National leader Simon Bridges has provided the basis for claims the threat of prison stops crime - an unpublished study that has not been peer-reviewed that suggests the worst offenders aren't put off by the prospect of prison.
The paper studied a mass pardon in Italy and found reoffending rates showed certain inmates were put off further offending by the prospect of prison.
The research paper was provided by Bridges after the NZ Herald asked for the evidence to support his claim the Three Strikes legislation deterred people from committing further crime.
This month has seen two judges question the value of longer sentencing.
Crime & punishment
• Judge: Make New Zealand safer with shorter prison sentences
• High Court judge's challenge: Show me the evidence long sentences work
• Justice path and bulging prisons - will NZ listen to scientist or sceptic?
• Andrew Little: 'Longer sentences, more prisoners - it doesn't work'
Court of Appeal president Justice Stephen Kos said greater flexibility in sentencing, shorter sentences and open prisons could be the answer to a safer New Zealand.
Before sentencing two drug couriers, High Court judge Matthew Palmer asked prosecution and defence lawyers for evidence that longer sentences actually put people off further crime.
The Crown didn't produce any evidence they did - and the defence produced evidence they did not.
The study - Criminal Discount Factors and Deterrence - said "imprisonment does have the potential to deter crime".
But it showed any deterrent effect was not in areas likely to impact on New Zealand's recurring offender or prison population.
It found there was a strong potential deterrent effect on prisoners who were educated and older.
Of those in New Zealand prisons, 71 per cent of people do not have literacy levels sufficient to cope with daily tasks and 65 per cent were aged under 40.
It also found there was a minimal deterrent effect for those imprisoned on drug or violence charges - such inmates make up half of New Zealand's prison population.
It also said it might not be the best choice because money put into prison could be better used for "policies aimed at increasing the certainty of punishment either through a higher probability of apprehension or through improved efficiency of the criminal justice system".
The paper was produced for an internal thinktank as a "working paper" and came with the caveat such studies "often represent preliminary work".
Victoria University criminologist Dr Liam Martin said it was surprising Bridges had supplied a paper which had not been peer-review or published.
He said there had been peer-reviewed published academic research on the mass Italian pardon of 22,000 prisoners.
The pardons were to relieve pressure on prisons and saw inmates with fewer than three years to serve freed on condition they did not offend again. If they did, the old sentence would be added to any new sentence.
Martin said the data from the Italian pardons appeared to show those who had spent less time in prison were put off returning while those who served lengthy prison sentences were less troubled.
He said it lined up with research showing those who spent long periods in prison became used to it.
Martin said: "The literature does suggest some kind of deterrence effect but that works against the fact that sending people to prison makes them more likely to reoffend."
He said overall research showed prisons did not work as the negative aspects outweighed any positive benefit to the community.
Bridges initially offered Three Strikes statistics as evidence underpinning his belief in deterrence, despite the Ministry of Justice repeatedly saying the data should not be used that way.
When this was pointed out, the National Party leader's office provided a statement and a link to the Italian study.
In the statement, Bridges said: "This paper from 2016 shows that harsher sentences work as a deterrent, but deterrence is only one factor in support of tough prison sentences."
He said longer prison sentences were also used to punish people, provide drug and alcohol dependency programmes and to keep criminals from committing crime against the public.
Bridges said: "That's why we believe in Three Strikes. We believe in giving people second chances, but the really bad guys on their third strikes have proven they can't be trusted in society."
Bridges said most of those in prison were locked up for offences punishable by two years in prison or more.
"These include murder, manslaughter, rape, aggravated assault and sexual violence," he said.
Other crimes in this category - not mentioned by Bridges - include false accounting, forging documents and computer hacking.
Bridges said in his statement: "National believes that our worst offenders should be punished and the community kept safe".
Minister of Justice Andrew Little has also said the worst offenders should not be released.
In February, Little offered assurances to the mother of a murdered teenager that criminal justice reforms would not see her son's killer - or other dangerous criminals - released early.
"The safety of the community is paramount. That's why there are going to be people in prison still."
Little's intended reform of the criminal justice system has brought a focus on whether evidence or public emotion should lead criminal justice policy.
The Office of the Prime Minister's Chief Science Adviser issued a study in March in which it said New Zealand's prison population had been driven up by "shifting policies and media panics".
The former adviser, Sir Peter Gluckman, said it was time for "data over dogma" and to use "evidence-based approaches to prevention, intervention, imprisonment and rehabilitation".
"This does not diminish the importance of incarceration for a subset of individuals so as to protect the public."
Statistics show New Zealand has one of the lowest homicide rates in the OECD even though people feel more at risk here than most countries. While having a lower crime rate, 65 per cent of New Zealanders feel safe walking alone at night against an OECD average across 38 countries of 69 per cent.