Verified Survey Respondents in India: What Verification Proves, and What It Does Not

There are four steps. The first happens once. The rest happen on every study.

The point of the design is separation. The identity check happens once, with the member. The researcher only ever sees that the person was verified, never who they are.

A zero-knowledge proof lets one party convince another that a statement is true without revealing why it is true. In the standard definition, a prover convinces a verifier that a statement holds, and the verifier learns nothing beyond the fact that it holds (Wikipedia, zero-knowledge proof). The idea was developed by Shafi Goldwasser, Silvio Micali and Charles Rackoff (a16z crypto).

Think of a nightclub that needs to know you are over 18. Showing your whole ID tells the bouncer your name, address and date of birth. A zero-knowledge check would tell him only "over 18: yes". Our case is the same. The statement is "this respondent passed identity verification". The proof confirms that and nothing else.

We make no claim here about the cryptographic scheme beyond what the methodology page states.

Identity documents are checked once, at joining. Researchers never receive them.

Religion, caste and sexuality are never collected. Because age band, gender, city and tier, and NCCS are already known, you can cut results by them without asking. That keeps the survey short and free of personal data. See targeting fields for the values you can set.

A verified person can still click through a survey. These checks catch that.

• Attention checks. Inserted automatically, roughly one for every seven questions, for example a prompt to select a named colour or animal.
• Speeders. Respondents who finish faster than a careful read allows are flagged.
• Straight-lining. The same answer down a grid is flagged.
• Open-end scoring. Text answers are scored for gibberish, copy-paste and off-topic replies.
• Deduplication windows. Repeat respondents are excluded within set windows.
• Screener integrity. Category screens are asked neutrally among decoys, and key criteria are re-checked in different words.

Flagged responses are removed before analysis. We do not publish removal rates, so ask any provider, us included, to show the share removed on your own study. For the wider picture, read our guide to survey data quality in India.

• A verified identity is not a guarantee of attentive answers. Verification shows a real, unique person. It says nothing about how carefully that person reads question 18. That is why the checks above exist, and why you should read open-ended answers yourself.
• Smartphone users only. SuperJ is app-based. Offline or very low-connectivity households are under-represented and need face-to-face fieldwork.
• India only. The panel is not global.
• Eligible panel is a subset. The 25M+ figure is the consumer footprint of the SuperJ network. The verified, survey-eligible panel is smaller, and availability depends on your targeting.
• What JMDT is. The JMDT record stores that a member passed verification. We make no claims about its speed or tamper-proofing, and you should not assume more than that.

Verification is one layer of data quality. Study design, screener wording and analysis matter as much. See consumer panel India for how panel types differ.

The industry body ESOMAR publishes a set of questions for buyers of online samples, meant to help buyers evaluate different sample providers (Ipsos, June 2025). The list below is our own, narrower and focused on verification. Use it in a call or a written tender.

• How is a member's identity checked at joining, and by what document or method?
• Is the check once, or repeated? How often is it refreshed?
• How do you stop one person holding several accounts?
• What does the researcher receive: a flag, an ID, or personal data?
• What proof can you show me that a given respondent was verified?
• Where is that record stored, and who can audit it?
• Do you recruit from other panels or exchanges? If so, are those respondents verified the same way?
• How do you flag speeders and straight-lining? What do you remove, and what share was removed on my last study?
• How are open-ended answers checked?
• What is the duplicate exclusion window across studies?
• Which groups does the panel under-represent, and what do you suggest for them?
• Can I run a small pilot and inspect the raw responses before committing?

A good provider answers each point with a specific method. Be wary of a reply that says "we use advanced AI" and stops.

If your job is a survey among verified Indian consumers, this is what a 400-response study costs at list price.

Pro is sold in bundles of 400 to 1,000 responses and costs 20% less per response than Plus. Enterprise plans come with a dedicated research team and are priced on request.

The SuperJ network has a 25M+ consumer footprint, and the survey-eligible panel is a subset. Members complete Aadhaar-based KYC once, and each study receives a zero-knowledge proof. Surveys run in 8+ Indian languages, and flagged responses are removed before analysis. SuperJ is on Google Play and the App Store. See /panel-methodology and pricing.

The limit: SuperJ is app-based, so it reaches smartphone users only. Offline households need face-to-face fieldwork.