Research suggests just over half of all suspects interviewed inside a police station will either confess or make damaging admissions. Surprisingly, police interviewing tactics appear to have only a limited effect on the decision making of suspects. Instead, the strongest predictor of a confession is the strength of evidence against the accused. This paper examines the links between strength of evidence and the outcome of interviews. Data is drawn from previously published studies of police interviewing practices in the USA, UK, and Australia. This review shows how when evidence is strong prior to questioning, such as a suspect having being found in possession of stolen goods, then confessions are the most likely outcome. However, some apparently strong forms of evidence, such as testimony from police officers, can be used inappropriately making confessions less likely. The findings are discussed in relation to training in investigative interviewing skills and investigative practices. The paper also identifies a need for research into the conditions which give rise to true confessions, as opposed to a contemporary focus on the conditions that can result in false confessions.