Case File 067 · Investigation and Psychology

Open-Ended Versus Leading Questions

How open-ended and leading questions differ, what memory research shows about suggestive wording, and why interview guidance puts open questions first.

By Murder Book Editorial TeamPublished 16 September 2026Updated 16 September 2026

An open-ended question invites a person to describe something in their own words, such as "Tell me what happened." A leading question suggests the answer or assumes a fact that has not been established, such as "You saw the gun, didn't you?" The difference matters because memory can be influenced by the way questions are asked. In a classic 1974 experiment, Elizabeth Loftus and John Palmer showed that changing a single verb in a question changed people's speed estimates, and even made some of them more likely to "remember" broken glass that was never there. That is why interview guidance for police in England and Wales tells interviewers to start with open-ended questions and to use leading questions only as a last resort. Open questions do not guarantee an accurate account, but they give the person's own memory the best chance to speak for itself.

Key takeaways

Open-ended questions let the person control what they report, which lowers the risk that the interviewer's assumptions shape the answer.
Leading questions imply an answer or assume a disputed fact, and research shows they can change what people later report.
UK interview guidance ranks question types: open-ended first, then specific-closed, with forced-choice and leading questions kept for a last resort.
The cognitive interview, built on memory research, increases correct details with only a small rise in errors, according to a 2010 meta-analysis.
An answer to a leading question is weaker evidence, whether or not the person is telling the truth.

What is an open-ended question?

An open-ended question is framed so the person can give an unrestricted answer. "Tell me about the evening", "What happened next?" and "Is there anything else you can remember?" are typical examples.

The Ministry of Justice guidance Achieving Best Evidence in Criminal Proceedings (updated 2023) calls the open-ended question "the best kind of question for gaining good quality information". It gives two reasons. The witness controls the flow of information, and the style "minimises the risk that the interviewer will impose their view of what happened on the witness."

The guidance also notes a common misunderstanding. Questions that start with "who", "what", "where", "when" or "why" are often described as open, but in practice they are more often specific-closed questions, because they ask for a narrow piece of information.

What is a leading question?

The same guidance defines a leading question as one that "implies the answer or assumes facts that are likely to be in dispute." Whether a question counts as leading also depends on what the witness has already said. Asking "What colour was the car?" is not leading if the witness has already mentioned a car. It is leading if they have not.

Loftus and Palmer's 1974 paper gives a similar definition: a question that, by its form or content, suggests to the witness what answer is desired.

Some leading questions are more suggestive than others. According to Achieving Best Evidence, tag questions such as "You did see the gun, didn't you?" are thought to be the most suggestive. Even small word choices matter: asking about "the" object rather than "a" object has been found to produce more erroneous answers, because "the" presupposes that the object exists.

The spectrum of question types

Interview guidance does not treat questions as simply "good" or "bad". It places them on a scale from least to most directive.

Question type	Example	How the guidance treats it
Open-ended	"Tell me everything you remember about the car."	Use first and most often
Specific-closed	"What time did you leave?"	Second choice, to fill gaps the account left
Forced-choice	"Was the car large or small?"	Avoid if possible; last resort
Multiple	"Where did he go and what was he carrying?"	Avoid
Leading	"He had a knife, didn't he?"	Last resort only, in the least-leading form possible

The guidance recommends that questioning should "whenever possible, commence with open-ended questions" and move to specific-closed questions only if needed. It also warns that opening with a run of closed questions teaches the witness to give short answers, which can carry on even when the interviewer later asks for a full account.

What did the Loftus and Palmer experiments show?

Loftus and Palmer ran two experiments with students who watched short films of traffic accidents.

In the first, 45 participants answered the question "About how fast were the cars going when they ___ each other?" The blank was filled with smashed, collided, bumped, hit or contacted. The verb changed the estimates. The mean was about 41 mph for smashed and 34 mph for hit, with contacted lowest of all. The paper also noted that people are generally poor at estimating speed: for four staged crashes filmed at known speeds, the average estimates did not track the real speeds closely.

In the second experiment, 150 participants watched a film of a multiple car accident. Fifty were asked the smashed question, fifty the hit question and fifty were not asked about speed. A week later everyone was asked, "Did you see any broken glass?" There was no broken glass in the film. Sixteen of the 50 people in the smashed group said yes, compared with 7 in the hit group and 6 in the group with no speed question.

The authors concluded that questions asked after an event can cause "a reconstruction in one's memory of that event". The question did not just colour the answer given at the time; it appeared to change what some people reported later.

Why can a leading question change a memory?

Loftus's later review of three decades of work describes the misinformation effect: memory for the past can be impaired after exposure to misleading information. The review notes that warnings about possible misinformation help only under limited circumstances, and that some people are more susceptible than others.

Achieving Best Evidence reflects this research directly. It states that responses to leading questions "tend to be determined more by the manner of questioning than by valid remembering", and that leading questions can "significantly distort the witness's memory in the direction implied".

This links closely to why memory is not a perfect recording and to the risks of repeated questioning, which can pull an account in the same direction.

What does research say about open questions in practice?

Research on child witnesses has been especially influential, because children can be particularly sensitive to suggestion. A 2003 study by Michael Lamb and colleagues analysed 130 forensic interviews with children aged 4 to 8. Across age groups, about half of the details children reported came in response to open invitations.

A 2007 review by Lamb and colleagues of research on the NICHD Investigative Interview Protocol, a structured approach that relies heavily on open-ended prompts, concluded that controlled studies have repeatedly shown that interview quality improves when interviewers use the protocol.

What is the cognitive interview?

The cognitive interview was developed by psychologists Ed Geiselman and Ron Fisher to help witnesses remember more. As summarised in Achieving Best Evidence, the original version gave witnesses four instructions:

Report everything, even details that seem trivial.
Mentally reinstate the context of the event.
Recall the event in different orders, such as backwards.
Recall from a different perspective.

The enhanced version adds techniques from the psychology of communication, such as transferring control of the conversation to the witness. The guidance notes that not every technique suits every witness; for example, changing perspective is not recommended for young children.

A 2010 meta-analysis by Amina Memon, Christian Meissner and Joanne Fraser reviewed 25 years of laboratory studies. It found "a large and significant increase in correct details and a small increase in errors", with no difference in the rate of confabulated details. The technique appeared to benefit older adult witnesses even more than younger adults.

These findings concern witness interviews in research settings, and they measure how much accurate information is recalled. They are not a method for spotting lies. For the difference between gathering information and pressing for a confession, see interview versus interrogation.

Are leading questions ever acceptable?

In a witness interview, the guidance allows them only when every other approach has failed, and even then in the least-leading form possible. It also stresses that the interviewer should never be the first to suggest that a particular offence was committed or that a particular person was responsible, unless there is absolutely no alternative.

If a leading question does produce new information, the guidance advises returning to open-ended or specific-closed questions rather than following up with more leading ones. It also cautions that an answer prompted by a leading question is likely to be of limited evidential value.

This is a point about evidence, not honesty. A truthful witness can accept a suggestion, and an answer that matches the interviewer's theory is not proof that the theory is right.

How does this connect to Murder Book?

Murder Book: The Art of Interrogation is an upcoming narrative interrogation game from Gungrounds in Zagreb. Its cases are built around people's stories: unreliable accounts, narrative gaps, avoidance and contradictions. Development screenshots show a technique menu with options including Build Trust, Bluffing, Baiting and Confrontation.

That is a useful place to remember what the research shows. How a question is phrased can shape the answer, so what a character says under pressure is something to weigh, not a verdict. A changed detail may point to a lie, a mistake or a suggestion. Murder Book presents these as fiction to interpret, not as rules that work in real interviews. You can read more in how Murder Book turns interrogation into gameplay.

Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027.