OPINION
Gunman Matu Reid, who died and was responsible for two other deaths in Thursday’s Auckland shooting, had been assessed as being at “low risk” of reoffending by a probation officer. How did that happen?
The first thing to consider is that hindsight is a wonderful thing – it would be easy to look back and say that, surely, someone should have predicted that he might do such a thing.
Actually, no. At least not based on the information about Reid which is in the public domain.
Nothing of what we know about his past suggested he would be a danger to the general public.
A danger to women, yes, given his history of domestic violence, but Thursday’s tragedy was off the scale for someone with his criminal profile.
I conducted hundreds of risk assessments on criminals myself during four years of working as a probation officer.
Using the information publicly available about Reid, I have used that training and experience to look at his circumstances and gauge if I would have predicted things differently.
Probably, I wouldn’t have.
To give credit where it is due, the probation officer who assessed him as having a low likelihood of reoffending got one thing absolutely right – the parallel assessment that Reid was capable of causing a high degree of harm if he ever did do something wrong.
That assessment would have been based on an explosive act of violence in 2021, for which he was convicted of strangulation, assault with intent to injure and assault on a female, after an attack on his girlfriend.
Strangulation – commonly carried out by men seeking to dominate their intimate partners – is a horrific offence.
It is a terrifying experience for the victim and, quite apart from the emotional trauma, can lead to brain injury and potentially death.
That is why the law was changed five years ago to give it a maximum penalty of seven years in prison.
Judge Stephen Bonnar sentenced Reid to five months of home detention, in line with the recommendation in a probation officer’s report.
The Sentencing Act 2002 requires judges to impose the “least restrictive” sentence appropriate in the circumstances.
“I do not want to send a young man like you, with a limited history, to prison,” Judge Bonnar told the 24-year-old Reid. “I think that could be counter-productive and actually set you down the wrong path.”
In making sentencing recommendations to the judge about Reid, the probation officer would have drawn on two types of risk assessment – one of them statistical and based on long-term factors, and the other “dynamic” and more subject to changes in the short term.
The statistical measure is what is known as a “ROC*ROI”, which means “Risk of (re)Conviction multiplied by Risk of Imprisonment”.
The ROC*ROI is an automated tool that draws on data about people’s interactions with the criminal justice system, taking into account variables such as age, number and type of convictions, and even the length of time between convictions.
It was developed in the 1990s by Corrections psychologists and statisticians from the University of Canterbury and can be likened to the actuarial assessments used by insurance companies to calculate risks when making up policies.
The ROC*ROI comes up with a number to predict the likelihood of someone reoffending within five years – so if the result is 0.8500, the chance is considered to be 85 per cent.
Over the decades it has been in use, with adjustments from time to time, there has been a high correlation between the predictive tool and actual reconviction rates.
In Reid’s case, however, the ROC*ROI would have been of limited use. His risk assessment was made after the strangulation incident but it is possible that conviction had not been entered in the calculation system yet.
Even if it had, Reid had a relatively short criminal history – it has been reported there was only one other conviction, for assault in 2020 – so there would have been little data to identify his true potential risk.
Reid would also have been assessed using a different tool – the DRAOR model, which stands for “Dynamic Risk Assessment for Offender Re-entry”.
DRAOR is a more subjective system, that considers a person’s risk factors under “stable” and “acute” headings, and assigns a score of 0, 1, or 2 for each, depending on whether they present no problem, a slight problem, or a definite problem.
“Stable” factors include a person’s peer associations, attitudes to authority, impulse control, problem-solving ability, sense of entitlement, and attachment to others.
In Reid’s violence against his partner, there are definite red flags in the seriousness of the strangulation, his reported threats to other members of her family and the statement, “You don’t know what I am capable of”.
This would have called into question his impulse control, cast questions over his attachments to other people and indicated a sense of entitlement – all things that would count against him in assessing risk.
But comparatively little is known of other matters – his peer associations, his attitude to authority (he reported to probation on Wednesday), and his problem-solving ability. More positive assessments in these areas would help to offset the red flags.
Acute factors in the DRAOR focus on what is happening in an offender’s present – things like substance abuse, accommodation, anger or hostility, access to victims, negative mood, employment, and interpersonal relations.
We know that Reid was drinking before strangling his partner, but it has been reported that he has recently passed two drug tests. He was capable of irrational anger, but may not have shown it in everyday life.
Also offsetting any negative assessments, he would have been given credit for having a job - he had been working on the construction site.
In regard to access to victims, given Reid’s profile, an assessor would be looking at women – his ex-girlfriend, any new sexual partners - not workmates on a construction site.
The DRAOR also has a list of protective factors, which work in a person’s favour – their responsiveness to advice, any pro-social identity or high expectations of themselves (again, employment often comes into play here), and social support or social control.
The fact that Reid was on home detention means that he was able to sustain a place to serve that sentence, or that someone considered “pro-social” was willing to take him in. That address would have been visited and any such people would have been interviewed and vetted.
These are all factors that are deemed to reduce a person’s risk.
Looking over the information in the public domain, I probably would not have identified any unusual risk in Reid’s case.
The violence against his ex-partner was horrible but, sadly, it was not out of place in the context of New Zealand’s appalling domestic violence landscape. There is nothing stand-out about Reid in that regard.
His ROC*ROI score would probably have been below 0.2999, meaning he would have been statistically considered “low risk”.
The dynamic DRAOR factors might have added context but probably not enough to override his limited criminal history and the actuarial scoring system when it came to making recommendations to a judge.
The management of Matu Reid on his home detention sentence is subject to a Department of Corrections internal review.
Ric Stevens is an Open Justice reporter for NZME based in Hawke’s Bay. Previously, he worked as a probation officer in Christchurch and Hastings. He never met Matu Reid. As far as he is aware, he does not know any probation officers involved in managing Reid.
Police have advice for people who come across a crash scene.