One headline measure
OverallOverall satisfaction on a labelled 5-point scale, or likelihood to recommend on 0–10. Keep the same measure across rounds.
Guide · Questionnaire design
From objectives to a frozen, translated and piloted instrument
This guide explains how to design a survey questionnaire for households, customers, employees or programme beneficiaries in India. You will turn objectives into indicators, write neutral questions, order modules, translate and back-translate, add skip logic and pilot test before fieldwork. The result is an instrument enumerators can read aloud and analysts can trust.
- Eight steps from objective to frozen form
- Question wording rules with examples
- Outlines for household, socio-economic, customer and employee surveys
- A pilot checklist
Survey 3/6 · AmenitiesOffline
How is drinking water in your area?
- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
✓ Section saved on phone · autosave
BackNext
Home #17 · Intro & consent
Hello, I am Ram Kumar. This survey is run by Sample Organisation. Purpose: to learn about amenities in your area. It is voluntary, and you can say no at any time.
Consent v2 · 09:44
Agreed — start surveyDeclined
One idea per questionFixed recall periodsNeutral wordingBack-translationSkip logicRange checksPilot with real respondentsFrozen version
Quick answer
A good survey questionnaire asks only what the analysis needs, in words every respondent understands the same way, in an order that builds trust before sensitive topics. Each question maps to an objective, has a fixed recall period and clear answer options, and has been tested on real respondents before the main fieldwork begins.
Most questionnaire problems are visible only in the field. A question that reads well in an office in English can become ambiguous in Bhojpuri or Tamil, too long when spoken aloud, or embarrassing when a neighbour is listening. Designing for spoken delivery by an enumerator is different from designing a form someone fills in alone.
Start from the decisions the data will inform, not from a list of interesting topics. If nobody can say which table a question will feed, cut it. Every extra minute costs field money and lowers attention for the questions that matter.
Inputs you need first
- Study objectives in one page
- Indicator list or analysis plan
- Target respondent definition
- Languages of the field area
- Mode: face-to-face, phone or online
- Maximum interview length
Step by step
Follow these steps in order. Skipping the pilot or the translation check is where most instruments fail.
- 01
1. List objectives and indicators
Write each objective, then the indicators that measure it, with numerator, denominator and recall period. This list becomes your question bank's filter.Research lead3,000Target outlets2,184Visited1,962Audited1,871AcceptedDaily accepted auditsLast 12 days - 02
2. Draft questions against each indicator
Reuse tested questions from national surveys where possible so results are comparable. Mark every question with the indicator it serves.Research leadSurvey 3/6 · AmenitiesOfflineHow is drinking water in your area?- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
✓ Section saved on phone · autosaveBackNext - 03
3. Choose answer formats
Use closed options where categories are known, numeric fields for counts, and a few open questions only where you will code them.Research leadSurvey 3/6 · AmenitiesOfflineHow is drinking water in your area?- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
✓ Section saved on phone · autosaveBackNext - 04
4. Order the modules
Consent and screening first, easy factual questions next, sensitive topics later, demographics that feel intrusive near the end.Research leadHome #17 · Intro & consentHello, I am Ram Kumar. This survey is run by Sample Organisation. Purpose: to learn about amenities in your area. It is voluntary, and you can say no at any time.
Consent v2 · 09:44Agreed — start surveyDeclined - 05
5. Translate and back-translate
One translator writes the local version; another converts it back to English. Resolve every difference in meaning, not just words.Translators - 06
6. Program skip logic and checks
Add skips, range limits, consistency checks and required fields in the app so errors are caught during the interview.ProgrammerSurvey 3/6 · Public amenitiesHow is drinking water in your area?- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
BackNext✓ Saved on phone · autosaveSync statussample6Safe on phone8Synced to server#BS-0413 · In area · ±12 m Added to accepted count - 07
7. Pilot test in the field
Interview respondents like your sample but outside it. Time each module and note every hesitation, clarification and refusal.Field team
- 08
8. Revise, freeze and version
Fix what the pilot showed, re-test changed questions, then freeze the version with a number and date before training.Research leadOfflineNo network6Safe on phone8Synced to serverSyncs when network returns
3,000
Target outlets
2,184
Visited
1,962
Audited
1,871
Accepted
Daily accepted auditsLast 12 days
Survey 3/6 · AmenitiesOffline
How is drinking water in your area?
- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
✓ Section saved on phone · autosave
BackNext
Survey 3/6 · AmenitiesOffline
How is drinking water in your area?
- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
✓ Section saved on phone · autosave
BackNext
Home #17 · Intro & consent
Hello, I am Ram Kumar. This survey is run by Sample Organisation. Purpose: to learn about amenities in your area. It is voluntary, and you can say no at any time.
Consent v2 · 09:44
Agreed — start surveyDeclined
Survey 3/6 · Public amenities
How is drinking water in your area?
- Comes regularly
- Comes sometimes
- Rarely / never
- Don't know
BackNext
✓ Saved on phone · autosave
Sync statussample
6
Safe on phone
8
Synced to server
#BS-0413 · In area · ±12 m Added to accepted count
Offline
No network6
Safe on phone
8
Synced to server
Syncs when network returns
Wording
Most bad questions break one of a few rules. Check every draft question against this table.
| Rule | Weak question | Better question |
|---|---|---|
| One idea per question | Is the ration shop clean and open on time? | Split into two: cleanliness, then opening hours |
| Fixed recall period | Do you often visit the health centre? | In the last 30 days, how many times did you visit the health centre? |
| Neutral wording | Don't you agree the new road has helped? | Has the new road made travel to the market easier, harder or no different? |
| Answer options that cover all cases | Income: below 5,000 / 5,000–10,000 | Add higher bands, "no income" and "refused" |
| Words respondents use | What is your household's primary occupation? | What work brings in most of your household's money? |
| Avoid assumed behaviour | Which app do you use to pay bills? | Do you pay any bills on a phone? If yes, which app? |
Read every question aloud. If the enumerator needs to explain it, rewrite it.
Pilot
Pilot testing a survey instrument in India means interviewing people similar to your sample, in the real languages and settings, before fieldwork. It checks timing, comprehension, translation, skip logic, sensitive questions and device performance. Plan at least one round of revision and re-test any question you change substantially.
Run the pilot like the real survey
A desk review is not a pilot. Use trained enumerators, the programmed app and real households or customers outside the sample.
- Pick sites that resemble the hardest parts of your sample.
- Record module-wise timing for every interview.
- Ask respondents to explain difficult questions in their own words.
- Log every skip error, crash and forced answer.
- Hold a debrief with enumerators the same evening.
Signals that a question needs rewriting
Many "don't know" answers, long pauses, respondents asking for the question again, enumerators paraphrasing differently, or answers piling into one option all signal a problem. So do questions that make respondents look at family members before answering; move those later or find a private setting.
Pilot outputs
- Average interview length
- List of changed questions
- Updated translations
- Final skip map
- Version number and date
Our view on questionnaires
Household surveys
A household survey questionnaire in India usually opens with identification and consent, then a household roster, housing and assets, livelihoods, the study's core module and access to services. Using definitions from national surveys, such as the NSS household and NFHS modules, makes your figures easier to compare with official estimates.
Typical modules
Adapt the order so rapport is built before income or health questions.
- Identification: location codes, GPS, interviewer ID, consent.
- Roster: members, age, sex, relation, education, activity.
- Housing and assets: materials, water, toilet, fuel, listed assets.
- Livelihoods: land, crops, livestock, wage work, migration.
- Core module: your study's indicators asked as defined.
- Services and schemes: access, use and problems faced.
Who answers
Decide in advance who the respondent is for each module. The household head may answer the roster, but questions about a woman's health or a child's schooling should go to the person who knows. Record who answered each module so analysts can check consistency.
Socio-economic
A socio economic survey questionnaire measures a household's economic position and social profile so that results can be broken down by groups. Keep categories aligned with official classifications, and ask sensitive identity questions only when the analysis needs them and consent covers them.
Economic
Consumption and income
Consumption is usually easier to recall and more stable than income. If you ask income, ask by source and for a defined period.
Economic
Assets and housing
An asset list with house materials often ranks households better than a direct income question, and respondents find it less intrusive.
Work
Occupation and employment
Ask about main and secondary activity, days worked and type of employer. Use categories close to national labour survey definitions.
Education
Schooling and literacy
Record highest class completed for each member and current enrolment for children, rather than a vague literacy yes or no.
Social
Social group
Where the brief requires it, use official categories, explain why the question is asked and always include a refusal option.
Access
Schemes and services
Ask about specific entitlements received in a fixed period, not general awareness. Keep scheme names in the local form people use.
Customers
Customer satisfaction survey questions should measure a specific experience soon after it happened, then ask what drove the rating. A short overall score plus a few attribute ratings and one open reason is usually more useful than a long battery of agree-disagree items.
Four to eight drivers
AttributesRate speed, staff behaviour, product quality, price clarity and resolution of problems, each tied to a team that can act.
One open question
Reason"What is the main reason for your score?" Code the answers into themes; they explain the numbers.
Employees
Employee engagement survey questions cover how people feel about their work, manager, growth and workplace. Anonymity decides honesty: report only groups large enough that no one can be identified, offer the survey in the languages staff speak, and tell employees in advance what will happen with the results.
Manager and team
"My manager tells me clearly what is expected of me." Ask on a labelled agreement scale.
Work and purpose
"I have the tools and information I need to do my job well." Tie low scores to fixable resources.
Growth
"In the last six months someone has talked to me about my progress." A time frame keeps it factual.
Safety and fairness
Include questions on workplace safety and fair treatment, with a confidential channel for serious concerns.
Avoid these
These are the errors that show up again and again when a dataset reaches analysis.
Wording
- Double-barrelled questions
- Leading or loaded words
- No recall period
- Jargon and English terms in local versions
Structure
- Sensitive questions too early
- Missing "don't know" and "refused"
- Skips that trap respondents
- Interview far longer than planned
Process
- No back-translation
- Pilot skipped or done in the office
- Changes after training without re-briefing
- No version number on the final form
How it works
From questionnaire to verified data on the Bharat Survey platform.
Learn moreकैसे काम करता है (हिंदी)
यही प्रक्रिया हिंदी में।
Learn moreHow to conduct a baseline survey
Indicators, sampling and a household questionnaire outline.
Learn moreNABH patient satisfaction questionnaire
A hospital feedback questionnaire outline.
Learn moreCAPI, CATI and CAWI explained
How mode changes question design.
Learn moreSurvey cost in India
Why interview length drives the budget.
Learn moreFAQ
Straight answers to the questions people ask most about this topic.
How long should a survey questionnaire be?
As short as the analysis allows. Many household surveys run 30 to 45 minutes, customer surveys under 10 and phone surveys 10 to 15. Time it in the pilot, not on paper. If it runs long, cut questions that do not feed a planned table before you cut sample size.
Should I use a 5-point or 10-point scale?
For spoken interviews, a labelled 5-point scale is easier because respondents can remember every option. Longer numeric scales work better in written or online surveys. Whatever you choose, keep it the same across rounds so results can be compared over time.
How many people should a pilot test include?
Enough to see each skip path and each language used, often a few dozen interviews across different respondent types. The goal is to find problems, not to estimate results. If you make major changes, run a smaller second pilot on the revised questions.
Can we reuse questions from government surveys?
Yes, and it is often wise. Questions from national surveys have been tested at scale, and reusing their definitions lets you compare your results with official figures. Check the recall period and categories match, and credit the source in your methodology note.
Who should translate the questionnaire?
A translator fluent in the local spoken variety, not only the formal language, followed by an independent back-translation into English. Field supervisors from the area should review the final wording, because they know which words respondents actually use.
India Nationwide, Hindi-first
Send it with your objectives. We will review wording, flow and length, then program, translate and pilot it for your field area.