Hester van Eeren

| Chapter 3 3 | 46 prevention programs, the costs to victims, and the valuation of fear (Cohen & Piquero, 2009) are included in this valuation (Cohen & Piquero, 2009; Cohen, Rust, Steen, & Tidd, 2004). Therefore, further research is needed into which categories of costs of crime are included in a WTP value, before determining what society is willing to pay for one CAFY. The cost categories included in the cost-effectiveness model should be reflected in the WTP value and vice versa. Also, other estimations of the societal WTP might be considered. These could, for example, be based on the cost of crime using a bottom-up approach or a breaking-down approach (Moolenaar, 2009). These methods take into account only the costs of crime, not the willingness to reduce crime levels. Third, to estimate a WTP per CAFY, it should be known what type of criminal activity is avoided in a CAFY. The seriousness of the crime, the number of times the crime is committed and the types of criminal activity can also be taken into account in defining criminal activity. Furthermore, it is important to decide on how tomeasure criminal activity. The CAFY used in our study was based on the adolescents’ self-reported contact with the police. However, criminal activity may as well be determined on the basis of police registries (Slot et al., 1992), contacts with other judicial institutions (McCollister et al., 2003a; McCollister et al., 2003b; McCollister et al., 2004), rates of reconviction (Barrett & Byford, 2012), or a delinquency score (French et al., 2008). Different definitions of criminal activity can influence the model results. For example not all committed crimes are recorded in police registrations, while self-reported measures could yield socially desirable answers. Using the CAFY in further research thus requires a clear definition of criminal activity. A final remark on this analysis concerns the interventions chosen. FFT and the Course House were chosen to illustrate the analysis in the field of crime prevention. The interventions under study, however, could be replaced by other interventions aimed at reducing juvenile delinquency, such as Multisystemic Therapy, Multidimensional Foster Treatment Care or Multidimensional Family Therapy (Netherlands Youth Institute, 2014). Contrary to a broader range of cost-effectiveness studies in the UK and US (i.e. Aos, Lieb, Mayfield, Miller, & Pennucci, 2004; Cary et al., 2013), in the Netherlands, to the best of our knowledge, the cost-effectiveness of such interventions has not yet been investigated, except for a cost-benefit analyses of ‘Maatregel Inrichting Stelstelmatige Daders’ or a case study into ‘Strafrechtelijke Opvang Verslaafden’ (van Zutphen et al., 2014; Versantvoort et al., 2005), which are both aimed at adults. Furthermore, in studying interventions in this field, the context of the interventions under study is highly important. In our illustration, we assumed that in practice, the interventions would be applied completely equivalently. However, preferences for an intervention may influence the choice for a certain intervention in reality, such as earlier experience with an intervention, specific characteristics of an adolescent, or the availability of the intervention itself. In our illustration, FFT may be, for example, used more often to avoid committing crimes, whereas the Course House could be used as an addition to a punishment under juvenile justice law, where the adolescent has already committed a crime. There may then be a higher probability of recidivism if the adolescents already have a history of committing crimes (Donker & de Bakker, 2012). These non-equivalent baseline situations may influence the measured effectiveness of the intervention. Moreover, the situation after