Leading question examples
A leading question is one whose wording pushes the respondent toward a particular answer. The fix is to remove the praise word, the loaded term or the assumed premise, and to offer the full range of answers. Each example below shows the leading version, the neutral rewrite and what changed.
Four kinds of leading wording
- Praise or criticism words such as excellent, poor or hardworking, which tell the respondent what to think.
- Loaded terms where one word carries a view. Pew reports much greater public support for expanding “assistance to the poor” than for expanding “welfare”, which shows how much a single term can move the answer.
- Assumed premises, where the question takes for granted something the respondent may not believe.
- Unbalanced answers, where the options lean one way, or where the positive option is listed first and the rest are not.
Leading questions and neutral rewrites
| Leading | Neutral rewrite | What changed |
|---|---|---|
| How much did you enjoy our excellent service? | How would you rate our service? (very poor to very good) | Removed the praise word and the assumption of enjoyment |
| Don’t you agree that our new pricing is fair? | How fair is our new pricing? | Removed the request for agreement |
| How often do you use the premium features you love? | How often do you use the premium features? (never to daily) | Removed the premise that the respondent loves them |
| Would you support a reasonable plan to cut waste? | Do you support or oppose the plan to cut waste? | Removed the word reasonable, which answers the question |
| Do you find our busy, friendly staff helpful? | How helpful was our staff during your visit? | Removed the descriptive words and the yes or no frame |
These rewrites are illustrations of the principle, not findings from a study. Use them as a pattern for your own wording.
The effect is measurable
Pew Research Center’s guide to writing survey questions describes wording as critical to the meaning of a question. In a January 2003 survey, asking whether people favoured military action “to end Saddam Hussein’s rule” produced 68 percent in favour. Adding that U.S. forces might suffer thousands of casualties changed the result to 43 percent in favour and 48 percent opposed. The rest of the question was the same.
Do not measure your own assumption
The split test
When you cannot decide between two wordings, Pew describes writing two versions and asking each half of the sample one of them. If the answers differ significantly, the wording is the cause. You can run the same test on a small scale: send one version to half of a pilot group and the other version to the rest, and compare.
A worked example: a three-question block
Here is a hypothetical block from a feedback survey, with the wording a reviewer would flag first:
- Our friendly team clearly cares, so how well did we meet your needs?
- Wouldn’t you agree the new dashboard is an improvement?
- Which of our many useful features do you rely on most?
The first question praises the team before asking for a rating, so a respondent who wants to be agreeable has a reason to rate highly. The second asks for agreement, which makes disagreeing feel rude. The third assumes the features are useful and that the respondent uses them. The neutral versions ask the same things with the praise, the request for agreement and the assumption removed: how well did we meet your needs, how would you describe the new dashboard compared with the old one, and which features, if any, do you use?
The rewrite does not make the survey friendlier by making it vaguer. It makes each answer mean the same thing to every reader, and that is what allows results to be compared with the next survey.
Check before you send
- No question contains a praise word or a criticism word.
- No question asks the respondent to agree with your view.
- No question assumes the respondent has used a feature or believes a claim.
- Every rating has labels at both ends and a neutral midpoint where one is needed.
Build it in Seagit Forms
Seagit Forms does not split a survey into randomised versions, so run the test as two separate forms with the same questions and the two wordings, send each to half of your group, then compare the answers yourself. Start from the market research survey template, and write each question as a neutral statement with the full range of answer options. Check the fields documentation to pick a field type that gives every option its own label.
Sources
- Pew Research Center, “Writing Survey Questions” (wording effects, the 2003 Iraq question, the 2005 terminal illness question, the welfare and assistance comparison, two forms of a questionnaire).
Common questions
- What is a leading question in a survey?
- A leading question is worded so that it pushes the respondent toward one answer. The push can come from a praise word, a loaded term, a premise the respondent has not agreed to, or an answer that is easier to choose than the others.
- Does a small change in wording really change the answer?
- Yes. Pew Research Center reports that a question about making it legal for doctors to give terminally ill patients the means to end their lives drew 51 percent support, while a question using the phrase assist in committing suicide drew 44 percent. Both ask about the same thing.
- How do I check my questions for leading wording?
- Read each question twice: once as a supporter of the position and once as an opponent. If the supporter would find it easier to answer, rewrite it with neutral words. Then pilot the survey with a few people who did not write it.