How to Use Questionnaires for Effective Data Collection

Whether you’re studying maternal health in rural Bihar or measuring contraceptive awareness among college students, the questionnaire is the workhorse of quantitative research. It looks deceptively simple, just a list of questions, but a well-designed questionnaire is the difference between research that informs national policy and data that ends up in a forgotten folder. Let’s unpack how this method actually works, the different forms it takes, and how to design one that delivers reliable, unbiased results.

What is the questionnaire method? 🔗

A questionnaire is a structured tool consisting of a set of written questions used to gather information from respondents about a specific research problem. In population and family health studies, it is the most widely used instrument for collecting data on fertility, morbidity, contraceptive use, healthcare-seeking behaviour, and demographic patterns. Large-scale surveys like the National Family Health Survey (NFHS) rely heavily on carefully designed questionnaires administered to lakhs of households.

The strength of the method lies in its standardisation. Every respondent receives the same questions in the same order, which makes responses comparable and findings statistically meaningful. The weakness is equally real: a poorly worded question can quietly bias an entire dataset before analysis even begins.

Types of questionnaires 🔗

Questionnaires are broadly classified by how rigidly the questions and answers are structured. Researchers usually choose between structured, unstructured, or a hybrid semi-structured form depending on whether they need numbers, narratives, or both.

Structured questionnaires 🔗

A structured questionnaire has a fixed format with predetermined questions and a fixed set of answer choices. Respondents simply tick or select from the options provided. This format is also called a closed or restricted questionnaire because it limits the range of replies. Structured questionnaires collect quantitative data that can be measured and expressed numerically, which makes them ideal for population-level studies where the goal is to count, compare, and generalise.

For example, a structured question on family planning might read: “Have you ever used any modern contraceptive method? (a) Yes (b) No (c) Don’t know.” The responses are quick to code, easy to analyse statistically, and consistent across thousands of respondents.

Unstructured questionnaires 🔗

An unstructured questionnaire uses open-ended questions with no predetermined answer choices. Respondents reply in their own words, allowing them to express opinions, feelings, and experiences freely. This format generates qualitative data that captures the depth and texture missing from tick-box surveys. Open-ended questions allow respondents to address their views openly without being limited by answer format or choices.

A maternal health researcher might ask: “Can you describe your experience during your last antenatal visit?” The replies are rich and revealing, but they take much longer to analyse and cannot easily be converted into percentages.

Semi-structured questionnaires 🔗

Most real-world health research falls in the middle. A semi-structured questionnaire combines closed and open-ended questions, asking standardised demographic and behavioural items alongside a few open prompts that let respondents elaborate. This hybrid approach is common in family health studies because it balances statistical rigour with human nuance.

Closed-ended vs. open-ended questions 🔗

Within any questionnaire, the question itself is either closed or open, and choosing between them shapes the data you collect.

Closed-ended questions offer a fixed set of responses, dichotomous (Yes/No), multiple-choice, Likert scales (Strongly agree to Strongly disagree), or ranking items. They are quick to answer, easy to compare, and friendly to statistical software. Their drawback is that the researcher must anticipate every plausible answer in advance, and respondents may pick the “closest” option even when none fits.

Open-ended questions invite respondents to construct their own answers. They surface unexpected insights, particularly useful when studying sensitive issues like reproductive health stigma, where forced choices can hide more than they reveal. The trade-off is that responses are time-consuming to code and may suffer from low completion rates if respondents lose interest.

Methods of questionnaire administration 🔗

How a questionnaire reaches the respondent matters as much as what it asks. The same set of questions can yield very different response rates and quality depending on the mode of administration. Common modes include face-to-face, mail, telephone, email, and internet questionnaires, each suited to particular research contexts.

Personal interviews 🔗

In this method, a trained interviewer reads the questions aloud and records the respondent’s answers. It remains the gold standard for large household surveys in India, including NFHS and the Sample Registration System, because it works regardless of the respondent’s literacy level.

Implementation typically follows these steps: train interviewers on the questionnaire and ethical protocols, pilot the tool, obtain informed consent, conduct the interview in a private setting, record answers carefully, and back-check a sample to verify quality. Response rates with interviews are typically 70-80%, compared with under 50% for self-administered questionnaires, mainly because personal contact reduces refusals and the interviewer can clarify confusing items. The downsides are high cost, the time required, and the risk of interviewer bias if questions are not asked exactly as written.

Mail questionnaires 🔗

The researcher posts the questionnaire to selected respondents, who fill it in at their convenience and return it by post. Mail surveys are useful when the sample is geographically dispersed and the budget for fieldwork is limited. They also offer respondents anonymity, which can improve honesty on sensitive items.

The standard steps are: prepare a clear cover letter explaining the purpose, enclose the questionnaire with a self-addressed stamped return envelope, send reminders after two to three weeks, and track returns systematically. The major problem is low response, often under 30%, and there’s no way to confirm whether the intended person actually filled it out.

Telephone surveys 🔗

Telephone administration is faster and cheaper than face-to-face interviews and is widely used for opinion polling and rapid health assessments. With mobile penetration crossing 1.15 billion subscribers in India, telephone surveys have become more representative than they were a decade ago.

Implementation involves preparing a short, conversational script, training callers, using random-digit dialling or a pre-existing sample frame, and recording responses through computer-assisted telephone interviewing (CATI) software where possible. The format limits how many questions can be asked, usually 10 to 15 minutes’ worth before respondents disengage, and visual aids like showcards are impossible.

Email and online questionnaires 🔗

Email and web-based surveys, distributed through platforms like Google Forms, SurveyMonkey, or KoBoToolbox, have transformed data collection in the last decade. They are nearly free, instantaneous, and automatically compile responses into a spreadsheet.

Implementation includes designing a mobile-responsive form, drafting a recruitment email or social media invite with a unique link, sending reminders, and ensuring data security. The catch is the digital divide: online surveys systematically miss respondents without smartphones, stable internet, or English/regional-language digital literacy, which is a serious limitation in population health research where the most vulnerable groups are often the least connected.

Designing effective questionnaires 🔗

The design phase is where most questionnaires succeed or fail. A few principles, applied consistently, make a substantial difference.

Start with clear research objectives 🔗

Every question should connect to a specific research objective. If you cannot explain why a question is in the questionnaire, delete it. Long questionnaires fatigue respondents and reduce data quality, so ruthless pruning is a virtue.

Use simple, unambiguous language 🔗

Write at the reading level of the least-educated respondent in your sample. Avoid jargon, technical terms, and double negatives. In multilingual contexts, translate the questionnaire into the local language and then back-translate to check for distortions, this is standard practice for WHO and UNICEF surveys.

Avoid leading and biased questions 🔗

A question like “Don’t you agree that immunisation is important?” practically tells the respondent what to say. Keep questions free of bias by avoiding language that pushes respondents to respond in a certain way, and remember that respondents tend toward socially desirable answers, especially on sensitive topics like sexual behaviour, alcohol use, or domestic violence.

Avoid double-barreled questions 🔗

“Are you satisfied with the cost and quality of maternal care?” merges two separate ideas. A respondent who finds it cheap but poor in quality has no honest way to answer. Split such items into two questions.

Order questions thoughtfully 🔗

Begin with easy, non-threatening questions to build rapport. Group related items together. Place sensitive questions (income, sexual practices, illegal behaviour) toward the end, when trust has been established. End with demographic items.

Design response options carefully 🔗

For closed questions, answer options should be mutually exclusive (no overlap) and collectively exhaustive (cover every plausible response). Always include “Don’t know” or “Prefer not to answer” for sensitive items. Balance Likert scales with equal positive and negative options around a neutral midpoint.

Pre-test and pilot the questionnaire 🔗

This step is the single best investment a researcher can make. Pre-testing focuses on specific questions to evaluate their effectiveness using techniques like cognitive interviewing, while pilot testing administers the entire questionnaire to a small, representative sample to simulate real-world conditions.

A typical pilot involves 20 to 50 respondents drawn from the target population. The researcher checks for confusing wording, ambiguous response options, skip-logic errors, time taken to complete, and the analysability of the data collected. Because it is rarely possible to foresee all the potential misunderstandings or biasing effects of different questions, it is vital for a well-designed survey to include provision for a pretest. Skipping this step is the most common cause of unusable data in student dissertations.

Strengths and limitations to keep in mind 🔗

The questionnaire method is cost-effective, scalable, and produces standardised, quantifiable data, all critical in population and family health research. At the same time, it depends entirely on what respondents are willing and able to tell you. Recall bias, social desirability bias, and selection bias (especially in online modes) can quietly distort results. The method also cannot capture context that the researcher did not anticipate, which is why mixed-methods designs combining questionnaires with focus groups or in-depth interviews are increasingly the norm in serious health research.

What do you think? If you were designing a questionnaire to study adolescent reproductive health awareness in your state, which mode of administration would you choose and why? And how would you handle the trade-off between getting a high response rate and protecting respondents’ anonymity on sensitive questions?

References

1. https://main.mohfw.gov.in/sites/default/files/NFHS-5_Phase-II_0.pdf
2. https://www.surveymonkey.com/mp/what-is-a-questionnaire/
3. https://www.cvent.com/en/blog/events/structured-vs-unstructured-questions
4. https://www.betterevaluation.org/methods-approaches/methods/questionnaires
5. https://www.sciencedirect.com/topics/computer-science/administered-questionnaire
6. https://aapor.org/standards-and-ethics/best-practices/
7. https://soundrocket.com/best-practices-for-questionnaire-design/
8. https://pmc.ncbi.nlm.nih.gov/articles/PMC6830238/