Exit polls in India are often treated as an election-night tamasha. This is not the case in many developed democracies, where the science and art of studying elections is becoming more refined and rigorous with every cycle. When polls fail there, it becomes a national scandal. In the UK, following a series of high-profile failures, such as the 2017 general elections and the 2016 EU referendum, the House of Lords established a Select Committee on Political Polling in June 2017. The committee’s final report in April 2018 called for tighter oversight of the polling industry.

If anything, India is moving in the opposite direction. Standards in the 1990s and 2000s were arguably much more robust. The tamasha has now reached such a low level in this round of assembly elections that any serious researcher of electoral politics would bow their head in shame. It is telling that even prominent voices in TV news have begun arguing that exit polls must exit if they do not disclose their methodology, warning against reducing them to a prime-time lottery with zero accountability. Election polling data is extremely crucial to understand the nature of Indian democracy for multiple reasons. In many conflict-ridden democracies, independent polling data acts as a credible device to check the robustness of the election process. In established democracies such as India, they help us in understanding the pulse of people regarding not just who they are voting for, but what is shaping their vote choice. The 2024 general election significantly undermined confidence in this enterprise. Not because pollsters got the numbers wrong — we have seen that before —but because there was no serious, scientific explanation of why they went wrong. No diagnosis. No attempt to improve the craft. An Aug 2016 issue of Seminar that I curated on the state of election surveys (available online) examined the challenges polling faces not just in India, but globally. The concerns of contributors remain relevant, but the institutional response in India has been strikingly absent.

Since then, the track record of even the best pollsters has not been modest to say the least. In this round, some of the more credible organisations were notably absent. Credibility in polling doesn’t mean always getting predictions right, but rather providing a lot of details regarding how their surveys are conducted. Details about sample size, sampling design, geographical spread, representativeness, weighting, and mode of interviewing are essential to evaluate reliability. Without these, projected numbers are not data but mere speculation.

Exit polls, and even more so post-poll surveys, are important not merely to forecast winners, but to understand the social and political basis of electoral mandates. They provide granular insights into voting behaviour that actual results alone cannot. Conducting such surveys is expensive. Costs depend on sample size, interview length, geographic spread, and fieldwork effort. Yet, many exit polls today do not even disclose who funded them. In the past, most exit polls were commissioned by established media houses. In the last few years, a proliferation of little-known firms has produced a flood of forecasts with no transparency and no scrutiny. This raises a basic question: who is commissioning these surveys, and why?

More importantly, India’s top TV networks are complicit in this decline. Instead of investing in credible, independent polling, many have outsourced their responsibility. If media organisations verify facts before publishing a story, why do they air exit poll numbers with no methodological disclosure? Why are viewers expected to trust surveys conducted by unknown entities? The mushrooming of polling firms with little track record has gone largely unquestioned. Is it not the responsibility of the media to ensure transparency? Should they not invest in building credible polling infrastructure? And why is polling confined only to elections? In many democracies, regular surveys track public opinion on key issues, offering a continuous understanding of citizen sentiment. In India, the absence of such investment has reduced polling to an episodic spectacle.

Then there are the ‘poll of polls’. In statistics, averaging improves accuracy only when errors are random and independent. In political polling, they are often neither. When a high-quality survey is averaged with a poorly designed one, the result is not balance, but distortion. Bad data does not cancel out; it contaminates. Moreover, aggregation encourages ‘herding’. Pollsters, wary of being outliers, adjust their estimates to align with prevailing trends. What emerges is a false consensus —not because multiple surveys independently converge on the projected estimate, but because they converge on each other.

Opinion polls and exit polls are vital tools to understand the health of a democracy. Reducing them to spectacle is a disservice. India does not merely have the problem of inaccurate poll projections, but a collapsing polling ecosystem marked by opacity, weak incentives, and declining standards. If this trajectory continues, exit polls will not just fail to predict elections but will fail the very idea of democratic accountability.

END OF ARTICLE