of respondents. This method is not new. It began in 1957 during the second Lok Sabha elections when the Indian Institute of Public Opinion conducted a survey.
But even the best guesses or presumptions cannot exclude necessary procedures. Without a structured questionnaire, data cannot be collected coherently or analyzed systematically. Since exit polls began in 1957, at least one aspect has improved tremendously, i.e. the sample size. Gone are the days when a national sample of 20,000 to 30,000 respondents was considered large. Today, survey agencies in the country are conducting exit polls with samples as large as 1 million. Exit polls of a few hundred thousand samples have become quite mediocre today. Large sample size is important in exit polls, but more important than sample size is how representative it is if it reflects a variety of voter profiles. Call-backs of respondents, images of being interviewed and phone calls from the field, WhatsApp groups and similar tools have helped to overcome such shortcomings, although there is no rule of thumb on how to estimate. Estimates are made and seats are predicted based on the results of previous elections. Calculating vote share is not an easy task given India’s diversity of location, caste, religion, language, education level, economic class – all of which affect voting behaviour. Over or under representation of any of these diverse segments of the electorate can affect the accuracy of estimates.
There are other difficulties as well. Since the swing model is applied to previous vote shares, changes in coalitions, or splits or mergers of parties between two elections, become difficult. For example, BJP and JDU in Bihar. When competition is limited to two parties, the measurement of swings and electoral change becomes easier. As more political parties and candidates are added, the complexity of the swing increases. Innovative exit polls are likely to be more accurate than estimates made using traditional methods. But exit polls using traditional methods predict vote share and help analyse voting behaviour based on different socio-economic backgrounds. Counting methods do not predict vote share, and any systematic analysis of voter behaviour can only be a pipe dream. Many exit polls predict the number of seats, but no vote share or no systematic details. Should we consider this as an exit poll? So now there is a need to distinguish between actual exit polls and predictions.
