Leading question
A question worded to nudge the respondent toward a particular answer.
What it means
A leading question is one phrased, framed, or presupposed in a way that suggests or pressures a particular answer, undermining the neutrality required for valid measurement or testimony. The bias can be lexical, as when a single loaded word shifts responses, or structural, as when a question smuggles in an assumption the respondent has not endorsed. In survey research, leading wording inflates agreement and distorts estimates, which is why item neutrality is a core principle of questionnaire design. The deeper danger appears in memory: because memory is reconstructive, leading questions can implant or alter what people believe they witnessed, a finding with serious implications for eyewitness testimony and forensic interviewing. Loftus and Palmer famously showed that merely changing the verb describing a filmed car crash changed both estimated speed and later false memories of broken glass. The remedy is open-ended, balanced, presupposition-free phrasing and trained, non-suggestive interviewing. It matters because how a question is asked can manufacture the very answer it appears merely to elicit, corrupting both data and memory.
Examples
Witnesses asked how fast cars were going when they 'smashed' into each other gave higher speed estimates—and more often falsely recalled broken glass—than those asked when the cars 'hit.'
Asking customers 'How much did you enjoy the new app?' collects warmer scores than 'How was the new app?' — the word enjoy plants the answer before anyone thinks.
A manager who asks 'You're happy with the new rota, aren't you?' in front of the team will hear yes, and learn nothing about whether the rota works.
First described in Loftus & Palmer (1974) for the memory effect.