The exit polls carried out on Sunday after the last vote was cast has predicted a win for the BJP-NDA alliance. Exits polls are the sum total of public opinion and what a voter informs the surveyor after he/she leaves the polling booth. Hence a lot depends on whether the voter actually gave accurate information. Exit polls somehow trample on the idea of the secret ballot. It presumes that the voter is speaking the truth. Exit polls involve several variables that need to be factored in to arrive at a reasonable rationale. Several exit polls in recent years have gone awry. Hence to rely on exit polls is risky business.
It is instructive to note however, that the Sensex jumped up over 1400 points after the exit polls predicted a win for the NDA. Since Sunday, critics belonging to the combined opposition, particularly the Congress party and the TMC leader, Mamata Banerjee have termed the exit polls as being overstated and “gossip.” The most striking example of an exit poll going horrible wrong was the one following the 2004 elections where the NDA was predicted to win between 230 and 275 seats, and the UPA roughly 176 and 190 seats. In the final result the NDA won 187 seats and UPA 219. On its own the BJP won 138 seats and the Congress 145 seats. In the 2009 elections the exit polls managed to get closer to reality. In 2014 the exit polls gave the NDA between 183 and 289 seats, while the UPA tally was projected to be between 92 and 120. When the results were declared on May 16, 2014, the NDA managed 336 seats while the UPA was reduced to just 60 seats. The BJP alone got 282 seats and the Congress plunged to 44 seats.
Exit polls rely heavily on vocal voters but miss out on the silent ones. In a country like India many voters give false responses because of fear of reprisals. However, exit polls have sometimes got it right as in the case of the elections in Assam, West Bengal and Kerala.
The margin of error in an exit poll also called confidence interval, tells us how much we can expect the survey results to reflect the views from the overall population. For instance a 60% “yes” response with a margin of error of 5% means that between 55% and 65% of the general population think that the answer is “yes.” Hence there is some amount of extrapolation which again relies on whether the informant is telling the surveyor the truth. Be that as it may, the NDA and its faithful are in a celebratory mood.