Definition
A leading question is one whose wording contains or suggests its expected answer. The guide to survey and questionnaire design lists it first among the wording traps, with its own example: asking how much someone enjoyed a new checkout assumes enjoyment, where a neutral version asks how they would rate it, on a balanced scale.
The lead can sit in different places. It may be an assumption built into the question, an adjective that signals the right answer, an option list that offers only the outcomes the team hopes for, or a single cause named for the respondent to agree with. In each case the question carries the asker's view into the answer.
It is a wording problem, not a sampling one. A survey can reach exactly the right people and still mislead if its questions lead them.
Why it biases results
The marketing research methods section names three biases that do most of the damage: selection bias, non-response bias and measurement bias. Leading questions belong to the third, the gap between what people say and what is true, alongside social desirability and the limits of memory. None can be removed entirely, and all should be stated in the findings.
What makes leading questions especially costly is that they are introduced by the people writing the questions, not by the people answering. The survey design guide compares a badly worded survey to a thermometer that reads high: worse than no thermometer, because people act on it. A survey with leading questions produces confident numbers pointing the wrong way, and the team that wrote it is the least likely to notice.
In practice: neutral wording in surveys, interviews and data stories
In a short on-site poll, the tutorial on on-page polls and surveys asks about what happened, not about what the visitor would do, with one open question. Its contrast is useful: "What is stopping you from choosing a plan today?" is open and neutral, while "Is our price too high?" leads. It also suggests reading the question aloud to someone outside the team before it goes live.
In an interview the lead can be spoken or unspoken. The guide to customer interviews asks for specifics about particular occasions, and warns against leading questions and against reactions that reward the answers the interviewer hoped for, as well as against explaining or defending the product mid-session.
In public-facing research the cost is reputational. The guide to digital PR campaign ideation and data stories notes that journalists increasingly check survey method, and a survey built on leading questions will be picked apart once published. It points to the codes and guidance that the Market Research Society and ESOMAR publish on question wording, and advises neutral questions from the start.
A leisure centre group rewrites a leading survey question
Snalpumb Leisure Centres changed its pool timetable across 4 sites and surveyed members a month later. The first question read: "How much do you love the new, more flexible pool timetable?", with five options from "Love it" to "Like it". Of 1,240 responses, 88% chose one of the top two. The operations manager, unconvinced, ran the survey again with a different sample of members, replacing the question with: "Which swimming sessions did you attend in the last two weeks?" followed by "Has the timetable change affected when you swim?". This time 31% said they swam less often, most of them former early-morning swimmers whose 6:30 session had moved to 7:30. The early session was restored at 2 sites the following month.
Related terms
A ScreenerThe short questionnaire that confirms each research candidate fits a study, asking about recent behaviour without revealing which answers qualify. has the mirror-image problem: if its qualifying answer is obvious, candidates who guess what the researcher wants get through, so screeners avoid leading wording too. In a Focus GroupA moderated discussion among a small set of people who share an experience, used to hear how people talk about a category together and react to material., the moderator's phrasing and reactions can lead the whole table at once, and participants already tend to follow whoever spoke last. And Sample SizeThe number of users or sessions an experiment needs to detect a given effect reliably. is a separate question entirely: a larger sample answering a leading question gives a more precise estimate of the wrong thing.
Common confusions
- A leading question is a loaded word. Loaded or unfamiliar terms are a separate trap; a question can be free of jargon and still assume its answer.
- Positive wording is just being friendly. Wording that signals the hoped-for answer changes the answer; friendliness belongs in the introduction, not the question.
- A bigger sample fixes it. More respondents make the biased result more precise, not more true.
- Only surveys can lead. Interviewers lead with follow-up questions, reactions and explanations; data stories inherit whatever bias the survey behind them contained.
- A leading question is the same as a double-barrelled one. A double-barrelled question asks two things at once; it can be neutral and still unanswerable, while a leading one asks one thing and supplies the answer.
