It is tempting to think that the only way to stop individuals committing crime is to get those who commit it to serve a sentence and/or make amends to the victim and/or undertake a rehabilitative programme. But that’s not what the evidence says. In many cases of low level or first time crime, being caught by the police will make a profound difference, regardless of the sanction applied.

The evidence that punitive conditions are sometimes unnecessary or even counterproductive is around, though little talked about. Recently the government has proposed changing the use of cautions. The simple caution (previously called a warning) has been around for many years as the lowest level of criminal sanction. It involves a formal version of a rap over the knuckles. The person who committed the crime has to admit their guilt and the crime is recorded on the police national computer. But the person doesn’t have to do anything after their knuckles are rapped. And they are not prosecuted.

The government is proposing to abolish the simple caution and replace it nationwide with the conditional caution. A conditional caution (which has been used by some police forces for a while) has to have conditions; examples cited by government are of proper rehabilitative programmes for drug addiction or to address offending behaviour. But not all conditions are rehabilitative – a condition could also be to pay compensation or abide by a curfew.

In 2014 the conditional caution was piloted and its results compared with the simple caution. Despite those on the conditional caution having a more onerous punishment, the drop in reoffending was exactly the same between the two sanctions. So less was the same. And both were very successful in reducing reoffending compared to court sanctions – 0nly 5% reoffended within three months of receiving the caution/conditional caution. And there was no statistically significant difference between those who complied with the conditions imposed on them and those who didn’t.

Another example of the “less is the same” rule applies to a well intentioned initiative by Thames Valley police. They wanted to offer better rehabilitative opportunities to those who admitted domestic abuse but whose actions did not meet the prosecution threshold. So they commissioned a 10 week course specifically designed to reduce someone’s propensity to commit abuse. The force evaluated its effect on reoffending compared to no course. Early comparative results were not great, so they gathered more data through putting more people on the course. But Thames Valley Police have recently pulled the plug after realising that the 10 week course did no more good than no course. The good news is that the people who had admitted abuse mostly turned a corner – the average reoffending rate for all fell considerably. The (police recorded) reoffending of those who engaged with the programme reduced 47%, while the reoffending of those who didn’t engage reduced 67% in the twelve months following risk assessment. And those who were referred to the programme but failed to complete it, did as well as those who did complete.

I take my hat off to Thames Valley for trying an approach which looked as though it would transform diversion for domestic abuse cases, for properly evaluating it and then letting go when it was clear it didn’t work any better than a metaphorical rap over the knuckles. But, as Dr Peter Neyroud points out, “one thing we need to test is whether the victim feels that a programme which is intended to encourage desistance is seen as a credible attempt at prevention and thus gives greater legitimacy to diversion”.

And for an example where more formal CJS sanctions are counter-productive? Seminal research led by Professors Lesley McAra and Susan McVie suggests that if a child who commits an offence is swept into the formal criminal justice system, they are more likely to reoffend than their peer who is dealt with informally.

Its hard for practitioners – police, YOT workers, probation officers – to absorb that fewer conditions may work just as well as more. That rehabilitative programmes don’t always improve things. But it makes sense. People unused to being apprehended by the police haven’t got hardened to the criminal justice system. For many, just being caught causes huge shame and guilt, and that prompts them to desist. They don’t need to do a programme or to pay a fine – the police are the hand on the tiller guiding them to safer shores.

I’d love someone to do ethnographic research on the first conversation the police has with someone who fesses up to a minor crime. Whether they mete out a caution or not, I’m sure that the relationship the officer creates and the words he uses can change history. Because in the end people support people to stop committing crime. Programmes and interventions can help but, for people who have made a stupid mistake or who have uncharacteristically lost their temper, one conversation with one police officer can make all the difference.