Why I do not believe in exit polls
Even if the pollsters are sincere about collecting information, large sections of voters do not really tell them who they really voted for
I voted in South Bombay for nearly 20 years up to 2004 when I moved residence to Central Mumbai from Colaba and then Churchgate. The delimitation of constituencies happened after that. So, although I was now a resident in a heavily Shiv Sena area from a previously cosmopolitan one, I found I was still a South Bombay voter for the Lok Sabha constituency, though my assembly constituency was now clearly working class as it was a segment of the textile mill land area which had now several upmarket residential and commercial buildings.
Close to elections n 2009, my domestic help, who was the wife of a former textile mill worker hailing from Andhra Pradesh, warned me that my milkman and newspaperwallah, who were Maharashtrian, would pressurise me to vote for the “dhanush-baan" – the Shiv Sena symbol. They worked for the local shakhas of the Shiv Sena and were tasked with making sure people voted heavily for the party.
“You tell them you will vote for them, just to avoid trouble. Later in the booth you can vote for whoever you want to,” she advised.
I bristled at both my maid and the thought of the milkman dictating my voting choices. Whereupon she got a wise expression on her face and said “Look, Madam, you are a new resident and voter here. Our Telagu Samajam held a meeting at 2am last night and decided we are voting for the ghadi (NCP) or panja j (Congress), whoever stands here. But we are telling those shakha guys hat we are voting for the dhanush-baan.”
That got me curious. Why 2am? And why mislead the Shiv Sena?
The post-midnight meeting was to avoid their radar - as milk and newspaper vendors they had to go to bed early to wake up at the crack of dawn. “Abhi to yeh logan (Sena) bhaiyya log (north Indians) ko maar rahe hain. Kal anna (south Indian) ko bhi maarna chaalu kar denge.” North Indians were then voting heavily for the Congress and the community was afraid they would be targeted too if the Shiv Sena got to know their choices.
There were huge communities of north and south Indians in the constituency in the old residences. The newer buildings had a more cosmopolitan mix and it is the latter which worried the Shiv Sena.
Right on cue, as my maid had warned, in a couple of days, I had the newspaper boy ringing my doorbell and politely asking after my well being. After some small talk he came to the point - have you registered and who are you voting for?
Having been warned by my maid, I was wise enough not to chew his head off. I haven’t decided, it will depend on the candidate, I told him. He knew I was a journalist and dared not bully me. “Please consider the dhanush-baan,” he said unhappily and left it at that as he rang my neighbour’s doorbell where he got a similar response.
That year the Congress candidate, Milind Deora defeated the five-time sitting Sena MP Mohan Rawle – after all the poll surveys, Deora had been unsure of his victory as he thought the huge textile area was stacked against him. But I had an inkling it might not go the way the pollsters had predicted. For my maid had again very informatively told me that they had misguided the surveyors at the exit polls as much as they had the Sena workers in the preceding weeks.
I have stopped believing in exit polls ever since. Even if the pollsters are sincere about collecting information, large sections of voters do not really tell them who they really voted for. The bulk of voting happens in slums and lower middle class areas in the cities which may be reached by these workers. But they find it impossible to penetrate the bastis and hamlets in the rural areas, particularly lower caste settlements and ghettos. So they fail to catch the sentiments of the people who really matter at elections.
Apart from being mistaken for government agents, In the cities and the villages, their suited-booted, laptop-carrying demeanour is most off-putting to the less privileged masses and this class divide too is a debilitating factor in getting their fractions right.
Then, again, from personal experience, I can say even if these poll surveyors are sincere, the agencies that hire them may not always be so. I have had one employee of one agency, seeking to rope in my newspaper as a partner in the opinion and exit polls, tell me confidentially that they could “arrange” the data in a manner suited to my paper’s political line. Even as I was shocked at that dishonesty, I came across a sting by a television channel where a top executive of one such agency repeated the same offer on camera to the reporter approaching them for a poll survey, confirming what the other executive had told had earlier.
But even without these hazards, for the past several elections almost every agency had got their numbers wrong. This year, the fantastic numbers being accorded to the BJP could be a sum of all the hazards plus an additional factor- to demoralise the opposition and to rally the markets before the actual results on May 23.
The reports from the ground do not suggest the BJP could be doing as well as the exit polls have predicted. One of my former editors Vir Sanghvi had once famously said about newspaper readership surveys that they are rather like exit polls - only presumptive. The actual circulation figure is what you get when the ballot boxes are opened. I would turn that on its head to say exit polls are like readership surveys and poll results give you what ultimately matters - the real amount of cash you have collected from the newspapers actually