Niels Raaijmakers

In the police context, the term vulnerability has become an important criterion to determine which victims are most in need of receiving protection. The identification of vulnerability, however, is complex given the heterogenous nature of the concept (Brown et al., 2017). It can have different meanings, such as the risk of repeat victimization or the impact of the experience of victimization. Furthermore, it can be recognized using a wide range of indicators, such as characteristics of the victim, the specific victimization-incident or the context in which victimization takes place. To aid police officers with the identification of vulnerability, several assessment instruments have been introduced in police practice. Examples are the Individual Assessment, as stated in the Article 22 of EU Victims’ Directive (2012/29/EU), and various risk assessment instruments, such as B-SAFER, SASH or DASH. The heterogenous nature of vulnerability, however, makes it difficult to address the concept’s full spectrum during assessments. Narrow interpretations of vulnerability in both the instruments’ designs and the practical use by police officers suggest that only a subset of vulnerable victims may be identified correctly. Partial and flawed identification of vulnerable victims may lead to overfocussing on one group of victims while overlooking others. This dissertation addresses the identification of vulnerability in the police context. Its central research question is: How and to what extent are vulnerable victims identified in daily police practice? This thesis has two aims. Firstly, it strives to advance our understanding of how police officers identify victim vulnerability when a victim reports a crime. Secondly, it aims to evaluate tools that may help to assess the risk of repeat victimization more effectively. Both aims are addressed from multiple perspectives, each highlighting relevant aspects of police practice. MAIN FINDINGS PER CHAPTER Chapter 2 examines how police officers identify and respond to vulnerable victims. Using the Articulated Thoughts in Simulated Situations approach (ATSS; Davison et al., 1983), it analyzes Dutch police officers’ responses to two fictional vignettes portraying a victim reporting a crime. Findings demonstrate that officers primarily relied on incident-related factors, such as the type of crime and the immediate consequences of victimization. Additionally, the study revealed that officers proposed protection measures, such as external referral or stop-conversations in stalking cases, regardless of the outcomes of their vulnerability assessment, which points at taking a ‘better safe than sorry’ approach. This seems to contradict worries in earlier research that victims in need may be overlooked, but it raises a different issue: the risk of over-protection. Providing measures to victims with minimal or no protection needs may strain scarce resources and disrupt natural coping mechanisms. This chapter provides relevant insights into how officers assess vulnerability based on their professional judgement (and thus without assessment instruments). It also demonstrates the shortcomings of this assessment method. Chapter 3 addresses the development and validation of ProVict: a risk assessment instrument for assessing the risk of repeat victimization, based on structured police registration data, more specifically victims’ prior contacts with the police. Findings demonstrate that prior contacts with the police could be used to develop a valid risk assessment instrument, as the instrument could correctly identify a majority of the repeat and non-repeat victims. However, the instrument also produces some false negatives in which repeat victims were placed in the low-risk category, and a considerable number of false positives in which non-repeat victims were placed in the high-risk category. Both misclassifications can be problematic. Overlooking at-risk victims leaves them vulnerable to new harm, while allocating measures to those with minimal needs risks over-protection. This chapter demonstrates the potential of using structured police data for risk assessment, while also showing significant limitations. Chapter 4 builds upon Chapter 3 and incorporates unstructured police data in ProVict, such as police officers’ notes and interview reports. It explores to what extent unstructured police data, both independently and combined with structured data, could be used for assessing the risk of repeat victimization. Findings show that unstructured data independently yielded acceptable to excellent predictive validity. An instrument that combines both structured and unstructured data also shows acceptable to excellent predictive validity and correctly identifies the majority of repeat and non-repeat victims. The findings, however, still carry a risk of misclassifying victims. ProVict can be promising, but should not be used as a stand-alone assessment method. To reduce the risks of misclassifications in the future, it’s important to also include additional risk factors and protective factors from outside the police context in the instrument. Chapters 2 through 4 showed that the risk of under- and over-protection could occur in methods applied to assess vulnerability. To address this further, Chapter 5 critically reviews how five key shortcomings of vulnerability assessment instruments contribute to the risk of under and over-protection. Fragmented definitions, narrow interpretations by officers, biases and the withholding of adequate protection measures may lead to under-protection and thus to the risk of overlooking victims who have an increased risk of repeat victimization. In contrast, neglecting resilience factors when assessing vulnerability, overriding assessment outcomes and providing measures to victims with minimal or no protection needs, may result in over-protection. Both misclassifications reflect a discrepancy between the risk for a victim to become confronted with new harm and victimization and the police assessment of this risk. This difference may expose victims either to further harm or undermine empowerment and natural coping mechanisms. This chapter calls for greater awareness of under- and over-protection and advocates for a more nuanced understanding of vulnerability. CONCLUSION AND DISCUSSION To answer the central research question, this dissertation highlights two key findings. First, it concludes that officers assess vulnerability from an incident-focused perspective, thereby overlooking non-incident related indicators. Second, across different assessment methods, this dissertation demonstrates that not all vulnerable victims and non vulnerable victims are identified correctly. Different assessment methods all demonstrated the risks of under- and/or overprotection. Both types of misclassifications can have serious consequences for how victims process victimization-incidents. This challenges police organisations to find a better balance between protecting those in need while simultaneously avoiding unnecessary interventions for those with minimal and no needs for additional protection. This requires a broader and more nuanced approach to victim vulnerability. Additionally, this may require more extensive collaboration between the police and other organizations that provide support to victims. Together with new research on the nature and consequences of misclassifications, a more nuanced understanding of vulnerability can contribute to an enhanced identification of vulnerable and non-vulnerable victims.