As a casual observer of the news media, in the past few days I have been surprised to see the eruption of accolade for Modi’s (supposed) win in Gujarat. The electronic and print media alike have arrived at the shared conclusion that the BJP is certain to – in fact, it already has – won Gujarat with a 120-140 seats. What is the basis of all this jubilation? Exit polls. This is India’s “exit poll” moment – the moment when the media, in its eternal folly, initially goes overboard over some new way of forecasting the future and later backtracks when its predictions turn out to be very different from the reality.
The US media has experienced such gaffes several times before in various Presidential elections (trust me, or Google it!) and having learned the bitter truth about the inaccuracy of exit polls, is now much more cautious in using exit poll data to make electoral determinations.
There are several reasons that exit poll data cannot be considered reliable and especially in the present situation. First – exit polls are voluntary. It cannot be otherwise without violating the sanctity of a voter’s right to privacy over his/her electoral choices. Second – by definition, these polls are conducted on voters after they exit the voting booth, having already cast their ballot. Most voters are just happy to go home after having endured long waits to cast their vote in an election with as big a turnout (71%) as this one. Only a small number are likely to stick around to answer the exit pollster questions. There is no guarantee that those who do take the exit poll, constitute a fair and representative sample of all the voters who come to a particular election booth. Third and last, there is no way to determine how close the answers given by a voter to an exit pollster are to the choices that voter has just made. In other words, there is no way to check whether the voter is telling the truth. Voters who cast their ballots for a party which happens to be in the minority at present, are unlikely to be completely honest about their electoral choices one they leave the voting booth.
Thus there is more than enough reason to be cautious about taking the predictions of exit polls too literally. So I’m not surprised that those who actually have the incentive to get their predictions right – i.e., those in the business of making monetary bets over electoral various outcomes – should reach a different conclusion from exit polls. If I were a gambling man, I would think very carefully before placing my bets in accordance with the predictions of exit poll data.
Freebits 