Leading questions can manufacture the expected answer
Aliases: leading question · suggestive wording · question wording effect
What it is
Leading question bias occurs when a question presupposes a fact, signals a desirable stance, restricts causes, or makes one answer easier to produce. “Is the new flow clearer?” assumes improvement and reveals a positive frame; “Where is it unclear?” presumes a problem. Leading can also enter through paraphrase, tone, facial response, options, and repeated probing.
Why it happens
Answers are constructed within the frame supplied by a question rather than retrieved intact. Participants treat wording as information about what the researcher knows and socially expects, search for matching experiences, neglect accounts outside the frame, and avoid appearing uncooperative. Vague memories and weak attitudes are especially susceptible. Prompted evidence can be fluent and consistent precisely because all participants received the same cue.
Studying it
Cognitive pretests ask how respondents interpreted a question and which assumptions they noticed. Critical items can randomize neutral and directional wording and compare distributions, themes, and nonresponse. Interviews begin with concrete episodes and open accounts, then use symmetric probes for supporting and contrary evidence. Code whether a concept first appeared in participant speech or in the interviewer’s prompt so elicited content is not reported as spontaneous.
Where it stops holding
Every question frames a response; perfectly influence-free language does not exist. Clarification, safety-event confirmation, and standardized closed items may be necessary when their frames match the construct and treat alternatives symmetrically. Agreement with a leading statement is not proof of falsehood, but it cannot separate an existing view from the wording effect.
Applying it
- Remove evaluative language and unverified premises; ask what happened before naming an attribute.
- Probe positive and negative directions symmetrically and permit no-change, uncertain, and out-of-frame answers.
- Ask a colleague to predict the easiest answer to each item; rewrite questions that make the contrary answer awkward.
- Sample recordings for supplied words, completed sentences, and evaluative reactions, reporting highly prompted material separately.