The development of a risk and needs assessment instrument for truancy

The aim of this study was to develop a risk and needs assessment instrument for truancy for use by juvenile probation. No validated tools are yet available for this purpose, while about half of all juvenile probation cases involve truancy in the Netherlands. It was examined whether risk/protective factors for general delinquent behavior were predictive for truancy and whether a risk assessment instrument could be developed based on these factors with sufficiently high predictive power. The sample consisted of 10,233 juveniles between 12 and 18 years of age (Mage = 16.21; SD = 1.34) who had been convicted for committing an offense. CHAID analyses were performed to develop (a) a risk model containing both static and dynamic risk/protective factors to estimate the level of risk of truancy and (b) a needs model containing only dynamic risk/protective factors to gain insight into the extent to which (and the way in which) the risk can be reduced. A split-sample validation was applied, with half of the sample being used to construct the models and the other half to validate the models. The AUC values of both risk models corresponded with a medium effect size (AUC = 0.69 for the risk model; AUC = 0.67 for the needs model), which can be considered sufficient to justify their use as instruments for risk and needs assessment for truancy for use by juvenile probation. Future research should examine whether and how the instrument (which was called Actuarial Risk and Needs assessment Instrument for Truancy [ARNIT]) can be further improved by adding specific truancy risk/protective factors.