On the evening of 5 October 1974, bombs in two Guildford pubs killed five people. Four people — Paul Hill, Gerry Conlon, Patrick Armstrong and Carole Richardson — were convicted on the basis of confessions, and served fifteen years before their convictions were quashed in 1989. Six men convicted of the Birmingham pub bombings served sixteen years before their convictions were quashed in 1991. In both cases the confessions were false and the investigations were found to have been improper.
British interrogation practice was rebuilt as a direct consequence.
The Police and Criminal Evidence Act 1984 came first, and its most important provision was mechanical: interviews with suspects must be recorded in full. Everything Part 4 has described so far — the unrecorded hours, the shaping, the fed-back detail — becomes visible under that requirement, and much of it stops.
Then, in 1992, the Association of Chief Police Officers introduced a national interviewing model developed jointly by police officers and psychologists. It is called PEACE, and the acronym describes the sequence: Planning and preparation. Engage and explain. Account, clarification and challenge. Closure. Evaluation.
Its defining feature is what it removes. There is no confession objective. The interview is an information-gathering exercise, and the interviewer's task is to obtain a full account and then test it — not to obtain agreement.
The account phase is where the psychology sits. The interviewee is invited to give an uninterrupted free narrative, at their own pace, without interruption, before any challenge. That instruction comes directly from the memory research in the next chapter: free recall produces the most accurate reporting, and every interruption both truncates it and risks contaminating it. Challenge comes afterwards, and consists of putting inconsistencies and evidence to the interviewee, which is chapter 100's technique.
Because the interview is low-threat, the interviewee stays in a state where retrieval works. Chapter 105 covers what stress hormones do to recall; PEACE is the practical answer, and the reasoning is that an interrogator who terrifies a suspect is damaging the memory system they are trying to read.
The obvious objection is that a technique which does not aim at confession will not get confessions. Clarke and Milne's national evaluation of PEACE found that interview quality improved and that admission rates did not fall. Subsequent meta-analytic work comparing information-gathering with accusatorial approaches finds that information-gathering produces more true confessions and fewer false ones.
That comparison is the most important empirical result in Part 4 of this guide, and it is not close.
PEACE or something like it is now used in England and Wales, Norway, New Zealand, and increasingly elsewhere. Norway's version, developed after its own miscarriage of justice, is more explicitly investigative still. The United States has no national standard and the Reid technique remains dominant, though Wicklander-Zulawski, one of the largest American training firms, announced in 2017 that it would stop teaching it.
Cross-ref: cognitive interview chapter