Survey question types: the complete guide with examples

The two families: closed-ended and open-ended

Closed-ended questions give respondents predefined answers to choose from; they're fast to answer and easy to chart, but only measure what you thought to ask. Open-ended questions collect free text; they're slower to answer and require coding to analyze, but they catch the unknown unknowns. Choose the type per question based on what you'll do with the answer — every type below exists because it fits a specific analytical job.

1. Multiple choice (single select)

Example: "How did you hear about us? Search engine / Social media / A friend or colleague / Advertisement / Other." The workhorse of surveys: one question, one answer. Use it whenever options are known and mutually exclusive. Watch out for incomplete option lists — always test whether an "Other" is needed. Analysis is simple percentages.

2. Checkboxes (multi-select)

Example: "Which of these features do you use? (Select all that apply.)" Use when several answers can be true at once. The analysis trap: percentages sum to more than 100, so report "% of respondents selecting each option" and never use a pie chart. Long checkbox lists invite lazy top-of-list selection — consider randomizing option order.

3. Dropdown

Example: "Which country do you live in?" Functionally a single-select, best when the option list is long (countries, states, industries) and would clutter the screen as radio buttons. Avoid dropdowns for short lists of 5 or fewer — visible options are faster and get more deliberate choices.

4. Rating scale

Example: "How would you rate your checkout experience?" on a 1–5 star or numbered scale. Use for quick quality judgments of a single item. Decide the number of points deliberately (5 is the everyday default; see our rating scales guide) and label at least the endpoints ("1 = Very poor, 5 = Excellent"). Analyze with means plus top-box percentage.

5. Likert scale

Example: "The onboarding made it easy to get started." — Strongly disagree / Disagree / Neither agree nor disagree / Agree / Strongly agree. A Likert item measures agreement with a statement; it's the standard for attitudes and perceptions (engagement, satisfaction drivers). Keep statements one-directional and neutral, and keep the scale balanced. Report distributions and top-two-box, not just means.

6. NPS (Net Promoter Score)

Example: "How likely are you to recommend us to a friend or colleague?" on 0–10. A standardized loyalty metric: 9–10 are promoters, 7–8 passives, 0–6 detractors, and the score is % promoters minus % detractors. Use the exact standard wording so scores stay comparable over time. Always pair it with an open follow-up ("What's the main reason for your score?").

7. Ranking

Example: "Rank these five features from most to least important to you." Forces trade-offs that rating questions hide — respondents can rate everything "very important," but they can't rank everything first. Keep lists to 5–7 items; ranking 12 things is guesswork after the top few. Analyze with % ranked first and average rank together.

8. Matrix / grid

Example: rows of items (Support, Pricing, Ease of use, Documentation) each rated on the same Satisfied–Dissatisfied scale. Compact for desktop respondents and consistent across items, but the most fatigue-prone format: big grids trigger straight-lining and are painful on phones. Keep matrices to 5–7 rows or split them into individual rating questions.

9. Semantic differential

Example: "How would you describe our brand?" with a 7-point slider between opposite adjectives: Traditional ↔ Modern, Complicated ↔ Simple. Unlike Likert (agree/disagree with one statement), semantic differential positions perception between two poles — ideal for brand image and product personality work. Make sure the adjective pairs are true opposites.

10. Open-ended (free text)

Example: "If you could change one thing about the product, what would it be?" The only type that captures reasons in respondents' own words. Use sparingly — one or two per survey — and place them after the related closed question. Analysis requires coding answers into themes, so budget time for it.

11. Demographic questions

Example: age bracket, role, company size, location. Not a distinct input format but a distinct discipline: ask only what you'll use for segmentation, use broad non-overlapping buckets, place them at the end, and always include Prefer not to say. Sensitive demographics handled carelessly are a top cause of abandonment.

12. Date, number, and validated fields

Example: "When did you attend the event?" (date picker) or "How many employees does your company have?" (number field). Validated inputs — date, number, email — prevent the formatting chaos of collecting these via free text. Use them whenever the answer has a natural data type.

Summary: which question type should you use?

How do you combine types in one survey?

1. 1Open with an easy single-select or rating question tied to the survey topic.
2. 2Use closed questions (choice, rating, Likert) for everything you need to quantify.
3. 3Follow your most important closed question with one open "why" question.
4. 4Use ranking once, where priorities genuinely conflict.
5. 5Close with optional demographics in broad buckets.

Most modern builders cover this full toolkit. Formkii's survey maker includes multiple choice, checkboxes, dropdowns, open text, rating scales, NPS, drag-and-drop ranking, and date questions — and its form builder adds 18 field types with conditional logic if you need validated inputs like email, number, or file upload.

Frequently asked questions