Why Exit Polls So Often Get It Wrong And Why We Still Believe Them

‘The truth is rarely pure and never simple.’ - Oscar Wilde

In every election season, when the last vote is cast and the electronic beeps fall silent, a peculiar suspense grips the country. It is the interval between faith and fact, between the act of choosing and the act of counting. Into that silence rush the exit polls, armed with percentages, pie charts, and promise.

This year in Bihar, as the second and final phase of polling concluded, the familiar ritual began again. Channels went live, analysts dusted their bar graphs, and the screen filled with numbers predicting who might form the next government. Most of them pointed towards a comfortable return for the National Democratic Alliance (NDA); others left a small window open for the Mahagathbandhan. And yet, under that bright studio light, everyone knew a quieter truth: the exit poll is not the election.

What Exit Polls Really Are

An exit poll is meant to be a mirror, quick, statistical, and seemingly precise. It asks voters, as they walk out of the polling booth, one simple question: whom did you vote for? The answers are then expanded through mathematical modelling to estimate how an entire state or nation might have voted. It looks scientific. It sounds certain. But the certainty is often illusory.

In truth, the science of exit polling rests on fragile ground. It depends on honesty of responses, representativeness of samples, and neutrality of interpretation, three conditions rarely guaranteed in India’s vast and uneven democracy. Here, elections are not just political contests; they are emotional reckonings. They carry the weight of identity, caste, kinship, grievance, and hope. The voter’s choice is often private, sometimes protective, occasionally deceptive.

India’s tryst with exit polls began tentatively in the 1970s, but they came into their own with the rise of private television in the 1990s, when journalists like Vinod Dua and Prannoy Roy transformed election coverage into a national spectacle. Their broadcasts turned numbers into narratives, introducing viewers to the grammar of psephology and making the act of voting part of popular culture. Yet, the intellectual rigour behind this practice came from the Centre for the Study of Developing Societies (CSDS), where Yogendra Yadav and later Sanjay Kumar built the Lokniti programme into one of India’s most credible research networks. They trained field investigators, constructed rural samples, and refined the post-poll method to explore not only who people voted for but why they did so.

If Dua and Roy gave Indian elections their voice on television, Yadav and Kumar gave them their vocabulary in social science, together shaping a field that turned political gossip into grounded understanding.

But the rise of 24-hour news changed everything. Television needed instant conclusions; elections needed interpretation on the run. The exit poll became a spectacle, a race for who could get closest to the final number before the actual count began. Research agencies such as C-Voter, Axis My India, Today’s Chanakya, and News18 Mega Poll began to dominate this field, deploying thousands of enumerators and algorithmic models. Yet, despite their sophistication, the basic fault lines persisted.

India’s multi-phase voting, its layered caste coalitions, regional identities, and linguistic silos make it one of the hardest countries in the world to poll. A booth in Aurangabad cannot explain the mood in Arrah; a swing in Siwan says little about Saharsa. Even a perfectly designed questionnaire can falter against the hesitation of a villager who believes revealing his vote might invite trouble, or a city voter who offers a false answer out of amusement or distrust.

Exit polls, in theory, claim to read the people’s verdict before it is declared. In practice, they capture the nervous weather of democracy, a climate that changes by the hour. The numbers they project are not false; they are simply incomplete. They reflect a fraction of the country, at a particular hour, under a particular emotion.

And yet, every election night, we turn to them again, not because they are infallible, but because they offer something rarer: a glimpse into the political psyche of a restless republic. Exit polls tell us less about who will win, and more about how uncertain our democracy still is, vast, vibrant, and ultimately beyond the grasp of neat prediction.

When Numbers Lied

The history of Indian elections is also a history of misplaced forecasts.

In 2023, Chhattisgarh, almost every major survey predicted a smooth victory for the Congress. When the votes were counted, the BJP swept back to power with more than 50 seats, a stunning reversal that humbled pollsters and pundits alike.

In Delhi, 2015, the exit polls anticipated a strong showing by the Aam Aadmi Party (AAP) but none imagined the tidal wave that followed. AAP won 67 out of 70 seats, a mandate that mocked mathematical probability.

Bihar, 2015 told a similar story. The exit polls projected a neck-and-neck contest between the NDA and the Mahagathbandhan. The actual results delivered a landslide to Nitish Kumar and Lalu Prasad Yadav’s alliance, leaving analysts searching for explanations.

Even the 2024 Lok Sabha election, the largest democratic exercise in history, exposed the fragility of political forecasting. Most exit polls saw the NDA soaring past 400 seats. The final tally stopped at 293, with the BJP alone dropping to 240. The opposition’s INDIA alliance performed far better than anticipated, proving once again that numbers can echo, but they do not always tell the truth.

And in Haryana, 2024, the exit polls handed the Congress a majority on paper. When the counting ended, the BJP had captured 48 seats, and the Congress fell short.

Perhaps the most dramatic of all was Uttar Pradesh, 2017. Every major poll pointed to a hung assembly. The BJP stormed to over 300 seats, a landslide that rewrote both state politics and the credibility of surveys.

Each of these episodes reminds us that exit polls are less about prophecy and more about psychology, a mirror reflecting what we wish to see, not always what is.

The Thrill and the Trap

Why, then, do we keep believing?

Because politics, like sport, feeds on suspense. Exit polls are the teaser trailers of democracy. They promise a glimpse of the climax before the curtain lifts. The voter, the party worker, the anchor, all crave early reassurance. In that collective impatience, exit polls thrive.

Television networks need drama; data gives it shape. Pollsters need visibility; predictions bring them fame. Political parties need momentum; favourable surveys provide morale. Everyone, in some way, benefits from the illusion of certainty.

But beneath the thrill lies a trap. When exit polls become the main story, the real story, the patience of counting, the dignity of verification, begins to fade. The spectacle of prediction overtakes the substance of process.

Between Arithmetic and Emotion

At heart, elections are not arithmetic; they are emotion disguised as arithmetic. A percentage can measure turnout but not yearning. A sample can capture caste, but not conviction. The Indian voter is not a data point; she is a mood, shifting quietly between loyalty and loss, memory and hope.

This is why even the best-designed exit poll falters. It cannot see the late-night conversation that changed a vote; it cannot read the hesitation of a first-time voter who pressed a button and then walked away thinking differently. The distance between the polling booth and the pollster’s assumption is not just geographical, it is psychological.

The Bihar Wait

As Bihar heads into its verdict day on 14 November 2025, the familiar tension builds again. The exit polls are out, their numbers already circulating across channels and social media threads. Predictions are made, seats tallied, alliances imagined. And yet, the wiser heads are waiting.

They remember 2015 and 2020, the years when Bihar taught India that it has its own arithmetic, often beyond the comprehension of television studios. They know that between a survey’s microphone and a voter’s silence lies a vast territory of human unpredictability.

When the counting begins, the only numbers that matter will be the ones emerging from the district centres, handwritten, verified, and final.

The Humility of Uncertainty

In the end, exit polls are not the problem; our faith in them is. They are snapshots, not verdicts, approximations in a land that defies neat equations. Their value lies in provoking debate, not in deciding destiny.

The wiser approach, perhaps, is to treat them as weather forecasts, useful for conversation, but not for travel plans. The real monsoon of democracy arrives only with the last count.

‘The truth is rarely pure and never simple.’ Wilde’s words feel truer than ever in the age of televised politics. The truth of democracy lies not in the projections we chase, but in the patience we keep.

When the EVMs open on the morning of November 14, Bihar will speak again, quietly, decisively, beyond the noise of prediction. And that, not the exit poll, will be the only truth that counts.

(Ashutosh Kumar Thakur writes regularly on society, literature, and the arts, reflecting on the shared histories and cultures of South Asia.)