Why I ... believe universities could help transform the police service

The recent debate on the effectiveness of police patrols has been dominated by the views of two distinguished chief constables, Sir John Stevens and Sir David Phillips. But relevant scientific evidence has been conspicuous in its absence. In healthcare, a similar debate over the effectiveness of a particular treatment would be dominated by trial results. The opinions of National Health Service managers would barely register.

One chief officer recently said he could not think of an example of police policy that had been affected by social science research. This suggests a serious problem, not simply with the quality of evidence but with its credibility.

Police work is almost alone among the major public services in not being rooted in university schools fit for purpose, namely to generate and disseminate knowledge of importance to policing. A few institutions, such as Canterbury Christ Church, Teesside and Portsmouth, offer degrees in police studies but none has a police school.

Professions such as medicine and dentistry depend on reliable evidence and it is impossible to qualify as a practitioner without university skills.

There have been steps in the right direction as regards the police. The new Police Skills and Standards Organisation is promoting foundation degrees as part of training, while aspiring chief officers are seconded on to courses in applied criminology at Cambridge. University departments of social science are contributing to the understanding of crime and police practice.

But their emphasis, setting and organisation are wrong. The problem is dislocation of practice, carried out by operational police units, and research, carried out on campuses. University police schools would bring these together.

In health, the randomised field trial has become the gold standard for practice evaluation. The random assignment of a new, promising approach, or an existing practice, reduces the bias and ambiguity of research findings to a minimum. Few police interventions have been subject to this rigour. So it is impossible to say whether a project in Manchester designed to tackle city-centre disorder has led to changes in crime that are distinct from what has happened in areas where no initiative has been introduced.

There is convincing scientific evidence that targeted patrols are effective and that the more precisely police presence is targeted, the greater crime reduction. But some generic measures, such as blanket increases in the number of officers, and the New York "Compstat" system of comparing precinct statistics have not yet been proved effective.

A solid foundation of evidence would make police policy less vulnerable to populism and the next trendy theory. It would also mean harmful or ineffective interventions could be identified and discarded. Medical research has demonstrated that interventions that seem rational can sometimes do more harm than good. Blood-letting and putting babies to sleep on their fronts were once seen as good practice, but research has proved them to have potentially lethal consequences.

Research skills are different from practitioner skills. High-quality evaluation must precede implementation in practice and, in the context of policing, should draw on statistics, psychology, geography, law and economics. These disciplines can be found in universities, but are absent from police training institutions. The John Jay College of Criminal Justice at the New York City University goes part of the way to providing this expertise, but it is almost alone.

University police schools would train officers who would manage investigations, develop policy and lead command units and police forces.

The universal probationer training of constables would remain and frontline recruitment would be improved as opportunities arose for career development in and around the schools.