# How do you write a good survey questionnaire?

> Written with AI · [Communications: research methods in communication](https://lessoons.com/en/communication-research-methods) · Grade 12 · 2026-11-13 · https://lessoons.com/en/communication-research-methods/how-do-you-write-a-good-survey-questionnaire

To write a good survey questionnaire, decide exactly what you want to measure, ask one thing per question in simple neutral words, avoid leading, double-barreled and loaded questions, give balanced answer options such as a five-point Likert scale, order the questions from easy to sensitive, and test the draft on a few people before sending it. Respondents should know the survey is voluntary and anonymous.

## The topic explained

### Start with what you want to know

Writing a good survey questionnaire starts before the first question: decide exactly what you want to measure. “What do students think about the news?” is too vague. “How often do 12th graders check the source of a news story before sharing it?” can be measured. Every question in the questionnaire should serve that aim; a question you will not analyze is a question that tires respondents for nothing.

Then choose the form of each question. A closed question offers fixed answers, yes or no, a list or a scale, and is quick to answer and easy to count. An open question lets people answer in their own words; it brings reasons and surprises but takes longer to analyze. Most questionnaires are mainly closed, with one or two open questions where the researcher wants to hear explanations.

### The traps

A leading question suggests the answer: “Social media harm young people, don't they?” A double-barreled question asks two things at once: “Is the website fast and clear?” A person who finds it fast but confusing cannot answer. A loaded question uses a word that judges: “How much time do you waste online?” assumes the time is wasted. Unbalanced options tilt the result: if the answers are good, very good and excellent, no one can say poor.

Other traps are jargon that respondents may not know, double negatives (“Do you not disagree…?”), and questions that ask about the distant past or about behavior people prefer to hide. The common mistake is to think that a questionnaire measures what people do; it measures what people say they do, which is why neutral wording and anonymity matter so much.

### Scales and scores

Attitudes are often measured with a Likert scale, named after the psychologist Rensis Likert, who described the method in 1932. The respondent reads a statement and marks how far they agree, usually on five points from strongly disagree (1) to strongly agree (5). Several statements on the same subject can be averaged into one score. For example, a made-up student who answers 4, 5, 3 and 4 to four statements about checking sources has an average of (4 + 5 + 3 + 4) ÷ 4 = 4.0, near “agree.”

A scale should be balanced, with as many negative as positive options and a clear middle. Researchers sometimes write some statements in the opposite direction, so that a person who ticks “agree” to everything without reading can be spotted.

### Order, testing and ethics

Put easy, neutral questions first and sensitive ones, such as income or personal habits, later, when the respondent is engaged. Keep the questionnaire short. Before sending it, run a pilot: give it to a few people like your respondents and ask what was unclear. Almost every pilot finds a confusing question. Finally, respondents must know what the survey is for, that it is voluntary and that their answers are anonymous or confidential. A strong exam answer explains the traps with examples, rewrites a bad question, and calculates a scale score.

Online forms make it easy to build and send a questionnaire, but they do not solve the problem of who answers. A link shared on social media reaches whoever happens to click, often the people who care most about the subject, not a random sample of the population. A well-written questionnaire sent to the wrong people still gives a misleading picture.

## Practice: writing a questionnaire

For 12th grade communications students. Every question and answer here is made up. Answer from what is printed on the page, then check yourself against the answers.

1. **Open or closed? (a) “How many hours a day do you watch videos? 0-1, 1-2, 2-3, more than 3”; (b) “What makes you trust a news site?”; (c) “Do you have a social media account? yes / no.”** — (a) closed; (b) open; (c) closed.
2. **Rewrite the loaded question in the table as a neutral question, and say which word you removed.** — For example: “How much time do you spend online on a school day?” The new question removes the judging word “waste.”
3. **What is wrong with the question “Do you read the newspaper and watch the news every day?” Rewrite it.** — It is double-barreled: it asks about two habits. Rewrite as two questions: “Do you read a newspaper every day?” and “Do you watch the news every day?”
4. **Which option is missing on the scale, and what score does it carry?** — Disagree, with a score of 2, between strongly disagree (1) and not sure (3).
5. **A made-up student answers 4, 5, 3 and 4 to four Likert statements about checking sources. What is the average score, and what does it mean?** — (4 + 5 + 3 + 4) ÷ 4 = 4.0, which corresponds to “agree”: the student reports usually checking sources.
6. **Five made-up students answer one statement with 2, 3, 2, 1 and 2. What is the class average?** — (2 + 3 + 2 + 1 + 2) ÷ 5 = 10 ÷ 5 = 2.0, which corresponds to “disagree.”
7. **Which step is missing in building the questionnaire, and why is it important?** — The pilot with a few people: it reveals unclear or confusing questions before the real survey, when they can still be fixed.
8. **Write three good questions for a made-up survey on how 12th graders get their news: one closed, one Likert statement and one open.** — For example: closed, “Where do you get most of your news? television, websites, social media, friends, other”; Likert, “I trust the news I see on social media” (1-5); open, “What makes you trust a news source?” A good answer avoids leading and double-barreled wording.

## Common questions

### What is a Likert scale?

A rating scale on which respondents mark how far they agree with a statement, usually from strongly disagree to strongly agree on five points. It is named after Rensis Likert, who described the method in 1932, and the scores can be averaged.

### What is a leading question?

A question whose wording pushes toward a particular answer, such as “Don't you agree that…?” It makes the results reflect the researcher's view rather than the respondent's, so it should be rewritten in neutral words.

### What is a double-barreled question?

A question that asks about two things at once, such as “Is the app fast and easy to use?” A person who thinks it is fast but hard to use cannot answer accurately, so it should be split into two questions.

### How long should a questionnaire be?

As short as the research question allows. Long questionnaires tire respondents, who then answer carelessly or stop. Every question should serve the aim of the study; a pilot test shows how long it really takes.

### Why should a survey be anonymous?

Because people answer more honestly about sensitive subjects when they cannot be identified, and because research ethics require protecting participants. Respondents should also know the purpose of the survey and that taking part is voluntary.
