AI generated image showing an abstract representation of the discrepancies between the Exit Poll results and the actual results declared on June 4.
The marathon elections are finally past us. But not without a result not many saw coming when the elections season started three months ago. The results on June 4 came as a surprise to many, particularly because a plethora of exit polls funded by mainline news channels of the country had predicted a resounding victory for the Narendra Modi-led NDA, with almost all of them giving the alliance 350+ seats, and some even predicting ‘400 paar’.
As we know now, even ‘300 par’ didn’t materialise. Since then, there has been intense debate about exit polls, and whether their inaccurate predictions are simply a case of incompetence on the part of multiple agencies, or a result of malafide intent to influence the stock market. I have been asked this question more than a few times over the last 10 days, in my capacity as the head of a media research firm.
While I’m no stockmarket expert, the scam allegations seem a bit far-fetched. It would take multiple agencies to comply with the perpetrators of the scam, put their reputation on the line, and hope that they make some illegal money off it. The history of scams suggest they are ‘designed’ in a way that they are under the radar, far from the public eye. Exit polls were anything but that. In any case, I don’t see how we will know any more on this topic anytime soon.
The incompetence argument is a lot more persuasive one. Despite large sample sizes (while not all polls reveal their methodology and sample design, some do), and the claim to have covered all 543 constituencies, and represented different demographic segments adequately, how do so many polls get it wrong? In my opinion, the incompetence doesn’t lie in their ability to conduct field surveys, but in their lack of confidence to look at the findings dispassionately. All quantitative research that comes with the responsibility of predicting an outcome will operate on ranges, rather than exact numbers. And it’s now evident that the exit polls were leaned towards the higher ends of their ranges, and probably stretched them further. One poll made sure its upper end was 401, and another went for the round figure: 400!
Over the course of this year’s election coverage, we have seen many pollsters become election experts, going beyond analysing their data, and entering domains of political analysis that’s best left to journalists with their ears to the ground. Bafflingly, many editors of mainline news channels have encouraged this, by giving pollsters a platform on their shows every night, even during the period when the Election Commission embargo on exit poll results was in play.
It may be hard to resist fame, but if it comes at the cost of objectivity, a pollster must examine if it’s worth it. As it is, our news channels operate like echo chambers, and it is hard to not get influenced by their narratives if they are platforming you as an important talking head.
So, the pollsters have gone wrong in their minds, probably working backwards from a pre-decided outcome they talked themselves into believing. It’s impossible to say what conclusions a more objective analysis on their data would have led them to conclude.
In any case, exit polls seem to quite a wasteful indulgence for news channels. But it’s a vicious loop, because no channel wants to miss out on an evening’s hype. Just like no pollster wants to be the only one who got it wrong!