Can early computer based screening predict tomorrows offenders

Can the likelihood of a child growing up to become an offender be established by computer software packages? Anabel Unity Sale reports

There was one international story in August that at first glance seemed to be a classic silly-season report: a New Zealand government taskforce charged with reducing prisoner numbers was said to be considering introducing a screening programme for pregnant women to identify unborn children at high risk of becoming criminals.

Surely no government could seriously want to label foetuses as potential young offenders? But it turns out the proposal is a serious attempt to help reduce youth offending in the country by targeting preventive services at children early on. The thinking is that enough is known about the traits of people who go on to commit serious crime to identify them before they are born. New Zealand researcher Kaye McLaren, author of Tough is Not Enough: Getting Smart About Youth Offending, says: “A lot of things that put people at risk of offending are characteristics of their family and parents.” She suggests new mothers be surveyed in hospital, and children identified as at risk should go on a database where they would be singled out for extra intervention from services to try and help them before they get into trouble.

The plan to feed information into a database which would highlight young people likely to commit crime begs the question of whether technology can be used as an accurate measure for predicting behaviour.

Use of such technology is already well advanced in this country. In England and Wales the Youth Justice Board uses two computerised assessment tools to record and analyse information about children and young people and their offending behaviour. The first is called Asset and was introduced by the YJB in April 2000 to youth offending teams (YOTs) to assess their work with convicted young offenders so that those who may reoffend may be targeted. The second, Onset, is a referral and assessment framework first used in 2003 by the 13 youth inclusion and support pilot panels dealing with young people who may engage in offending behaviour in the future (see Knowing the Score).

A final report is due next year on the pilots’ findings. All YJB-funded prevention programmes and youth inclusion programmes now use the Onset predictive approach. Onset and Asset allow youth justice practitioners to electronically record, store and share the information they gather after assessing a child or young person.

The two software systems were created by the Centre for Criminology at Oxford University on behalf of the YJB. Kerry Baker, the centre’s research officer, says Asset prompts practitioners to consider a range of practice issues when looking at a young person’s offending behaviour and the likelihood of them reoffending.

She says: “YOTs are multi-disciplinary teams and initially there was a concern that different practitioners might have been making assessments focusing only on subjects they knew about. Asset makes them look at all factors.”

Although English and Welsh YOTs have to use Asset with convicted young offenders, some teams in Scotland and Ireland have adopted the software voluntarily.

However, there are dissenters from these methods. Eileen Munro, a reader in social policy at the London School of Economics, says there are cases of children as young as nine months being identified as a future delinquent. “Scientific evidence does not back it up. Human development is varied and people do not follow a fixed path from birth. The danger comes if we start thinking they do.” She adds anyone can make a prediction about a child “but it’s about making an accurate prediction. If it’s not accurate it means it is acceptable to label a child on the basis of that [false] reasoning.”

The result of labelling a child – be it as “antisocial” or as “liable to truant” – can clearly be damaging to their self-image and self-esteem and may become a self-fulfilling prophecy. Munro says: “There is no way you can say ‘this child is a future villain’ without it having a negative impact.”

Patrick Tomlinson, director of practice research and development at childcare services provider Saccs, warns that all predictions must be supported by statistical research and evidence showing significant correlation between assessment results and potential behaviour. “Otherwise we could end up with simplistic systems to deal with complex problems that haven’t been fully underpinned with research,” he says.

Any type of assessment carries with it a potential for labelling, according to a spokesperson for the YJB. He says unstructured assessments can result in bias and prejudice and consequently “inappropriate labelling”. However, he defends the assessment approach adopted by the YJB and says it has built safeguards into these mechanisms to prevent this. Asset includes “evidence boxes” where practitioners are expected to explain the nature of any problems they identify, show how this may be contributing to the young person’s offending behaviour and demonstrate the basis for the scorings they give.

So could the use of such assessment tools lead to lazy social work practice? Munro believes that if practitioners rely solely on a tick-box approach instead of creating a relationship with the young person and their family and carers, then it would. Tomlinson adds that it might become too easy for practitioners to rely on an electronic system for answers if they do not discuss the case with their relevant colleagues.

This is a concern the creators of the tools are aware of and is the reason why guidance has been issued about how to use them and make sure the evidence boxes are filled in correctly. Baker says: “To complete a good Asset assessment takes a lot of professional skill and does not diminish social work practice in any way.” She adds that practitioners need to understand how the tools fit in with their own practice and build an open culture with their colleagues where their approach is debated and justified.

A spokesperson for the YJB agrees, and emphasises the need for practitioners to continue face-to-face work with clients.

“Asset and Onset are not a replacement for professional skills but tools to help promote defensible decision-making and better targeting of interventions.”

With this in mind, if UK social care practitioners wish to avoid the Orwellian route being considered in New Zealand to address youth offending, it is best to remember that other approaches do work. Clearly there is space for electronic assessment tools but nothing beats preventive, client-focused practice.

Knowing the score
Operating in a similar way to Asset, which is used by youth offending teams to identify those who may reoffend, Onset goes further by using a scoring mechanism to rate factors with young people who have yet to be convicted of an offence. The Onset referral and assessment framework can identify the potential risk factors of offending that need to be reduced and the protective factors that need to be enhanced. It is a four-stage process covering consent, referral and verification, assessment (including the child and parents’ or carers’ self-assessment) and planning, action and review.

To have your say, go to our discussion forum

Contact the author

More from Community Care

Comments are closed.