Explained : What’s Wrong With ‘Modi is the World’s Most Popular Leader’ Survey Claim?  and Its Impact

Every few months, Indian media proclaims that “Modi is the world’s most popular leader,” citing approval numbers from an American firm called Morning Consult. These headlines rest on a single proprietary online tracker whose methods and limitations are rarely scrutinised. It is important to unpack who runs this survey, how it works, and why its numbers – and the screaming headlines it spawns – should be treated with extreme caution.

Who runs the survey?

The figures come from the Global Leader Approval Rating Tracker produced by Morning Consult Political Intelligence, a commercial product of Morning Consult, a US‑based business‑intelligence and polling company founded in 2014. Morning Consult is a privately held, venture‑backed, for‑profit firm; its CEO and co‑founder is Michael Ramlet.

The tracker covers leaders in more than 20 countries, including India, and publishes rolling approval/disapproval numbers based on daily online surveys of adults in each country.

How does their Modi rating work?

The familiar claim that “Modi is the world’s most popular leader” originates in this Global Leader Approval Rating Tracker. For each leader, Morning Consult reports a 7‑day rolling average of responses to a standard question on whether respondents approve or disapprove of the leader’s job performance. The product is designed and marketed as “political intelligence” for corporate and political clients, with only a subset of the data and methodology available freely on their site.

Methodologically, Morning Consult relies on large online surveys drawn from multiple panel providers. These are non‑probability samples: respondents volunteer for online panels rather than being randomly selected from the entire population. The firm then applies weighting by demographics (age, gender, region, education, etc.) using official statistics to make the sample resemble the country’s adult population.

What is wrong with the “Modi is the world’s most popular leader” headline?

Two separate leaps are being made in this claim by the Indian media. One from a single commercial firm’s online tracker to a precise estimate of Indian public opinion, and the other from that one country estimate to a cross‑country league table of global popularity. Both rest on shaky grounds.

First, Morning Consult’s Modi approval series measures the views of online, panel‑reachable Indian adults, not the entire populace. In a country with India’s digital divides, this is a crucial distinction. The numbers are therefore an imperfect and poor proxy of a section of Indians at best, not an accurate measure of India’s vast and diverse population.

Second, cross‑country comparisons such as “Modi vs Trump vs Carney” assume that the survey in each country has comparable coverage, panel quality and political context. In reality, countries differ hugely in internet penetration, panel infrastructure, media environments and response patterns. Treating all these numbers as directly comparable is methodologically sloppy.

Finally, the exact levels – 68% vs 75% approval – are often over‑interpreted as precise point estimates when, for non‑probability online polling, the real uncertainty is much larger and not fully quantified. In practice, the tracker is best used for within‑series trends (does their measured approval for a given leader rise or fall after a major event?) rather than for absolute levels or global rankings.

Whenever you’re reporting on survey results, always include a note on methodology.

So I dug around and found that the survey org Morning Consult only has a sample of 300-500 for India. Also the surveys are done using “general online or literate population”. Then the sample is… https://t.co/ptDXcYFIpc

— Vasundhara Sirnate (@vsirnate) March 25, 2026

How should one then treat the Modi numbers?

A reasonable working rule is to treat Morning Consult’s Modi approval series as one noisy indicator of attitudes among India’s online, reachable population, and not as the definitive voice of 1.4 billion Indians. One needs to be extremely cautious about exact levels (e.g., 68% vs 75%) as population point estimates, and the global popularity league tables, which ignore deep cross‑country differences in measurement quality, politics and media. These are the two things most popular in the Indian media.

How reliable is Morning Consult generally?

Morning Consult is a serious, data‑heavy operation, but its record is flawed, and its methods attract criticism even in the US context. For the 2016 US presidential election, Morning Consult’s national polling showed Hillary Clinton ahead by more than her eventual popular‑vote margin, in line with a broader industry tendency to overstate her advantage. Various assessments of pollster performance in subsequent US midterms have placed Morning Consult below average in accuracy compared to traditional pollsters using probability‑based methods.

Morning Consult has heavily publicised work on phenomena like “shy Trump voters” and mode effects (online vs phone), but subsequent election outcomes and outside analyses have cast doubt on the scale and political significance of the effects they described. The deeper issue is not unique to Morning Consult: non‑probability online polling itself is contested among survey methodologists.

Is this an issue with non-probability online polling in general?

Experts have pointed to several recurrent problems with non‑probability online pollsters, which includes Morning Consult. These concerns apply even more strongly in low‑to‑uneven‑internet environments, particularly a structurally unequal digital landscape like India.

No true sampling error

Classical margins of error assume a random sample from a known population. Non‑probability samples do not satisfy this assumption, so quoting a “margin of error” under‑states total uncertainty.

Heavy reliance on weighting

Statistical weighting is used to compensate for imbalances in the raw sample. If the weighting variables (age, gender, region, etc.) fail to capture politically relevant differences (media exposure, caste/community, offline mobilisation, fear of expressing dissent), then model‑based bias can persist or even get worse.

Limited transparency and replication

Because these surveys are proprietary products, microdata, full codebooks, and weighting schemes are rarely released. This prevents independent academics or journalists from testing robustness or re‑analysing the data.

But are there any methodological red flags or caveats specifically about this survey?

There are some red flags that need to be marked when interpreting the high Modi numbers in the survey. 

First, Morning Consult explicitly describes its work as “online polling” using non‑probability panels rather than random digit dialling or sampling from a population register. In India, this can bias the sample toward urban and peri‑urban residents, smartphone and high‑speed internet users, and more educated, higher‑income and more politically engaged respondents. These groups are not politically neutral; most analyses show that they skew toward the ruling party and its preferred narratives relative to the poorest, least connected citizens.

Second, Online panels work best in countries with near‑universal internet coverage. Even though India’s digital access has grown, large gaps persist by income, gender, region, religion and caste. Under‑coverage of offline, low‑income and older Indians systematically reduces exposure to citizens who may be less enthusiastic about Modi.

Third, Morning Consult publishes high‑level methods – daily online interviews, rolling averages, weighting – but does not provide a full India methodology note in the public tracker. Missing or opaque details include the panel composition and vendor mix in India, rural/urban stratification and quotas, language coverage (which Indian languages are used, how translations are tested), and response rates and attrition patterns. Without these, it is impossible for external experts to judge whether the Indian sample meaningfully represents poorer states, non‑Hindi regions or politically marginalised groups like Dalits, Adivasis and Muslims.

Fourth, the public‑facing tracker typically shows only topline numbers. It does not provide the exact question wording used in Indian languages, the order of questions in the survey (which can prime responses), and any split‑sample experiments or robustness checks. Subtle differences in wording – “approve of performance” vs “satisfied with the way he is doing his job” – can materially affect approval rates, especially in highly polarised environments.

Finally, political Intelligence is sold to corporate and political clients as a near‑real‑time dashboard of public mood across dozens of countries. There is a built‑in commercial incentive to present a clean time‑series charts with stable, comparable trends or simplistic narratives like “world’s most popular leader.” This often comes at the expense of prominent uncertainty communication, nuanced caveats about representativeness and comparability and full technical disclosure that would allow external audit.

In short, the headline “Modi is the world’s most popular leader” should be ideally written as: “According to one US‑based, online, non‑probability polling firm’s proprietary tracker – subject to significant methodological caveats – Modi currently records the highest measured approval among the leaders it tracks.” That is a far more modest claim than the screaming headlines in the Indian media suggest.

This article went live on March twenty-sixth, two thousand twenty six, at forty-three minutes past eleven in the morning.

The Wire is now on WhatsApp. Follow our channel for sharp analysis and opinions on the latest developments.