Leading Questions: Definition, Types and 30 Examples

Good user research depends on asking questions that let respondents describe what they actually think, remember, or experienced.

A leading question pushes someone toward a particular answer through its wording or framing. But biased survey results can also come from response options, question order, interviewer behavior, social desirability, and the wider context of the questionnaire.

That distinction matters. Not every biased survey question is technically a leading question, but the result can be similar: data that reflects the survey design as much as the respondent’s actual opinion.

Decades of survey-methodology research show that wording, response format, question order, and context can all change the answers people give.

This article breaks down where those biases come from, how to spot them, and how to design questions that reduce them.

TL;DR

• A leading question steers respondents toward a preferred answer through wording or framing.
• Loaded and double-barreled questions are related problems, but they are not the same thing.
• Question order, response options, interviewer behavior, and social desirability can bias answers even when the wording looks neutral.
• Open-ended and closed-ended questions solve different research problems; neither is inherently better.
• Randomize unordered response options when order could influence selection, but keep naturally ordered scales in their logical sequence.
• Pretest important questionnaires before relying on the results.

How can a single phrase change survey responses?

Small wording changes can change what respondents think a question means, which part of the topic they focus on, and which answer feels most appropriate.

Early in Norman M. Bradburn’s classic, Asking Questions: The Definitive Guide to Questionnaire Design—for Marketing Research, Political Polls, and Social and Health Questionnaires, a core source for this post, the author illustrates how a subtle shift in language affects responses:

Two priests, a Dominican and a Jesuit, are discussing whether it is a sin to smoke and pray at the same time. After failing to reach a conclusion, each goes off to consult his respective superior. The next week they meet again.

The Dominican says, “Well, what did your superior say?”

The Jesuit responds, “He said it was all right.”

“That’s funny,” the Dominican replies. “My superior said it was a sin.”

The Jesuit says, “What did you ask him?”

The Dominican replies, “I asked him if it was all right to smoke while praying.”

“Oh,” says the Jesuit. “I asked my superior if it was all right to pray while smoking.”

The figures below are historical survey results. Their value here is not as current public-opinion data, but as examples of how wording can produce materially different answers.

Example 1

• in cases of incurable disease, doctors should be allowed to “assist the patient to commit suicide”: 51% agree
• in cases of incurable disease, doctors should be allowed to “end the patient’s life by some painless means”: 70% agree

Example 2

• “having a baby outside of marriage” is morally wrong: 36% agree
• “an unmarried woman having a baby” is morally wrong: 26% agree

Example 3

• “Do you think the United States should allow public speeches against democracy?” 21% agree
• “Do you think the United States should forbid public speeches against democracy?” 39% agree

The point is not that every question has one perfectly neutral wording. It is that every wording choice carries assumptions and cues.

That matters most when attitudes are weakly held, the topic is unfamiliar, or the respondent is trying to infer what the researcher means.

Research on questionnaire design has repeatedly found that wording, question form, and placement can change response distributions.

What is a leading question?

A leading question is worded or framed in a way that steers a respondent toward a particular answer.

For example:

• Leading: “How much easier is the new dashboard to use?”
• Neutral: “How would you compare the usability of the new dashboard with the previous version?”

The first assumes an improvement. The second allows the respondent to report that the experience became easier, harder, or stayed the same.

The same issue appears in B2B customer research:

• Leading: “How much time does our product save your team?”
• Neutral: “How, if at all, has the product affected the time your team spends on this task?”

A test is simple: Does the question contain an assumption or preferred direction that the respondent has to work against?

Types of leading questions (with examples)

When you think of leading questions, you likely think first of the language—the words and phrasing—of those questions. But the question type, topic, and order can be equally influential.

Question language

Which of the following is a leading question?

• “Does your employer or his representative resort to trickery in order to defraud you of your part of your earnings?”
• “With regard to earnings, does your employer treat you fairly or unfairly?”

The former, a blatantly leading question, was how Karl Marx framed it in early surveys of workers. The latter, defanged of words like “trickery” and “defraud,” offers a more neutral question.

Intentionally leading questions such as Marx’s are unlikely to plague your survey. But subtle choices can be influential. For example:

• Is the new design easier to use than the old one? The use of “new” and “old” cues respondent expectations, which are also primed to consider whether the changes make the website “easier” to use.
• Was one design easier or harder to use than another? This phrasing eliminates the bias introduced by old vs. new and gives equal weight to a positive or negative experience.

Additionally, some words, though seemingly interchangeable, have connotations that skew results. For example, asking respondents about “welfare”—a politically charged topic—yields far different levels of support compared to “assistance for the poor”:

In the context of web design, it’s easy to think of similar examples: calling content an “ad” instead of “sponsored” or identifying an element as a “pop-up” rather than a “lightbox” may shape responses.

Similar to leading questions, two other types of questions can also bias response data:

Double-barrelled questions

Words such as “and” and “or” can be warning signs for double-barreled questions, but they do not automatically make a question double-barreled. The real problem is asking respondents to evaluate two different things while giving them only one answer.

• “Are you satisfied with the pay and benefits at your office?”
• “How would you describe your experience trying to find blog or webinar content?”

While double-barreled questions are technically distinct from leading questions, they create a similar measurement problem: one answer may not accurately represent the respondent’s view of both things being asked.

Loaded questions

Unlike leading questions, which suggest the desired answer, loaded questions assume one:

• “Was it easier to navigate the new design?” (leading)
• “Which of the design improvements was your favorite?” (loaded)

A loaded question contains an assumption the respondent may not accept. In the example above, the question assumes the changes were improvements and that the respondent had a favorite.

The choice of language becomes especially critical when questions tackle sensitive topics.

Question topic

Although respondents are motivated to be ‘good respondents’ and to provide the information that is asked for,” writes Bradburn, “they are also motivated to be ‘good people.’

In surveys, social desirability bias can skew responses even when the question itself is neutrally worded, and, as a result, skew results. The effect can vary substantially depending on the topic, survey mode, and how sensitive the behavior feels to the respondent.

Take a simple question: “On average, how much time do you spend on social media each day?” Social desirability bias may lead to underreporting—devoting whole evenings to scrolling through Facebook’s News Feed isn’t something to brag about.

To get more accurate responses, the language of the question needs to offer “outs”—ways to mitigate the impact of social desirability bias. The best way to do so depends on whether the concern is underreported or overreported behavior, or for knowledge-based questions.

Underreported behavior

Researchers sometimes use face-saving wording to make socially undesirable behavior easier to report. That can improve disclosure, but it needs to be designed carefully because the wording itself can introduce a benchmark, anchor, or expectation.

For example:

“During the past seven days, about how much time per day did you spend using social media?”

is usually safer than:

“The average person spends more than two hours per day on social media. How much time do you spend?”

The second version may reduce embarrassment, but it also supplies a numerical reference point before the respondent answers.

Face-saving techniques can be effective, but they are not universally reliable. A recent systematic review found they performed well overall, while effectiveness still varied by topic and implementation.

Overreported behavior

“Do you jog?” “How many books did you read this year?” Neither question is of great consequence, but both risk overreporting due to social desirability bias.

Countering that bias can be as simple as adding a short phrase to normalize a negative response:

• “Do you happen to jog, or not?”
• “How many books, if any, did you read this year?”

Similar phrases are useful for other question types. Consider the difference between these two:

• “What do you like about…?”
• “What, if anything, do you like about…?”

The latter, Estée Lauder learned, led respondents to choose “nothing” more often.

For similar questions, Bradburn offers another solution—provide reasons why someone may not perform the behavior:

Overreporting can be especially visible when a socially desirable behavior is relatively uncommon. As a classic study demonstrated, a question as basic as “Do you own a library card?” vastly inflated reporting.

Owning a library card was socially desirable but, in the location of the study, a minority of the population had one. The broader point is that respondents may be more likely to overreport behavior they believe they should perform, particularly when the behavior is socially desirable.

Knowledge-based questions

No one wants to come off as an idiot. That desire can lead to overreporting for knowledge-based questions, such as brand-awareness surveys. Respondents believe that they should know an answer, so they’re more likely to check yes.

Neutralize knowledge-based questions with phrases that suggest the knowledge is not expected:

• “Do you happen to know…”
• “As far as you know…”
• “Can you recall offhand…”

If you are measuring knowledge, do not turn the question into an opinion question simply because opinion is easier to answer. That changes what you are measuring.

Where lack of knowledge is a meaningful result, make it acceptable for respondents to say they do not know rather than pressuring them to guess.

The breast-cancer example below is still useful, but it should be understood as an example of changing the measurement from knowledge to belief, not simply as a better version of the same question.

Question type

Agree/Disagree, Yes/No, and True/False questions can be affected by acquiescence bias: the tendency for some respondents to endorse the assertion presented in the question regardless of its content.

That means a question can measure both the respondent’s actual opinion and their tendency to agree with statements.

• “Ads are the best way for news websites to earn money.”

How would you respond if that same prompt were changed to a forced-choice format?

• “Ads are the best way for news websites to earn money.”

OR

• “Paywalls are the best way for news websites to earn money.”

Acquiescence bias aside, the initial statement likely triggers negative feelings. (Who likes ads?) The second option that includes the primary alternative—one that requires you to open your wallet—may cause respondents to reconsider.

A Pew Research Center poll highlights the potential divide:

Presenting competing positions explicitly can reduce the tendency to simply endorse the first assertion, although the alternatives still need to represent genuine choices rather than a false dichotomy.

Question order

Question order can influence later answers by changing which information is most salient when respondents interpret the next question. These are known as context or order effects.

A preceding question can change the frame respondents use for the question that follows, even when neither question is leading on its own.

General vs. specific questions

“When a general question and a more specific-related question are asked together,” explains Bradburn, “the general question is affected by its position, whereas the more specific question is not.” The specific question, if it comes first, may lead respondents to their answer for the general one.

For example, take two questions on marketing knowledge, ordered from general to specific:

• “How would you rate your marketing team’s overall knowledge?”
• “How would you rate your marketing team’s knowledge of multivariate testing?”

Putting the general question first reduces the chance that the specific question will redefine what respondents consider when answering the broader one.

1. A lack of knowledge of multivariate testing may make responders more pessimistic about their teams’ overall knowledge.
2. The general question may be misinterpreted as referring to everything except knowledge of multivariate testing.

A secondary benefit of asking the more general question first is that it makes responses comparable to other surveys (assuming the other surveys asked the more general question first as well).

Underreported behaviors

Question order can also change the context in which respondents evaluate sensitive behavior.

Bradburn shares an example from a survey seeking to learn about shoplifting. The order of questions starts with more serious criminal behavior to make the real target of the survey, shoplifting, appear less deviant:

In this example, asking about more serious behavior first makes shoplifting appear less exceptional by comparison. That may reduce social-desirability pressure, but it is still a deliberate context effect rather than a neutral change in wording.

In some instances, the only way to eliminate a leading question is to “make the biased choice that is implicit in the question wording explicit in a list of choices that include alternatives.”

How can answer options create leading questions?

Answer options can bias a survey even when the question itself is neutral. The choices you provide, how they are worded, and the order in which they appear can all influence what respondents select.

To some extent, the distinction between question formulation and techniques for recording answers is an artificial one, because the form of the question often dictates the most appropriate technique for recording the answer—that is, some questions take on their meaning by their response categories. – Norman M. Bradburn

A well-worded question can still produce poor data if the available responses omit realistic answers, overlap, imply what is normal, or make some options easier to select than others.

The biggest structural decision is whether respondents answer in their own words or choose from a predefined set of responses.

When should you use open-ended survey questions?

Use open-ended questions when you need respondents’ own language, explanations, motivations, objections, or answers you may not have anticipated.

Open-ended questions let respondents formulate an answer rather than selecting from categories chosen by the researcher.

That makes them particularly useful for exploratory customer research and Voice of Customer work. They can reveal terminology, concerns, use cases, or motivations that would never have appeared in a predefined list.

But open-ended does not automatically mean unbiased.

Respondents have to recall an answer, decide what is relevant, articulate it, and type or say it. That creates more cognitive effort than selecting from a list, and some answers may simply fail to come to mind.

Research comparing open- and closed-ended formats consistently finds that they can produce different response distributions because the two formats ask respondents to perform different cognitive tasks. A current survey-research reference guide notes that closed options can remind people of answers they would not otherwise volunteer.

A historical split-ballot experiment demonstrates the effect. When respondents were asked about the most important issue in choosing a president, explicitly presenting the economy as an option made it much more likely to be selected than when respondents had to recall an issue unaided:

The experiment does not tell us that one format produced the “correct” answer. It shows that recognition and recall are different tasks.

If you ask respondents to recall an issue unaided, you learn what is most readily accessible to them. If you show them a list, you learn which of the supplied options they recognize as relevant.

If you do not yet know the likely response categories, start with qualitative or open-ended research. Use those responses to create the closed-ended options, then test whether the list captures the range of answers people actually give.

Open-ended questions are also useful when you want to avoid biasing respondents by normalizing a range. Respondents, Bradburn notes, tend to avoid extreme answers, so they’re less likely to choose the top-end of a range for undesirable behaviors or the bottom end for desirable behaviors.

The open-ended strategy may work well when respondents aren’t sure of the “normal” range, like the frequency with which people eat beef for dinner:

Open-ended responses also take more effort to analyze. Answers need to be coded or categorized, and that process can introduce another layer of interpretation.

They can also increase respondent burden. Recent experimental research on web surveys found that adding open-ended probes increased survey break-off and backtracking, illustrating the trade-off between richer qualitative information and respondent effort.

When should you use closed-ended survey questions?

Use closed-ended questions when you already know the meaningful response categories and need answers that can be compared, segmented, or quantified consistently.

Closed-ended questions reduce the effort required to answer and make quantitative analysis easier. But the response categories become part of the measurement.

Check four things:

• Coverage: Can every reasonable respondent find an answer that fits?
• Exclusivity: Can one answer accidentally fit more than one category?
• Neutrality: Do the choices suggest what counts as normal or desirable?
• Granularity: Can respondents genuinely distinguish between the levels you provide?

Research on questionnaire design shows that incomplete response sets can push respondents toward options that do not accurately represent their views, while mismatches between a question and its response options can affect both response distributions and data quality.

• A Yes/No question works when the underlying state is genuinely binary. It is less useful when you care about degree, frequency, likelihood, or intensity.
• Numerical ranges can imply what counts as low, typical, or extreme, so choose the boundaries deliberately. Include a midpoint when neutrality is a meaningful response.
• Removing the midpoint from a forced-choice scale does not eliminate neutral opinions; it simply prevents respondents from expressing them directly.

Consider the implicit judgment in these two potential response ranges:

These two ranges demonstrate why the boundaries matter. The same amount of viewing can appear relatively moderate in one scale and extreme in another.

That does not mean you should manipulate the range to make a behavior easier to admit. It means the range itself is part of the question and needs to reflect meaningful, defensible categories.

How should you handle “don’t know” answers and guessing?

Make lack of knowledge a legitimate response when knowledge is what you are measuring, and use quality-control techniques to detect respondents who claim familiarity with things they cannot actually know.

One way to detect false recognition is to include a fictitious brand, concept, or answer that respondents cannot genuinely know.

If respondents claim to recognize it, that shows that at least some of your awareness responses contain guessing or overclaiming.

If 20% claim to recognize a fictitious brand, that does not mean you can automatically subtract 20 percentage points from every real brand. The false-recognition rate is evidence of measurement noise, not a universal correction factor.

When lack of knowledge matters, make “I don’t know” or an equivalent response available rather than forcing respondents to guess.

Be deliberate about it, though: explicitly offering a “don’t know” option can itself increase the number of people selecting it. The important point is to decide whether ignorance is substantively meaningful for the question you are asking.

Does the order of response options affect survey answers?

Yes. Response-order effects differ depending on whether respondents read the options or hear them.

In visual surveys, including most online questionnaires, respondents can be more likely to choose options near the beginning of a list. This is known as a primacy effect.

In oral surveys, such as telephone interviews, options heard later can be easier to remember and therefore more likely to be chosen. This is a recency effect.

The difference makes sense: someone reading a list can scan the options, while someone listening to a list has to retain them in working memory.

Experimental survey research has repeatedly documented this mode-dependent pattern.

For unordered response options, randomize or rotate the list when position itself has no meaning.

Do not randomize naturally ordered categories such as:

age or income ranges.

Strongly disagree > Strongly agree;

Never > Always;

Low > High;

Bradburn also describes deliberately ordering responses to sensitive questions so that less socially desirable answers appear first. The intention is to make those responses easier to consider rather than letting respondents immediately settle on the socially desirable option.

That is a deliberate context intervention rather than a neutral rule, so use it cautiously and test whether it improves reporting for the behavior you are measuring.

The larger context of the survey plays a role, too.

How can the survey context influence respondents?

Respondents react not only to the questions themselves but also to who is asking, why they think the research is being conducted, and how the interviewer behaves.

Additionally, respondents’ desire to be good people and respondents may bias them toward telling you what you want to hear. If the UX designer is also conducting interviews about the design—and interviewees know it—they may feel compelled to offer positive feedback.

Interviewers can introduce additional cues through tone of voice, facial expressions, follow-up questions, or visible approval and disapproval.

If you need standardized quantitative data, keep administration as consistent as possible and avoid signaling which answer you expect.

Qualitative interviews work differently from standardized surveys. Follow-up, clarification, and researcher interaction are often part of the method. In that context, Irving Seidman recommends:

1. Making your perspective known upfront.
2. Asking the respondent what they think about that assertion.

That should not be treated as a rule for standardized surveys. If the objective is comparable quantitative responses, introducing the researcher’s own position would itself change the measurement.

Pretest important surveys before launch. Ask people similar to your intended respondents what they think each question means, pilot the complete questionnaire, and test alternative wording when two versions could reasonably produce different interpretations.

How do you avoid leading questions in research?

Remove unnecessary assumptions from the wording, ask one concept at a time, provide realistic response options, control for order and context effects, and pretest the questionnaire before trusting the results.

Leading questions are only one way a survey can manufacture the answer it appears to measure.

A neutral-looking question can still produce biased data if its response options are incomplete, if earlier questions change its context, if respondents feel pressure to present themselves positively, or if the interviewer signals which answer they expect.

Before publishing a questionnaire:

• ask one thing at a time;
• remove assumptions about what happened or what respondents should think;
• make undesirable, neutral, “none,” and “don’t know” responses legitimate where appropriate;
• make sure closed-ended options cover the realistic possibilities without overlapping;
• separate knowledge from belief;
• consider how earlier questions may change later responses;
• randomize unordered response options when position could matter;
• pretest the questionnaire before treating its output as reliable evidence.

You cannot remove every influence from a survey. The objective is to reduce the influences that systematically push respondents toward a result you wanted to find.

Frequently asked questions about leading questions

What is a leading question?

A leading question is worded or framed in a way that steers respondents toward a particular answer.

What is an example of a leading question?

“How much easier is the new product to use?” is leading because it assumes the product became easier to use.

What is the difference between a leading and loaded question?

A leading question suggests a preferred answer; a loaded question contains an assumption the respondent may not accept.

What is a double-barreled question?

A double-barreled question asks respondents to evaluate two things while allowing only one answer.

Can response options make a question leading?

Yes. Missing, overlapping, ordered, or suggestively worded options can bias responses even when the question itself is neutral.

Are open-ended questions less biased?

Not necessarily. They avoid predefined response categories but require more recall, articulation, and interpretation from respondents.

Does the order of answers matter?

Yes. Visual lists can produce primacy effects, while orally presented lists can produce recency effects.

How do you test a survey for leading questions?

Use cognitive interviews, pilot testing, expert review, and randomized wording tests where necessary.

How can you improve your customer research?

Avoiding leading questions is only one part of collecting useful customer evidence. Participant selection, research method, survey design, and how you interpret what people say all affect the quality of the result.

Learn how to run stronger user research: CXL’s User Research course covers research design, interviews, recruiting, and interpreting user evidence.

Turn customer feedback into marketing insight: The Voice of Customer Data course covers surveys, motivations, perceptions, customer language, and open-ended responses.

Use research to generate better experiments: The Strategic Marketing Experimentation course covers customer surveys, on-site polls, usability research, and turning research into test hypotheses.