You Must Ask the Right Question

Prothom Alo’s recently published opinion poll has already sparked widespread conversation.

At first glance, many news outlets and social media users have treated it as if it were a rigorous and scientific survey of public opinion. It is not.

The design of the poll, especially the core question at its heart, is fundamentally flawed.

What makes matters worse is that much of the media has uncritically accepted it and amplified it, giving the impression that it meets accepted standards of survey research. Rather than clarifying the political landscape, it has deepened public confusion at a time when reliable information is desperately needed.

In recent years, only a few large-scale surveys, such as those conducted by IRI and Innovision, have come close to maintaining international standards in Bangladesh. People have been waiting for more research of this quality. Instead, they have been presented with a poll built on weak methodology and guesswork, wrapped in scientific language but lacking scientific rigor.

One of the most glaring problems is the core question of the poll. Around the world, credible election polling typically asks a straightforward question: “If the election were held tomorrow, who would you vote for?” This captures the respondent’s own preference.

Prothom Alo did not ask this. Instead, it asked people to predict which political party they think will win the most seats. This shifts the focus from personal choice to national forecasting, a task that even experienced analysts struggle with in Bangladesh’s complex political environment.

Ordinary citizens, many without access to reliable political information, cannot reasonably be expected to make accurate constituency-level predictions. The result is not a measurement of public opinion but a collection of speculative guesses.

Compounding this problem is the sample selection. According to Prothom Alo, the survey included respondents from five cities and five rural or upozila areas, totaling 1,342 people aged 18 to 55, almost evenly split between men and women. Yet the most critical question remains unanswered: how were these locations chosen?

Without clarity on whether the selection was random or whether the areas represent specific political strongholds, the poll cannot credibly claim to reflect national sentiment.

Bangladesh’s political landscape varies immensely from region to region. Some districts are dominated by one party, others by religion-based politics, and still others by strong independent figures or intense polarization.

To treat the opinions of people from only a handful of locations as representative of 170 million citizens is statistically indefensible. It is like trying to forecast the entire country’s weather using temperature readings from only five stations.

The survey’s respondent pool further undermines its credibility. Prothom Alo states that the poll reflects national opinions but admits that all respondents were people who read online or print newspapers and are likely to vote. This automatically excludes large segments of the population, especially lower-income citizens, people with limited formal education, and rural communities with less access to news media.

There is no indication of rural-urban balancing, regional weighting, or any attempt to correct for sampling bias. In a diverse country of this size, relying on the views of 1,342 relatively informed respondents to predict national seat outcomes is neither statistically sound nor methodologically serious.

The poll’s claim of 99 percent reliability further misleads the public. In proper survey research, confidence levels require random sampling, well-structured questions, and strict assumptions. When the core question is flawed and the sample is narrow and unverified, such confidence claims become meaningless.

There is no evidence that respondents’ understanding of the prediction question was checked, no system for filtering out random or careless responses, and no mention of phone or digital verification. Without these safeguards, the poll lacks the foundations necessary for a valid reliability claim. It is branding, not science.

The analysis presented by Prothom Alo is equally superficial. Credible opinion polls typically disclose their sampling frame, weighting methods, margin of error, and demographic breakdowns. They show how opinions differ across gender, age, income, region, or education. Prothom Alo offers none of this depth.

Instead, it interprets respondents’ predictions as meaningful indicators of political reality. Assumptions are presented as insights, while the limitations of the data remain unacknowledged.

The absence of safeguards against manipulation further weakens the survey. High-quality polling includes checks for repeated answer patterns, coordinated responses, and internal inconsistencies. Prothom Alo has provided no indication that any such measures were used. Without these checks, the survey could easily be influenced by organized groups seeking to sway the results.

All of this has consequences beyond methodological critique. Public trust in opinion polls in Bangladesh is already fragile. Many citizens believe surveys are biased or manipulated. When a leading newspaper publishes a poll built on such shaky foundations, it reinforces those suspicions. People become more confused, false narratives spread, and the credibility of both research and journalism suffers. A poll that lacks methodological rigor does not clarify the political landscape; it distorts it.

A trustworthy opinion poll requires a trustworthy method: the right questions, a sample that reflects the country’s diversity, strict data verification, and transparency about what the poll can and cannot reveal. Without these basics, no amount of promotion or visual appeal can turn a weak survey into meaningful data. At best, it entertains. At worst, it misleads.

Prothom Alo’s poll fails to meet these essential standards. Compared with the more rigorous efforts by IRI and Innovision, it appears strikingly weak. In a politically sensitive moment, Bangladesh deserves better than guesswork presented as research.

Dr. Nakibur Rahman is a Professor of Finance at University of North Carolina and the US spokesperson for Jamaat-e-Islami Bangladesh.