Beyond the exit poll: Can SCORE restore faith in India's election surveys?

SCORE is a public-opinion platform launched in Bengaluru on October 16. It plans to measure political attitudes constituency by constituency and make the findings available directly to citizens. This differs from a typical election-time opinion poll, which offers a snapshot before voting, or an exit poll, which questions voters after they cast their ballots.

Its first exercise covers all 117 Punjab Assembly constituencies. Researchers interviewed more than 28,000 voters between September 13 and October 13, 2026. SCORE plans to publish a new Punjab tracker on the 15th of every month. It will also begin an Uttar Pradesh tracker on November 20.

The main difference is continuity and detail. Repeated surveys can show whether views are changing and whether a finding lasts over time. SCORE also intends to track assessments of MLAs, public services, and government schemes, giving voters more information than a simple seat projection. Repetition alone, however, does not guarantee reliable research.

The first Punjab survey put the ruling Aam Aadmi Party, or AAP, in a strong electoral position. It estimated that AAP would win between 69 and 82 of the state's 117 Assembly seats if voting had taken place on October 15. Punjab's majority mark is 59 seats, so the projected range would have given AAP a comfortable majority.

The survey also estimated AAP's vote share at approximately 38%. That was lower than the party's performance in the 2022 Assembly election, when it won 92 seats. The comparison suggests continued strength, but not a repeat of its earlier seat total. The article's supplied text ends before giving the survey's estimates for other major parties.

This result is presented as a verdict before the verdict, rather than a final prediction. Monthly tracking could show whether AAP's lead expands or narrows. It could also test whether voter concerns about governance affect the party's position over time. The article does not provide enough figures to compare every party.

SCORE's Punjab exercise is statewide and unusually broad in geographic coverage. It includes all 117 Assembly constituencies, rather than treating Punjab as one single electoral area. Researchers interviewed more than 28,000 voters between September 13 and October 13, 2026. This gives the platform a large base for examining local differences.

The planned schedule is also central to the project. SCORE intends to publish a fresh Punjab tracker on the 15th of every month. Results are expected to be available through its website and WhatsApp. An Uttar Pradesh tracker is scheduled to begin on November 20. The platform says its interface will make the results accessible and easy to analyse.

This scale can help reveal constituency-level variation that statewide headlines may hide. Regular measurement can also identify movement between surveys and test earlier findings. Yet a large sample and frequent polling do not automatically make results accurate. Survey design, sampling, questioning, and transparent methods still determine how much confidence the findings deserve.

A survey focused only on the likely winner gives voters a narrow picture of democracy. Tracking views about MLAs, public services, and government schemes can connect electoral preferences with everyday experience. It can show what constituents consider inadequate and whether promised benefits are reaching people.

For example, a constituency may support a party while expressing dissatisfaction with local services. Another may report that a government scheme is reaching beneficiaries, even if television debate focuses on a different issue. Repeated measurements can show whether these views are temporary reactions or lasting concerns. They can also reveal changes between constituencies instead of reducing Punjab to one statewide number.

That information could improve public discussion and give citizens more material for judging representatives. It may also encourage attention to performance between elections. SCORE's potential democratic value therefore lies less in naming a future Chief Minister than in supplying usable evidence. The benefit will depend on careful design, reliable data, and clear disclosure of methods.

India has many election predictions, but confidence in the organisations making them has weakened. The article says surveys have become television events rather than serious efforts to understand society. Credible researchers must compete with lesser-known operators that offer extravagant forecasts without adequately explaining their methods.

When sampling, questions, fieldwork, or calculations are not disclosed, audiences cannot assess how a result was produced. Pressure to convert complex findings into definite seat projections creates another problem. A survey may contain nuanced evidence about attitudes, identities, or local concerns, but television coverage can reduce it to a headline about who will win. Repeated misses then damage trust in polling more broadly.

SCORE responds by promising direct public access and regular constituency-level information. That may help if its methods are transparent and its findings are presented with appropriate uncertainty. The article also warns that monthly polling is not automatically more scientific. Credibility must come from sound design and openness, not frequency, polish, or dramatic predictions.

India already has a substantial tradition of electoral research. Eric D Costa and Prannoy Roy helped bring scientific election analysis to a wider audience. CSDS, especially through its Lokniti programme, studied electoral behaviour, social identities, political participation, and changing public attitudes. Yogendra Yadav and Sanjay Kumar helped emphasise why people vote, not only which party wins.

Commercial organisations such as CVoter and Axis My India later expanded polling's visibility and reach. Some built large field networks and tracking exercises. SCORE's stated contribution is to combine regular tracking with granular constituency data and public access in one platform. Its results are intended for the website and WhatsApp, not only media or political audiences.

The approach is therefore evolutionary, not entirely new. SCORE's difference lies in how these elements are packaged and delivered. Its monthly Punjab tracker may make local trends easier for ordinary voters to examine. Whether it adds lasting value will depend on reliable measurement, transparent methods, and the quality of its continuing data.

In general, pollsters select voters who resemble the wider constituency in important ways, such as location and social characteristics. They ask those respondents about their preferences, then use the sample's answers to estimate broader opinion. A well-designed sample can produce useful estimates without interviewing every voter. The method also requires careful weighting and honest reporting of uncertainty.

A larger sample reduces some random variation, but it cannot fix systematic error. If certain voters are missed, refuse to participate, or are included in the wrong proportions, the result can remain biased. Leading or unclear questions can distort answers. People may also change their minds, misunderstand the question, or report preferences they do not ultimately act on.

Repeated polling adds information only when the underlying process is sound. Monthly surveys can identify genuine movement, but they can also repeat the same error. The article makes this point directly: frequency does not automatically make research more scientific. Transparent methods are essential for judging whether a result is reliable.