Every survey question format trades off two things: how easy it is to answer, and how much you can do with the resulting data. A closed multiple-choice question takes five seconds to answer and produces instantly analyzable data; an open-ended question takes longer to answer and longer to analyze, but can surface something you didn't think to ask about directly. Picking the right format for each question, not defaulting to the same one throughout, is what separates a survey that produces a clear answer from one that produces a pile of ambiguous data.
8 types of survey questions
Multiple choice
Respondents pick one (or several) options from a fixed list. Example: "Which feature do you use most, Dashboard, Reports, or Integrations?" Fast to answer and easy to analyze, the most common format for a reason.
Open-ended (short text)
A free-text box with no predefined options. Example: "What's one thing we could improve?" Slower to answer and harder to analyze at scale, but captures detail and unexpected feedback that closed formats can't anticipate.
Closed-ended / yes-no
A binary choice. Example: "Did this resolve your issue? Yes / No." The fastest possible format to answer, best for a single, unambiguous fact you need to confirm.
Rating scale
A numeric or star scale, typically 1–5 or 1–10. Example: "How satisfied were you with onboarding? (1–5 stars)." Good for measuring intensity of opinion, not just direction.
NPS (Net Promoter Score)
A standardized 0–10 "how likely are you to recommend this" question, with a fixed formula for turning results into a single benchmarkable score. Useful when you want a metric you can track over time and compare against industry norms.
Ranking
Respondents order a list of options by preference, rather than picking just one. Example: "Rank these three roadmap items in order of priority to you." Captures more information than a single-choice question, at the cost of being slightly slower to answer.
Matrix / grid
Several related items rated on the same scale in one compact table. Example: rating five different product aspects, all on the same 1–5 scale, in a single grid. Efficient for the respondent, but can suffer from "straight-lining" (clicking the same column down the whole grid) if overused.
Categorical / demographic
Fixed, non-ranked categories, department, role, tenure, region. Example: "Which team are you on?" Usually used to segment other answers rather than as a standalone insight.
A well-designed survey mixes formats deliberately: closed formats for the bulk of the survey to keep completion rates high, with one or two open-ended questions reserved for where nuance genuinely matters.
Leading vs. neutral question phrasing
A leading question nudges the respondent toward a particular answer through its phrasing, "how much did you love the new checkout flow?" presumes a positive experience before the respondent has said anything. This produces data that looks clean and positive but is quietly unreliable, you're measuring how the question was worded, not what people actually think. The fix is straightforward: phrase every question as neutrally as possible. "How would you rate the new checkout flow?" gathers the same information without steering the answer. This matters even more for sensitive or evaluative questions, employee satisfaction, manager feedback, product quality, where a leading phrasing can meaningfully distort results in a direction that flatters whoever wrote the survey.
Choosing the right format for what you actually need
Start from the decision the data needs to support, not from a default template. If you need a single, trackable metric over time (is satisfaction trending up or down this quarter), a consistent rating or NPS question run on a fixed cadence beats a one-off detailed survey. If you need to understand why something is happening, not just whether it is, include open-ended questions even though they're harder to aggregate. If you're trying to prioritize between several options rather than just measure sentiment, a ranking question (or, for a live group decision rather than a survey, a structured vote) gives you far more useful signal than asking people to rate each option independently. Chooseday's forms support rating, NPS, multiple-choice, and open-text questions today; a continuous slider scale and image-based answer options aren't supported yet, if your survey specifically needs those, a dedicated survey tool may be a better fit for that piece.
Frequently asked questions
The main types are: multiple choice (pick one or several from a list), open-ended (free text response), closed-ended/yes-no (binary), rating scale (numeric or star rating), NPS (0–10 likelihood-to-recommend), ranking (order options by preference), and matrix/grid questions (rate several items on the same scale). Most real surveys mix several types, closed formats for fast, analyzable data, and a couple of open-ended questions for detail and nuance.
A leading question is phrased in a way that nudges the respondent toward a particular answer, for example "how much did you enjoy our excellent new feature?" assumes the feature was excellent before the respondent has answered. Leading questions produce data that looks clean but is quietly biased, always phrase questions neutrally: "how would you rate the new feature?" instead.
Closed-ended questions give the respondent a fixed set of options (multiple choice, yes/no, rating scale), producing data that's fast to answer and easy to analyze at scale. Open-ended questions ask for a free-text response, slower to answer and harder to analyze in bulk, but capable of surfacing detail and nuance that a fixed option list can't anticipate. Most good surveys are mostly closed-ended with one or two open-ended questions for texture.
Use NPS (Net Promoter Score, "how likely are you to recommend this to a friend or colleague, 0–10") when you want a standardized, benchmarkable loyalty metric you can track over time and compare against industry norms. Use a plain rating question ("how satisfied were you with X, 1–5") when you want feedback on something specific and don't need cross-company benchmarking.