Questionnaire – Free Builder, Types, and Examples

---

---

---

A questionnaire is a research instrument made up of questions or prompts used to collect information from respondents. It may contain closed-ended questions, open-ended questions, rating scales, or a combination of formats. Researchers use questionnaires to measure characteristics, experiences, behaviours, knowledge, attitudes, opinions and other variables in a consistent way.

Introduction

Questionnaires are used in academic research, healthcare, education, psychology, sociology, business, public policy and programme evaluation. Their apparent simplicity can be misleading. A questionnaire is not merely a list of questions: it is a measurement and data-collection instrument whose wording, response options, order, delivery method and testing can influence the quality of the findings.

This guide explains what a questionnaire is, how it differs from a survey and related instruments, the major questionnaire types, how to write effective questions, how to pilot and evaluate an instrument, and how digital tools and artificial intelligence can be used responsibly.

Key Takeaways

• A questionnaire is the instrument containing questions; a survey is the wider process of sampling, data collection, analysis and reporting.
• Questionnaires may collect quantitative, qualitative or mixed data.
• Every question should be connected to a research objective, variable or necessary classification purpose.
• Clear wording does not remove the need for expert review, cognitive interviewing and pilot testing.
• Reliability and validity apply differently to individual factual questions, multi-item scales and complete questionnaires.
• AI can assist with drafting and organization, but human expertise and participant testing remain necessary.

What Is a Questionnaire in Research?

A questionnaire is a systematically organized set of questions or prompts designed to obtain information from people. Respondents may complete it independently, or a researcher or interviewer may ask the questions and record the answers.

A questionnaire can collect information about:

• Demographic or background characteristics
• Knowledge and factual understanding
• Behaviours and experiences
• Attitudes, beliefs and perceptions
• Preferences and intentions
• Symptoms and self-reported outcomes
• Satisfaction and evaluation
• Explanations provided in respondents’ own words

Standardization is one of the questionnaire’s defining characteristics. When respondents receive the same carefully worded questions and response options, researchers can compare and aggregate their answers. Standardization does not guarantee valid results, however. Poor wording, incomplete response choices, inappropriate sampling and low participation can still introduce error.

Questionnaire, Survey, Interview, Scale and Inventory

These terms are related but are not identical.

Term	Meaning	Example
Questionnaire	A set of questions or prompts used to collect information	A 25-item student-experience questionnaire
Survey	The wider research process, including objectives, sampling, questionnaire administration, analysis and reporting	A national survey of student wellbeing
Interview	A data-collection interaction in which an interviewer asks questions and may probe responses	A semi-structured interview with doctoral students
Instrument	An umbrella term for a tool used to collect or measure data	Questionnaire, test, checklist or observation form
Scale	One or more items designed to measure a particular construct, often combined into a score	A five-item academic-confidence scale
Inventory	A broader standardized instrument that may assess several dimensions or traits	A multidimensional personality inventory

A questionnaire can contain several scales, but not every questionnaire is a scale. For example, a questionnaire may include age, programme of study, hours worked, satisfaction ratings and two open-ended questions. These items serve different purposes and should not automatically be combined into one score.

Similarly, a questionnaire may be used within a survey, experiment, programme evaluation, clinical study or mixed-methods project.

Is a Questionnaire Qualitative or Quantitative?

A questionnaire can collect quantitative data, qualitative data, or both.

Quantitative questionnaire data

Closed-ended questions produce responses that can normally be coded numerically. Examples include:

• Yes or no
• A selected category
• A frequency
• A rating
• A rank order
• A numerical amount

These responses can be summarized with frequencies, percentages, averages where appropriate, cross-tabulations, correlations or statistical models.

Qualitative questionnaire data

Open-ended questions allow participants to answer in their own words. For example:

What was the most difficult part of completing your research project?

Responses may be examined using qualitative content analysis, thematic analysis or another documented coding approach.

Mixed questionnaires

Many research questionnaires combine closed-ended items with a small number of open-ended questions. Closed items support comparison, while open responses may reveal explanations, experiences or answer categories that the researcher did not anticipate.

The choice should be driven by the research question rather than the assumption that one format is universally better.

Main Types of Questionnaires

Questionnaires can be classified by structure, administration method, timing, purpose and data format.

Types Based on Structure

Structured questionnaire

A structured questionnaire uses predetermined questions, wording, order and response options. It is commonly associated with quantitative research.

Example:

During the past seven days, on how many days did you study for at least one hour?
0 / 1 / 2 / 3 / 4 / 5 / 6 / 7

Structured questionnaires improve consistency and usually simplify coding. Their main limitation is that predefined choices may not capture every relevant experience.

Unstructured questionnaire

An unstructured questionnaire contains broad, open questions and gives respondents considerable freedom in how they answer.

Example:

Describe how your workload affected your academic experience this semester.

It can produce detailed information but requires more time to complete, code and interpret. In practice, a fully unstructured written questionnaire may resemble a set of written qualitative prompts.

Semi-structured questionnaire

A semi-structured questionnaire combines standardized items with open-ended questions.

Example:

How satisfied are you with the feedback provided by your supervisor?
Very dissatisfied / Dissatisfied / Neither / Satisfied / Very satisfied

Please explain the main reason for your answer.

This format supports comparison while preserving room for explanation.

Types Based on Administration

Self-administered questionnaire

Respondents complete the instrument without an interviewer. It may be delivered:

• Online
• On paper
• By email
• Through a mobile application
• By post
• Through a kiosk or tablet

Self-administration can reduce cost and interviewer influence, but respondents cannot always obtain clarification.

Interviewer-administered questionnaire

An interviewer reads the questions and records the answers:

• Face to face
• By telephone
• Through video communication
• In a field-data-collection application

This approach can help respondents with literacy, language or accessibility needs. However, interviewer behaviour and social presence may influence answers, so training and standardized instructions are important.

Types Based on Timing

Cross-sectional questionnaire

Data are collected at one point or during one defined period. The resulting information provides a snapshot.

Longitudinal questionnaire

Questionnaires are administered at multiple points to examine change. A longitudinal design may follow the same participants or repeatedly sample the same population.

Questions, response options, mode and order may need to remain consistent when the purpose is to compare results over time. Changes to wording can create apparent trends that reflect measurement changes rather than real changes.

Types Based on Purpose

Questionnaires may be:

• Exploratory: used to investigate an insufficiently understood topic.
• Descriptive: used to describe characteristics, experiences or opinions.
• Evaluative: used to assess a programme, service, intervention or course.
• Diagnostic or screening: used to identify possible characteristics or needs, subject to appropriate validation and professional limits.
• Explanatory or analytical: used to examine relationships among measured variables.
• Tracking: repeated to monitor change over time.

A questionnaire alone normally cannot establish causation. Causal conclusions require a research design that addresses alternative explanations, such as a randomized experiment or a strong quasi-experimental design.

Types of Questions in a Questionnaire

Question type	Example	Best used for	Main caution
Dichotomous	Have you used the library this month? Yes/No	Clear two-category classification	Reality may not be genuinely binary
Single-choice	What is your primary mode of study?	Mutually exclusive categories	Include all reasonable choices
Multiple-response	Which research tools have you used? Select all that apply	Several applicable categories	Analysis differs from single-choice data
Rating	Rate the clarity of the instructions from 0 to 10	Intensity or evaluation	Label scale endpoints clearly
Likert-type item	I feel confident interpreting statistical results	Agreement, frequency or similar ordered judgments	One item is not automatically a full Likert scale
Semantic differential	Difficult 1 2 3 4 5 Easy	Evaluation between opposite descriptors	Opposites must be meaningful
Ranking	Rank these five resources by usefulness	Relative priority	Becomes burdensome with long lists
Numeric entry	How many hours did you study last week?	Counts and quantities	Define units and realistic limits
Open-ended	What should be improved?	Unanticipated detail and explanation	Greater respondent and analysis burden
Matrix	Rate several services using the same scale	Repeated ratings	Large grids perform poorly on small screens
Contingency	If “Yes,” answer Question 8	Relevant follow-up questions	Routing must be thoroughly tested

Likert Items and Likert Scales

The terms are frequently used imprecisely.

A Likert-type item is an individual statement followed by ordered response categories, such as:

The questionnaire instructions were clear.
Strongly disagree / Disagree / Neither agree nor disagree / Agree / Strongly agree

A Likert scale more properly refers to a set of related items intended to measure the same construct, with responses combined according to a defined scoring method.

Before summing several items:

1. Establish that they represent the intended construct.
2. Specify whether any items are reverse-scored.
3. Examine the scale’s dimensionality and reliability.
4. Follow the scoring instructions for an established instrument.
5. Explain how missing responses are handled.

Adding unrelated rating items together does not create a valid scale.

Response Scales and Levels of Measurement

The form of the response affects what can be concluded and which analyses are defensible.

Nominal responses

Categories have no inherent order.

Examples:

• Department
• Country
• Employment status
• Preferred learning platform

Ordinal responses

Categories have a meaningful order, but the distance between categories is not known to be equal.

Examples:

• Never, rarely, sometimes, often, always
• Dissatisfied, neutral, satisfied
• Class rank

An individual Likert-type response is generally treated as ordinal.

Interval-level measurements

Values are ordered and interpreted as having equal intervals, but zero is not an absolute absence of the quantity. Many psychological scale scores are analyzed as approximately continuous when their development and distribution support that decision.

Ratio-level measurements

Values have meaningful intervals and a meaningful zero.

Examples:

• Age
• Number of publications
• Hours studied
• Income, when measured as an amount

Researchers should decide how variables will be analyzed before choosing question formats.

Components of a Research Questionnaire

A well-organized questionnaire commonly contains the following elements.

1. Title

Use a short, neutral title that does not reveal a desired outcome.

2. Participant information and consent

Explain:

• Who is conducting the research
• Why the study is being conducted
• What participation involves
• Estimated completion burden
• Whether participation is voluntary
• Possible risks or discomfort
• How data will be stored and used
• Whether responses are anonymous or confidential
• Whom to contact
• How consent is indicated

The precise requirements depend on the institution, jurisdiction, population and research risk.

3. Eligibility or screening questions

Use only when participation is limited to a defined group.

4. Instructions

Explain how to answer, whether multiple selections are allowed, which time period to consider and what unfamiliar terms mean.

5. Warm-up questions

Begin with relevant questions that are reasonably easy and non-threatening.

6. Main content sections

Group items by topic or construct. Provide transitions when the subject changes.

7. Sensitive questions

Place sensitive or demanding questions only where necessary. Explain why information is needed and permit participants to decline when appropriate.

8. Classification questions

Demographic or background items are often placed later unless they are needed for eligibility or routing.

9. Closing message

Thank respondents, provide submission instructions and repeat relevant support or contact information.

How to Design a Questionnaire

Questionnaire development should begin with the research problem, not with a survey platform.

Step 1: Define the research objectives

Write down exactly what the research must discover.

A vague objective such as “understand student experience” is too broad. A more useful objective is:

To examine postgraduate students’ satisfaction with the timeliness, clarity and usefulness of supervisor feedback.

Step 2: Identify the constructs and variables

Translate the objectives into measurable concepts.

Research objective	Construct or variable	Possible indicator
Assess feedback timeliness	Waiting time	Number of days before feedback
Measure feedback clarity	Perceived clarity	Multi-item rating scale
Explore improvement needs	Suggested improvements	Open-ended response

Abstract concepts such as motivation, trust, anxiety or satisfaction need clear conceptual and operational definitions.

Step 3: Search for existing instruments

Before creating new items, search academic databases, instrument repositories, published studies and official organizations.

For an existing instrument, check:

• What construct it measures
• The population in which it was developed
• Available languages
• Evidence of reliability and validity
• Scoring instructions
• Administration requirements
• Licensing or permission conditions
• Whether adaptation is permitted

Using a recognized instrument can support comparability, but a scale validated in one population is not automatically valid for every language, country, age group or mode of administration.

Step 4: Define the target population and mode

Consider respondents’:

• Age
• Language
• Literacy
• Subject knowledge
• Disability and accessibility needs
• Internet and device access
• Cultural context
• Willingness to discuss sensitive topics

The administration method influences wording, layout, routing and required instructions.

Step 5: Create a questionnaire blueprint

A blueprint prevents unnecessary questions and connects the instrument to the analysis.

Item	Objective	Variable	Question format	Planned analysis
Q1	Confirm eligibility	Current enrolment	Yes/No	Exclusion rule
Q5	Measure feedback delay	Days waited	Numeric	Median and distribution
Q8–Q11	Measure feedback clarity	Clarity score	Rating items	Scale evaluation and group comparison
Q15	Identify improvements	Suggested change	Open text	Thematic coding

If a question has no clear purpose, reconsider whether it should be included.

Step 6: Select suitable question and response formats

Choose formats based on:

• The information required
• Respondents’ ability to answer
• Sensitivity
• Desired level of measurement
• Analysis plan
• Device and mode

Do not choose a five-point agreement scale merely because it is familiar. Frequency, quality, likelihood or direct behavioural categories may better match the construct.

Step 7: Draft clear questions

Each question should:

• Address one idea
• Use familiar language
• Specify a time period where needed
• Avoid assumptions
• Avoid unnecessary technical terms
• Avoid suggesting a preferred answer
• Be answerable by the intended respondent
• Match the provided response options

Step 8: Design complete response options

Closed response categories should normally be:

• Mutually exclusive
• Collectively adequate
• Consistent with the question
• Ordered logically
• Clearly labelled

Consider “Other—please specify,” “Not applicable,” “Do not know,” or “Prefer not to answer” only when conceptually appropriate. These options are not interchangeable.

Step 9: Arrange the order and routing

A common sequence is:

1. Eligibility
2. Easy introductory questions
3. Core topic questions
4. More complex or sensitive questions
5. Demographic or classification questions
6. Closing comments

Use skip logic so respondents see only applicable questions. Test every route, including unusual combinations of answers.

Question order can influence later responses, so place broad measures before highly specific prompts when priming is a concern.

Step 10: Obtain expert and respondent feedback

Subject specialists can assess whether important content is missing. Methodologists can examine wording, measurement and analysis. Members of the target population can reveal unexpected interpretations.

Cognitive interviewing is particularly valuable. Participants may be asked:

• What does this question mean to you?
• How did you remember that information?
• What did you consider when choosing your answer?
• Was an appropriate response option available?
• What does a particular term mean in your own words?

Step 11: Pilot-test the questionnaire

A pilot examines the complete process rather than only individual wording.

Assess:

• Recruitment
• Consent
• Completion time
• Device compatibility
• Skip logic
• Missing responses
• Response distributions
• Technical errors
• Participant burden
• Data export
• Coding and scoring
• Preliminary measurement performance

Pilot participants should resemble the intended population. Pilot data should not automatically be merged into the main dataset when the instrument or procedures are changed.

Step 12: Finalize and document the instrument

Before launch:

• Freeze the approved version.
• Assign version numbers and dates.
• Test all routes.
• Check scoring and reverse coding.
• Confirm privacy settings.
• Prepare a codebook.
• Record the source of adapted items.
• Document permissions.
• Confirm the analysis plan.
• Obtain required ethics or institutional approvals.

How to Write Effective Questionnaire Questions

Use specific language

Weak:

Do you study regularly?

“Regularly” may mean daily to one respondent and once a week to another.

Better:

During the past seven days, on how many days did you study outside scheduled classes?

Ask one thing at a time

Weak:

How satisfied are you with the speed and usefulness of supervisor feedback?

A respondent may consider the feedback fast but unhelpful.

Better:

How satisfied are you with the time taken to receive feedback?

How useful was the feedback for improving your work?

Avoid leading language

Weak:

How beneficial was the university’s excellent research workshop?

Better:

Overall, how useful or not useful was the research workshop?

Avoid loaded assumptions

Weak:

Why did you fail to use the library’s research support?

This assumes failure and may not apply.

Better:

Have you used the library’s research-support service during the current academic year?

Respondents who answer “No” can receive an appropriate follow-up question.

Specify a reasonable recall period

Weak:

How often have you contacted your supervisor?

Better:

During the past four weeks, how many times did you contact your supervisor about your research?

Use a period respondents can reasonably remember and that matches the research objective.

Avoid double negatives

Weak:

Do you disagree that online classes should not be recorded?

Better:

Should online classes be recorded for later viewing?

Do not combine questions with explanations that influence answers

Background information may be necessary, but persuasive evidence or emotionally framed statements can prime respondents.

Avoid unnecessary forced responses

Making every online item mandatory may reduce missing data but can also produce false answers or prevent participants from completing the questionnaire. Sensitive questions should generally include a suitable way not to answer.

Example Questionnaire Template

Title

Postgraduate Research Support Questionnaire

Participant information

You are invited to participate in a study examining postgraduate students’ experiences of research support. Participation is voluntary. The questionnaire takes approximately [tested completion time] to complete. Your responses will be handled according to the study’s approved data-management plan.

Consent

• I have read the participant information.
• I understand that participation is voluntary.
• I understand how my information will be used.
• I consent to participate.

Yes / No

Section A: Eligibility

1. Are you currently enrolled in a postgraduate research programme?
   • Yes
   • No
2. Have you received academic supervision during the current semester?
   • Yes
   • No

Section B: Research support

3. During the past four weeks, how many scheduled meetings did you have with your supervisor?
   • 0
   • 1
   • 2
   • 3
   • 4 or more
4. How satisfied or dissatisfied were you with the availability of your supervisor?
   • Very dissatisfied
   • Dissatisfied
   • Neither satisfied nor dissatisfied
   • Satisfied
   • Very satisfied
   • Not applicable
5. Please indicate your response to each statement:
   • Feedback identified specific areas for improvement.
   • Feedback was written in language I could understand.
   • Feedback helped me decide what to do next.
   Response options:
   • Strongly disagree
   • Disagree
   • Neither agree nor disagree
   • Agree
   • Strongly agree

Section C: Resources

6. Which research-support resources have you used this academic year? Select all that apply.
   • Library consultation
   • Research-methods workshop
   • Statistical consultation
   • Writing centre
   • Online research guide
   • Peer research group
   • Other: ______
   • None of these

Section D: Open response

7. What single change would most improve the research support available to you?

Section E: Background information

8. What is your programme level?
   • Master’s
   • Professional doctorate
   • PhD or equivalent
   • Other: ______
9. What is your primary mode of study?
   • Full-time
   • Part-time
   • Other: ______

Closing message

Thank you for completing the questionnaire. For questions about the study, contact [research contact]. For concerns about your rights as a participant, contact [independent ethics contact].

This example must be adapted to the study’s objectives, population, jurisdiction, ethics requirements and analysis plan. It is not a validated measurement instrument.

Pilot Testing and Cognitive Interviewing

Pilot testing and cognitive interviewing are related but perform different functions.

Method	Main purpose
Expert review	Assess content coverage, relevance, wording and disciplinary accuracy
Cognitive interview	Discover how respondents interpret and answer each question
Technical test	Check display, validation, routing, links, data capture and export
Pilot study	Test the full recruitment, administration and data-management process
Psychometric study	Evaluate the properties of a multi-item measure

A pilot that produces no technical problems does not prove that respondents interpreted every item as intended. Likewise, a small group saying that the questionnaire “looks fine” does not establish construct validity.

Questionnaire Validity

Validity concerns whether the interpretation and use of questionnaire results are supported by adequate evidence.

Face validity

Face validity asks whether an instrument appears reasonable and relevant. It may improve respondent acceptance but is weak evidence by itself.

Content validity

Content validity concerns whether items adequately cover the intended construct. It includes:

• Relevance
• Comprehensiveness
• Comprehensibility

Subject experts and members of the target population may contribute different but complementary evidence.

Construct validity

Construct validity examines whether score patterns are consistent with the theory of the construct.

Evidence may include:

• Internal structure
• Relationships with related measures
• Differences between groups expected to differ
• Lack of strong relationships with unrelated constructs
• Factor analysis
• Measurement invariance across important groups

Criterion-related validity

Criterion evidence examines relationships with an appropriate external criterion or outcome. A suitable criterion may not exist for abstract constructs such as attitudes or perceived belonging.

Cross-cultural validity

A translated or adapted instrument should measure the construct comparably across languages or cultural groups. Literal translation alone is not sufficient.

Questionnaire Reliability

Reliability concerns the consistency or precision of measurement.

Internal consistency

Internal consistency concerns relationships among items intended to measure the same construct. It is not relevant to an arbitrary collection of unrelated questions.

A high coefficient does not prove that:

• The scale is unidimensional.
• The content is valid.
• The wording is unbiased.
• The instrument is appropriate for a different population.

Test–retest reliability

The same measure is administered again after a suitable interval to examine score stability when the underlying construct is expected to remain stable.

Inter-rater reliability

This may be relevant when interviewers classify answers or researchers code open-ended responses.

Measurement error

Researchers should also consider the precision of individual scores and whether observed changes are larger than expected measurement error.

When Does a Questionnaire Need Psychometric Validation?

Full psychometric development is most relevant when researchers create or substantially adapt a multi-item scale intended to measure a latent construct.

It may involve:

1. Defining the construct.
2. Generating an item pool.
3. Expert and respondent assessment.
4. Cognitive interviews.
5. Pilot administration.
6. Item analysis.
7. Exploratory factor analysis.
8. Confirmatory factor analysis in separate data where feasible.
9. Reliability assessment.
10. Relationships with relevant variables.
11. Cross-group or longitudinal equivalence.
12. Independent replication.

A questionnaire containing age, employment status, factual behaviours and one open comment does not become a psychometric scale merely because it contains several questions.

Ethical and Privacy Considerations

Researchers should follow applicable institutional, professional and legal requirements.

Informed participation

Participants should receive sufficient information to decide whether to participate. Consent language should be understandable to the intended population.

Voluntary participation

Researchers should avoid undue pressure. Students, employees, patients and other dependent groups may feel unable to refuse when recruitment is conducted by someone with authority over them.

Anonymity and confidentiality

These terms should not be used interchangeably.

• Anonymous data cannot reasonably be linked to an identified participant by the research team.
• Confidential data may be identifiable, but access and disclosure are restricted.

An online platform may record email addresses, login identifiers, IP information or other metadata. Researchers should examine the actual settings before promising anonymity.

Data minimization

Collect only information needed for the research. Avoid gathering names, precise locations or contact information without a justified purpose.

Sensitive questions

Explain why sensitive information is needed, use appropriate categories, permit nonresponse where suitable and provide support information when questions could cause distress.

Research approval

Research involving human participants may require review by an Institutional Review Board, Research Ethics Committee or equivalent body. Approval requirements differ by institution and jurisdiction.

Accessibility and Inclusive Design

A questionnaire should be usable by the intended population.

Good practices include:

• Plain, direct language
• Readable font size
• Sufficient contrast
• Keyboard navigation
• Proper labels for screen readers
• Alternatives to visual-only information
• Short matrix questions
• Mobile-responsive layout
• Clear error messages
• No unnecessary time limit
• Language support where needed
• Accessible consent materials

Accessibility should be tested rather than assumed.

Translation and Cross-Cultural Adaptation

Translation should preserve meaning, not merely substitute words.

A rigorous process may include:

1. Review of the construct in the target culture.
2. Independent translation or a team translation process.
3. Review by language and subject specialists.
4. Reconciliation or adjudication.
5. Review of response scales and examples.
6. Cognitive interviews with target-language participants.
7. Pilot testing.
8. Documentation of every adaptation.
9. Examination of cross-language measurement properties when scores will be compared.

Previously validated questions should not be casually rewritten. Even small wording changes may alter interpretation and comparability.

Administering a Questionnaire

The administration method should match the population and research aims.

Online administration

Advantages include automated routing, rapid distribution, validation rules and direct data export.

Risks include:

• Exclusion of people with limited internet access
• Device-related display problems
• Duplicate responses
• Automated or fraudulent submissions
• Unclear identity verification
• Low response rates
• Privacy concerns

Paper administration

Paper can be appropriate where internet access is limited or face-to-face distribution is practical. Data entry and routing errors require careful management.

Interviewer administration

This may improve inclusion and clarification but requires interviewer training, supervision and monitoring of interviewer effects.

Mixed-mode administration

Using more than one mode may improve coverage, but differences in visual layout, interviewer presence and response process can create mode effects. Questions should be designed and tested for comparability.

Analyzing Questionnaire Data

Planning should occur before data collection.

Create a codebook

For every variable, record:

• Variable name
• Question wording
• Response labels
• Numeric codes
• Missing-value codes
• Multiple-response rules
• Scoring instructions
• Reverse-coded items
• Derived variables
• Permitted ranges

Examine data quality

Check for:

• Missing responses
• Impossible values
• Inconsistent routes
• Duplicate submissions
• Extremely rapid completion
• Repeated response patterns
• Bot or fabricated responses
• Outliers
• Failed attention or logic checks, where justified

Exclusion rules should be defined transparently rather than selected only after seeing which removals produce preferred findings.

Analyze closed-ended questions

Depending on the research design and variable type, analysis may include:

• Frequencies and percentages
• Medians or means where appropriate
• Cross-tabulations
• Confidence intervals
• Group comparisons
• Correlations
• Regression models
• Scale scores
• Weighting for complex or unequal samples

The questionnaire does not determine whether findings can be generalized. Sampling, recruitment, response and weighting are also important.

Analyze open-ended questions

Researchers may:

1. Read and familiarize themselves with the responses.
2. Develop or apply a coding framework.
3. Code relevant segments.
4. Compare coding decisions.
5. Identify patterns or themes.
6. Select illustrative extracts carefully.
7. Document the analytic method.

AI-assisted coding or summarization should be checked against the original responses. Researchers must consider whether respondent data may legally and ethically be entered into an external AI system.

Handle missing data

Distinguish among:

• Question not shown because of routing
• Not applicable
• Do not know
• Prefer not to answer
• Accidental omission
• Participant dropout
• Technical failure

These categories should not automatically receive the same code.

Advantages of Questionnaires

Standardization

Respondents can receive consistent questions, which supports comparison.

Scalability

Online and paper questionnaires can reach geographically dispersed populations.

Efficiency

Closed-ended data can be collected and processed relatively efficiently.

Flexibility

A questionnaire can collect factual, behavioural, attitudinal and open-text data.

Reduced interviewer influence

Self-administration can reduce direct interviewer effects.

Anonymity potential

When genuinely anonymous procedures are used, participants may be more willing to report sensitive information.

Replicability

Documented questions and procedures can be reused or examined by other researchers.

Limitations of Questionnaires

Self-report error

Respondents may forget, misunderstand, exaggerate, minimize or provide socially acceptable answers.

Limited explanation

Closed questions can prevent participants from expressing unanticipated experiences.

Nonresponse bias

People who complete the questionnaire may differ systematically from those who do not.

Coverage problems

An online questionnaire excludes people who cannot access or use the selected platform.

Wording and order effects

Small changes in phrasing, context or response order can affect answers.

Response burden

Long, repetitive or cognitively demanding questionnaires may increase missing data, satisficing and dropout.

Weak causal evidence

Questionnaire data often describe variables or associations rather than demonstrating cause and effect.

False precision

Numeric response scales may appear exact even when the underlying concept is ambiguous or poorly measured.

Common Questionnaire Mistakes

1. Writing questions before defining objectives.
2. Using a new instrument without searching for established measures.
3. Asking two questions in one item.
4. Using undefined terms such as “regularly” or “often.”
5. Providing overlapping response categories.
6. Omitting “not applicable” when it is a genuine possibility.
7. Using agreement scales for every construct.
8. Mixing scale directions without a clear methodological reason.
9. Making all questions mandatory.
10. Using long matrix questions on mobile devices.
11. Changing validated wording without documenting the change.
12. Claiming validation based only on expert approval or Cronbach’s alpha.
13. Launching without testing routing and data export.
14. Collecting identifying information unnecessarily.
15. Choosing analyses after examining the results.

Questionnaires in Modern Research

Modern questionnaires are increasingly:

• Mobile-first
• Multilingual
• Administered across several modes
• Connected to electronic research databases
• Designed with adaptive routing
• Integrated with longitudinal panels
• Combined with interviews or administrative data
• Monitored for bots, duplicates and low-quality responses
• Developed with greater attention to transparency and reproducibility
• Analyzed with both statistical and qualitative techniques

Questionnaire versions, codebooks, scoring rules and analysis plans may be shared in repositories where consent, licensing and privacy conditions allow.

Digital Questionnaire Tools

The appropriate platform depends on the sensitivity, complexity and setting of the research.

Tool category	Examples	Appropriate use
Basic form builders	Google Forms, Microsoft Forms	Low-complexity questionnaires where institutional rules permit
Advanced survey platforms	Qualtrics, SurveyMonkey	Branching, randomization, panels and detailed survey workflows
Research data platforms	REDCap	Institutionally managed research and clinical data collection
Offline field tools	KoboToolbox	Fieldwork with intermittent internet access
Open-source platforms	LimeSurvey	Self-hosting, customization and greater infrastructure control

Before choosing a platform, assess:

• Institutional approval
• Data-storage location
• Encryption and access control
• Anonymous-response settings
• Export formats
• Audit trails
• Branching logic
• Accessibility
• Offline operation
• Multilingual support
• Data-retention controls
• Regulatory requirements

A familiar platform is not automatically suitable for sensitive research.

Can Artificial Intelligence Create a Questionnaire?

AI can assist with questionnaire development, but it should not independently determine what is measured or whether an instrument is valid.

Appropriate AI-assisted tasks

AI may help researchers:

• Brainstorm possible domains after a literature review.
• Identify repeated or unusually complex wording.
• Generate alternative plain-language drafts.
• Build an initial questionnaire blueprint.
• Suggest response-format options.
• Create a draft codebook.
• Test hypothetical skip-logic scenarios.
• Organize open-text responses for human review.
• Detect potentially duplicate items.

Tasks that still require human judgment

AI cannot independently:

• Define the construct correctly.
• Establish content validity.
• Determine whether wording is culturally appropriate.
• Replace cognitive interviewing.
• Grant ethics approval.
• Establish reliability or measurement invariance.
• Decide whether a licensed instrument may be reproduced.
• Guarantee freedom from bias.
• Protect confidential data unless used within an approved environment.

Researchers should verify all AI-generated items against theory, evidence, expert judgment and feedback from the target population. Prompts and AI outputs may also need to be documented when they materially influence the research instrument.

How to Report a Questionnaire in a Research Paper

A methods section should normally state:

• Whether the questionnaire was existing, adapted or newly developed
• The constructs and variables measured
• Instrument source and permission status
• Number and type of questions
• Response scales
• Administration mode
• Language and translation process
• Cognitive interviewing and pilot procedures
• Evidence of reliability and validity
• Recruitment and sampling method
• Completion and missing-data rules
• Scoring procedures
• Data-quality checks
• Ethics approval and consent process
• Where the questionnaire can be accessed, where permitted

Avoid writing only, “Data were collected using a researcher-made questionnaire.” Readers need enough detail to evaluate and reproduce the method.

Conclusion

A questionnaire is a structured research instrument used to collect information from respondents through closed questions, open questions, scales or mixed formats. Its usefulness depends on more than the number of responses collected. Strong questionnaires connect every item to a research objective, use suitable response options, minimize avoidable bias, protect participants, undergo testing and produce data that can be analyzed as planned.

For simple descriptive projects, this may require a clear blueprint, expert feedback and a careful pilot. Developing a new scale for a complex construct requires substantially stronger psychometric evidence. In either case, the quality of the eventual findings begins with the quality of the questions.