Doppler runs a national household panel and publishes the questionnaire, the sample design and the funder with every release. The Bangladesh Approval Rating is its first national reading of the government elected in February.
Doppler Opinion Research builds on the works of INNOVISION Consulting’s seminal People’s Election Pulse Survey and is founded as a dedicated and independent research entity to undertake public opinion research, market research and social research to inform national economic and social development policies of Bangladesh.
The first Doppler Approval Rating Survey finds strong approval for the Prime Minister, but lower ratings for the government and the country’s direction. People remain hopeful about the future despite acknowledging worsening economic conditions, with food and fuel prices and electricity emerging as the biggest concerns. The Computer Assisted Telephone Interview (CATI) was done on the panel of 10,413 samples and the success rate was 33.1%. In total, 3,443 samples completed the survey and they represent 489 wards and villages in all 64 districts. Data was collected from 6th August–15th August, 2026.
Published 19 August 2026 · first wave
One dot is one per cent of respondents, rounded to the nearest whole point.
“Overall, how well or badly would you say the government elected in February is doing its job?”
Very well · Fairly well · Fairly badly · Very badly · Does not know enough about it · Refuse or cannot say
Of the seven issues tested, only education gets a favourable rating.
No list was read out. Respondents named up to three problems in their own words.
Both are rated badly by fewer people than the government's 32.9%. But about a third of adults hold no view of either.
Five questions. People are hard on the past year and hopeful about the next.
The first row has no middle option: the question asks right direction or wrong direction.
Share saying the country is going in the right direction. The vertical line is the national figure, 48.6%.
Divisional figures show real differences between divisions; they are not exact levels. In Sylhet 26.4% will not answer at all, the highest refusal in the survey.
Both policies the government has brought in are more popular than the government itself, which stands at 54.5%.
The Family Card is still a pilot. High support is not evidence that households have felt the benefit yet.
Thinking a leader does his job well is not the same as feeling his party speaks for you.
11.0% say Jamaat-e-Islami speaks for people like them a lot, and 7.9% say the same of the NCP. Yet 47.7% and 39.8% say the two leaders do their job well.
In all six party groups the Prime Minister is rated above the government, and the government above the direction of the country.
Share holding no opinion on each item. On the question that reaches the home directly, the gap all but disappears.
Women are not more critical. On questions that reach the home directly they answer just as often: on the handling of electricity the gap is 0.9 points.
Six months after the election that chose them.
59.2% of graduates think their Member would listen, against 44.1% of adults with no schooling. Income changes nothing.
A different group feels less safe in each. The vertical line is the national figure.
Feels safe walking alone after dark · national 67.5%
Feels safe practising religion openly · national 95.2%
Never publish the 95.2% on its own; report the minority figures with it. 3.1% of women never go out after dark at all, so the real gap is wider.
DOPPLER · Doppler Approval Rating Survey 2026 · n=3,443 adults · fieldwork 6th August–15th August, 2026 ·
about ±2.6 points on a national figure at 95% CI · telephone interviews in Bangla, on the 10,413-person panel
first interviewed face to face in September 2025; 5,493 answered and 3,443 completed, a 33.1% response rate ·
64 districts across all 8 divisions, 489 sampling units · weighted to the census on age, sex, schooling,
division and town or village · self-funded by Doppler Opinion Research, with no party, ministry or campaign involved ·
full tabulations and questionnaire available on request
Attribute to: Doppler Approval Rating Survey 2026, Doppler Opinion Research. Quote figures to one decimal place. Free to report with attribution.
Bangladesh has 175 million people and 127.7 million registered voters, and no continuous independent measure of what they think. Official statistics command limited confidence: 29% of users reported satisfaction with Bangladesh Bureau of Statistics data in 2024, and the World Bank's statistical capacity score fell from 80 to 60 between 2014 and 2020. Governments legislate without it. Investors price risk without it. Development partners design programmes without it.
Every release carries its sample, weighting, field dates, mode, funder and full questionnaire in Bangla and English. A reader who distrusts the finding can audit the instrument that produced it.
Doppler asks identical questions and draws identical samples whoever the answer favours. The company runs no private polling for any party or campaign.
Every number is published with its margin of error. Differences inside that margin are reported as no change.
The company publishes readings, not endorsements, forecasts of preference or commentary on what the country ought to do.
20,000 consented households across all 64 districts, profiled and recontactable.
Door-to-door CAPI, a CATI call centre, and online instruments on one sampling frame.
Election-validated weighting, small-area estimation and quality protocols on every wave.
Questionnaire, sample design, funder and margin of error released with each finding.
Doppler measures the years between them. In August 2024, eight days after the fall of the previous regime, INNOVISION Consulting launched Bangladesh Speaks, the country's first micro-polling platform. What followed became the People's Election Pulse Survey: three national rounds, in February 2025, September 2025 and January–February 2026.
The first online and offline reading of the post-transition electorate. It identified the Islamist resurgence months before that shift entered mainstream commentary.
10,696 voters across all 64 districts. The findings became a national talking point, and competing agencies began fielding lookalike surveys within months.
10,413 eligible voters interviewed across 8 divisions, 64 districts and 521 primary sampling units, using two-stage stratified cluster sampling with probability proportional to size. The Economist cited the results and every major party planned against the numbers.
Panellists re-contacted ahead of the February parliamentary election. Of all published forecasts, Round 3 came closest to the result in the percentage distribution of votes.
The election instrument becomes a standing one. The September 2025 panel was called again and 3,443 adults completed a twenty-question interview. The survey is self-funded and run by the team that designed PEPS, with no party, ministry or campaign involved.
A consulting firm can run a survey programme when a client commissions one; only a dedicated research house can run a permanent one: monthly, quarterly, on a fixed calendar announced in advance, whether or not anyone has commissioned it that month.
The panel and the field force are fixed assets; every study runs across them. That is why a brand tracker, a policy evaluation and a national index can share one wave, and one field cost.
The measurement Doppler was built for.
Studies designed to a single client's question.
Doppler sells access to the infrastructure itself.
Doppler works with whoever needs evidence about this country, on the same method and the same disclosure standard for every client.
Doppler supplies the population evidence that programme design, targeting and measurement require, on a national frame that already exists. Commission a baseline, buy questions on an omnibus wave, subscribe to a national index, or run your own instrument on the panel.
Talk to us about: baseline and endline measurement, citizen feedback systems, hard-to-reach population studies, and national context data for country strategies.
Commission researchBangladesh will be the world's ninth-largest consumer market by 2030, and 34 million middle and affluent consumers are already navigating high inflation. Doppler measures what they think of your category, your brand and your price.
Talk to us about: brand health tracking, category barometers, pricing and concept testing, advertising effectiveness, segmentation, and omnibus questions.
Talk to our commercial teamDoppler measures how policy lands with the population it targets, on a method published in full so the finding survives scrutiny. Public institutions are served on the same terms as every other client, and none of them sees a public index before it is published.
Talk to us about: service delivery measurement, policy response research, citizen satisfaction tracking, and independent verification of programme reach.
Start a conversationDoppler publishes full data tables, questionnaires and methodology statements with every release, free to report on with attribution. Newsrooms can commission questions on an omnibus wave, or partner on a recurring index.
Talk to us about: data partnerships, commissioned polling, election-period research, and archive access for long-run analysis.
Media enquiriesDoppler holds a national sampling frame, a profiled panel and a longitudinal archive of Bangladeshi opinion from 2024 onwards. Researchers can field original instruments, embed experiments in a wave, license micro-data, or collaborate on published work.
Talk to us about: survey fielding for peer-reviewed research, panel and micro-data access, embedded survey experiments, replication data, and co-authored studies.
Partner with usDoppler is opening routes into the field: internships in questionnaire design, fieldwork and analysis; enumerator training and certification; and thesis support for students working with Doppler data.
Talk to us about: internships, the enumerator network, thesis data access, and training in survey methods.
Apply or askDoppler hires researchers, statisticians, field managers, enumerators, data engineers and product staff. Applications are open on a rolling basis.
Three national rounds, 26,256 interviews and a sampling frame mapped door-to-door across every district, designed and directed by the people who now run Doppler.
Founder and Chief Executive
Conceived Bangladesh Speaks and directed all three rounds of the People's Election Pulse Survey, the programme that produced the closest published forecast of the February 2026 parliamentary election.
A development economist and management consultant, and Managing Director of INNOVISION Consulting, he has spent more than 19 years designing and evaluating systemic solutions to poverty. At Doppler his role is the instrument itself: research design, sampling strategy, and the standard the company publishes against.
Md. Rubaiyath Sarwar is an Acumen Fellow. He completed his BBA and MBA from IBA, University of Dhaka, and an MSc in Economics from North South University.
Chairman
Chairman and Chief Executive of INNOVISION Consulting, built over two decades into a research and advisory firm working across Bangladesh, Ethiopia, Nepal, Nigeria, Afghanistan, Zambia and Yemen for DFID, SDC, the European Union, DANIDA, GIZ and USAID. He served as chief executive on the PEPS programme.
MBA and BBA from the Institute of Business Administration, University of Dhaka. At Doppler he chairs the board and holds responsibility for the governance that protects the company's independence.
Director of Research
Built the survey instruments for the People's Election Pulse Survey. A doctoral researcher in Political Science and Social Data Analytics at Pennsylvania State University, working on opposition parties, resistance to autocratisation and public policy in weak democracies.
Master of Public Policy from the University of Michigan as a Fulbright scholar; MSS and BSS in Development Studies from the University of Dhaka. He leads questionnaire design and the analytical framework behind the political and social series.
Whether or not they are a client. Full tables, questionnaires and methodology statements for every wave are published in Bangla and English.