Data Insight: What came first: Is there a causal role of child welfare service interventions on likelihood of having criminal cautions or convictions?

Note. Child welfare first = child welfare service intervention before criminal justice system recorded contact, Around the same time = child welfare service intervention occurred one year before or one year after criminal justice system recorded contact, Criminal justice first = criminal justice system recorded contact occurred before child welfare service intervention.

The descriptive statistics shown in the full publication show that for those with different orders of child welfare service and criminal justice system formal contacts. The table showed that there was little difference between order groupings (CWS first, around the same time, and CJS first) apart from in the group where child welfare service intervention came first, who had:

• Better outcomes for school absences
• Fewer school suspensions/fixed term exclusions
• A slightly lower proportion of permanent school exclusions
• A lower proportion of custodial sentences

Boys were more likely to have their first contact be a criminal justice system contact.

Including cases where the child welfare service intervention was not due in any way due to the child’s prior formal contact with the criminal justice system, children with a child welfare intervention were more likely to have a caution or conviction by early adulthood.

This increased risk of a criminal caution or conviction increased with the level of child welfare service intervention, but those with a referral but no intervention had a higher likelihood than those who were child in need or on a child protection plan (see full publication for the table). This pattern was different to the original analysis which found a stepped increase in risk of having a criminal caution or conviction in line with the level of child welfare service intervention. The difference between the two analyses is because this analysis used a smaller and selected sample, which did not affect the referral only group but did change the child in need and child protection plan groups. This means interpretations should be made with great caution.

The child looked after population was three times more likely than those without any child welfare intervention to have a criminal caution or conviction by 2020.

This is a slightly reduced estimate than in the full regression model, but the findings provide support for the potential adverse impact of being a child looked after on risk of being criminalised. It is important to consider that the child looked after population has experienced out of home placements because there is a significant concern for their safety and wellbeing. As such, the causal effect of being a child looked after is conflated with in many cases the impact of trauma, adversity, and maltreatment, and/or the absence of healthy caregiver attachments, nurturing and stable environments, or other enrichment experiences, which is not directly measured here.

All of these potential interpretations should be made with caution considering that these findings are based on an analysis that:

• Used a selected sample rather than a full sample of children.
• There was a lot of missing data for the dates of child welfare service or criminal justice system contacts.
• Children were excluded where the timing of the child welfare service intervention and criminal justice system contact occurred within the same two-year window, and due to data limitations, this meant more accurate event timings could not be calculated.

Why it matters

The data analysis shows an increased risk for children who are referred to or receive an intervention from child welfare services in childhood (aged 5-18 years) to receive at least one caution and conviction by early adulthood. This is after restricting the sample of children with a child welfare service intervention to only include those who received the intervention one year or more before the criminal justice system caution or conviction. The findings therefore provide further evidence of a causal association between the child welfare service intervention and the greater odds of a criminal caution or conviction.

There is substantial evidence that adversity in childhood, in particular maltreatment in the form of neglect and abuse, increases the likelihood of criminal justice system contact and vulnerability to criminalisation. The findings from this report demonstrate that child welfare services, designed to protect children from harm and uphold their healthy development, do not sufficiently protect them to reduce their risk of receiving a criminal caution or conviction (so that it is equal to the general population). In fact, the findings in this report show the highest risk for a formal criminal justice system outcome exists in the population who receive the ‘highest’ welfare intervention, in the form of an out of home placement (child looked after status), in line with existing evidence in other western nations and the findings of the full analysis.

New releases of the linked dataset include all episodes and periods of care, which will allow for more accurate judgement of highest level and first child welfare service intervention. A repeated analysis that includes all cases would be beneficial once this data is available. The new release of the linked dataset will also contain details of the child welfare service reasons identified at assessment, such as reports of drug and alcohol use by the child, or indeed reports of criminal or antisocial behaviour. This information may help to flesh out understanding of the association between care experience and criminal justice system contact.

The full care experience history may give better insights into the different care experiences of those in the child looked after category, such as details of care placements (including placement type and frequency of placement change). This may enable better understanding of the reasons behind the increased odds of criminal justice system cautions and convictions for this vulnerable group.