This article was first published in German in: Interventionen – Zeitschrift für Verantwortungspädagogik Nr. 20, 2026 and was automatically translated for international dissemination.
Original article by Maximilian Campos Ruf
Over two decades ago, Violence Prevention Network began its anti-violence work with violent right-wing extremist offenders in prisons. Soon thereafter, the approach was transferred to outside the prison system, now comprising additional measures, such as secondary prevention support for radicalised and at-risk young persons and their families. The overarching aim of this work has always been to support vulnerable individuals and radicalised violent offenders by helping them to change their behaviour, lead independent lives outside of extremist contexts, and (re-)integrate into democratic society, thereby preventing extremism of any kind. From its inception, this work has been accompanied and critically assessed by external academics and evaluators, as well as comprehensive internal monitoring, evaluation, accountability, and learning (MEAL) management processes. Looking back at more than 20 years of experience, the question arises: what results did these evaluations yield? And what does good evaluation practice require to better prepare P/CVE work for new challenges, political changes, and hybrid phenomena?
P/CVE practice and evaluation – a love-hate relationship?
Individualised social interventions, such as those commonly implemented in exit and disengagement support, are notoriously difficult to evaluate, especially regarding clear causal attributions of impact. In the past, practitioners and evaluators alike have often expressed frustration at the limitations of evaluation in this field. However, even though previous experiences with evaluations may have been difficult for all stakeholders involved, and contrary to widespread belief, practitioners in Germany are not generally opposed to well-designed and well-implemented evaluations.
Surveys conducted as part of the joint PrEval projects (1st iteration 2020-2022, 2nd iteration 2022-2025; funded by the German Federal Ministry of the Interior) clearly show that practitioners have a vested interest in good evaluation practice (Koynova et al. 2022, 21). For instance, 59% of the practitioners surveyed considered the current number and scope of evaluation processes to be appropriate, 33% even considered there to be too few evaluations, and only 8% found current efforts rather high or too high (ibid.). At the same time, it is evident that the specific needs and wishes tied to evaluation differ between stakeholders, and especially between funders and practitioners. While practitioners generally tend to prioritise evaluation formats that offer direct, practical benefits for their daily work (ibid.), funders show a greater interest in examining “achieved effects of funded projects and their sustainability” (GBMI 2019, 803).
The early days of evaluation at Violence Prevention Network
The first project evaluation at Violence Prevention Network was carried out in 2006. This first evaluation examined the then-pilot project ‘Preventive work with right-wing extremist youth in prisons in the state of Brandenburg’. The study reached very positive conclusions regarding the direct effect on participants (Lukas/Lukas 2007, 60). Participants rated the experience positively and were empowered to critically assess themselves, their criminal past, and acts of violence through the training. As a result, they were able to more carefully evaluate interactions with others, particularly by not automatically interpreting critical interactions as attacks, to listen to others, and to self-reflect (ibid.). Furthermore, their abilities to recognise and express their feelings and to verbally express themselves more articulately were strengthened through the project (ibid.). Success was even reported with regard to ideological and scene-related distancing:
“It can be noted that right-wing extremist positions were relativised and/or abandoned. This was evident, for example, from outward signs such as a change in appearance – for instance, growing out one’s hair, discarding scene-related clothing, or removing right-wing extremist symbols from their cells. There was also verbal distancing from opinion leaders or specific requests for help in leaving the right-wing extremist scene.” (ibid., 61).
Initial experiences in the follow-up care and stabilisation of participants who were released from prison also allowed for positive conclusions regarding their ability to apply what they had learnt outside the training context (ibid.).
Even then, however, a key challenge regarding the evaluation of prevention measures became apparent: one-off surveys at specific points in time may allow interesting insights into immediate effects and perceptions. Under certain circumstances, they can also provide conclusions around specific implementation challenges and possible improvements, thus unfolding a formative influence on a project, a factor highly valued by practitioners to this day. And yet, it became clear that in order to develop an even deeper, more complex understanding of how P/CVE interventions unfold impact in the long term, larger-scale studies with multiple data collection points over a significantly longer period of time are needed. This is not an easy feat. Such evaluations require time and, first and foremost, the funding institution’s political will and conviction to look beyond quick results and justifications within a short timeframe. Rather than presenting supposedly concrete but effectively insignificant results within short, 6-12-month-periods, such evaluations must engage more intensively and in a more nuanced manner with P/CVE practice. Only then can evaluation effectively contribute to a strengthened evidence base and ensure the work-field’s readiness to tackle future challenges.
Other studies from that early period also reached positive conclusions and provided supporting evidence for something that was to become the unique selling point of the German model of P/CVE: the division of labour between state institutions and civil society prevention organisations. Particularly in the context of the prison system, the author emphasises:
“The trainers have to enter the prisons as external staff so that they can operate relatively free from the usual hierarchy and the constraints of everyday prison life. Under these conditions, they offer a working partnership to participants which they can choose to join voluntarily, but which then commits them to binding, regular, and active participation.” (Lukas 2006, 73-74).
Even a further decade and a half later, this cornerstone of the German prevention landscape is still highlighted as particularly beneficial in a report by the European Commission’s Radicalisation Awareness Network:
“[CSOs] can bridge knowledge and capacity gaps that government actors lack and can provide support to target audiences that otherwise may not receive help. They have accumulated knowledge, expertise and skills in working with violent extremist clients and concerned individuals. CSOs provide exit services both within the criminal justice system and in society, using diverse approaches and methods. Their independence from government enhances their credibility and legitimacy in the eyes of their target audiences and promotes trust among affected communities. Due to their smaller organisational structure, CSOs have a great deal of operational flexibility and a wide window of availability, allowing for the provision of intensive care during sensitive periods. Many CSO-led interventions are designed to provide long-term, continuous support.” (RAN 2022, 33; emphasis added by the author)
In 2008, new approaches addressing broader target groups were evaluated in addition to the ongoing studies into right-wing extremism-targeted programmes (Lukas 2008a). These new programmes were implemented in the prison systems of Hamburg, Lower Saxony, and Berlin and for the first time also included participants with other radicalisation backgrounds, particularly Islamist extremism (Lukas 2008b). Although some challenges were noted, particularly regarding the adaptation of the original concept to other target groups, results were very positive (Lukas 2008a, 58-59; Lukas 2008b, 85-86). Yet, even though the overall positive assessment was important as a proof-of-concept, it was once again the specific references to challenges and opportunities for further development that proved to be particularly useful for future innovation. These first four evaluations formed the foundation of all future work and practical development at Violence Prevention Network in the years and decades that followed.
Recidivism rates as measures of success?
In stark contrast to an ideal, complexity-adequate evaluation design studying complex, multi-levelled social interventions such as disengagement support, two of the questions often asked are: What about re-incarceration rates, what about recidivism? Generally, studies and effectiveness assessments based on recidivism rates should be treated with caution as the underlying criteria, definitions, and study samples often vary considerably and actual comparability remains, at best, very limited. Nevertheless, in 2012 a study funded by the German Forum for Crime Prevention examined the re-incarceration and recidivism statuses of participants in Violence Prevention Network’s prison deradicalisation trainings. To this end, the author compared the recidivism data of 188 training participants imprisoned for violent offences within four years of their release with comparable recidivism data from other studies and concluded:
“The high proportion of non-recidivist violent offenders and the significantly lower re-incarceration rate among VPN training participants during the recidivism period suggest that participation in the training has a clearly positive effect in terms of post-release behaviour and the most severe subsequent sanctions.” (Lukas 2012, 21-22).
Compared with recidivism data for young offenders in 1994, the recidivism rate among training participants was 61% lower (ibid., 22). When compared with young violent offenders during the recidivism period 1995-1998, the re-incarceration rate among training participants was 58% lower (ibid.). As already mentioned, such comparisons should not be taken as absolute truth. The author himself notes that the lower proportion of custodial sentences without probation may partly result from changes in judicial sentencing practices since the 1990s (ibid.). Nevertheless, a positive trend can be inferred with regard to the pioneering work carried out by Violence Prevention Network at that time. The main function of such numbers is to justify and legitimise the money spent to set up such practical approaches. The many limitations attributed with the study of recidivism and especially their comparison across different target groups, settings, sometimes even countries, etc., however, raise the question of how relevant and useful their results actually are – and whether funds could not be used more efficiently to fund more complex and insightful evaluation approaches.
Evaluation practice throughout time
Building on these first tentative steps, a shift in evaluation practice has become evident since 2013: a greater number of evaluation and accompanying research projects were commissioned and set as requirements by funding bodies in general. This now also applied to primary and secondary prevention projects outside the prison context. In 2013, for instance, accompanying research was conducted on a project to strengthen intercultural and inter-religious dialogue in Berlin’s Wedding district and received consistently positive feedback from participants (Hayes 2013, 21-22). The process-oriented nature of this accompanying research provided particularly valuable insights for the organisation’s further development, including regarding diversity aspects, and for the future implementation of similar projects. The results were indispensable for the organisation’s continued prevention work in school settings.
As the organisation’s approaches were extended to other federal states, a combination of critical reflection, assessment of the suitability of measures in relation to the respective challenges, process-oriented and formative elements proved to be a very helpful approach to evaluation in P/CVE. Against this background, the evaluation report on the Hessen Advice Centre – Religious Tolerance Instead of Extremism was published in 2018 (Möller, Neuscheler 2018). This was based on a “mix of methods that, in a participatory manner, envisaged a process-accompanying, reconstructive-qualitative design of formative, that is design-oriented, evaluation” (ibid., 15). This evaluation focused primarily on whether the conceptually envisaged objectives could be achieved, to what extent, and on any possible unintended effects (ibid., 190). Overall, it reached the following conclusion:
“In summary, the work of the Hessen Advice Centre can be said to achieve a quality of results that not only meets the intended objectives but also produces outcomes which, particularly in work with radicalised individuals, extend to an improvement in living conditions and social status (Output III) (see also Beywl/Niestroj 2009 and Chapter 3 of this report).” (ibid., 2018)
With explicit reference to disengagement from extremism, the evaluation attests:
“The dismantling of extremist attitudes can be described as the most challenging objective in this field of work (see also the assessment by academic experts: e.g. Noricks 2009; Neumann 2013; Schmid 2013). Nevertheless, depending on the duration of the support and the individual progress made in reflection, staff are able to build on existing doubts and encourage clients to engage in critical reflection, so that extremist narratives and their inherent, strongly transcendence-oriented interpretations and courses of action that they offer can be reflected upon and dismantled. This can, among other things, ensure that clients no longer pursue efforts against the liberal democratic basic order or regard the active use of violence to achieve their goals as legitimate.” (ibid., 223)
In the spirit of realistic expectation management, however, the authors also note that “a complete and consistent reduction of facets of ideological representations across all individuals receiving support appears rather unlikely” (ibid., 223), a relevant limitation regarding the possibilities and boundaries of working with an often highly radicalised clientele. At the same time, they write:
“On the other hand, with reference to the achievement of objectives to date, it is also realistic that the previously existing [extremist] orientations lose their coherence and that, at most, isolated remnants of problematic attitudes remain, but that these no longer exert an influence on behaviour.” (ibid., 223)
The evaluation of the Berlin Advice Centre KOMPASS follows a similar line of reasoning (Schroer-Hippel 2019). It too concludes that the evaluated work, despite considerable hurdles, achieved success in terms of its objectives in the vast majority of cases:
“The evaluation shows that the counsellors are reaching the relevant target group and that the majority of them succeed in establishing counselling processes with clients who are often highly sceptical. Making initial contact and building a working relationship with clients is, in practical terms, the bottleneck for the actual deradicalisation work. In the majority of cases (64%), such a working relationship is successfully established. Where a working relationship has been established, it is found in the majority of cases (57%) – effectively as an interim assessment of the counselling work – that there is no longer a risk or that there is a clear willingness to disengage (8 out of 14 cases). In two (ongoing) cases, the level of risk had already decreased; however, the clients were still considered at-risk, albeit below the threshold for security relevance. Four further cases were still considered to require further assessment. Deradicalisation in the ideal-typical sense, as assessed using the indicator model, can be observed in a large number of cases – again as an interim result of the counselling – provided the hurdle of establishing a working relationship with the clients has been overcome (42%).” (ibid., 69)
These results suggest that counsellors succeed in “minimising risks or initiating processes of deradicalisation” (ibid., 69) among a large portion of ongoing, generally hard-to-reach cases. However, the evaluator also concludes that longer observation periods would be necessary for a conclusive assessment (ibid., 69).
Programme evaluation
During a time when a wide range of different prevention and disengagement measures are being implemented nationwide by various stakeholders, the need for and relevance of programme evaluations increases. Rather than focusing on individual projects, these evaluations examine entire funding streams relating to specific topics and seek to assess their achievement of objectives and impact. One example of this is the scientific monitoring of the pilot projects in the thematic area of ‘Prevention and Deradicalisation in the Prison and Probation System’ within the federal programme ‘Live Democracy!’ during the 2020-2024 funding period (Jakob et al. 2024). Naturally, programme evaluation results contain a much lower level of detail regarding individual projects. However, on an overarching level the study was able to draw positive conclusions regarding the development of the field over time, which can be attributed to the successful work of the organisations active in the programme. Despite significant challenges such as the COVID-19 pandemic during the survey period, structural innovations were initiated in prisons, the number of individual case consultations increased steadily over the funding period, and increasingly new target groups were reached (ibid., 118-119). With regard to the successes and impacts on the clients themselves, and while considering the difficult conditions within the prison system, the authors also draw positive conclusions:
“On the one hand, the [pilot projects] were able to conclude many individual counselling sessions with positive outcomes; on the other hand, they provided lasting impetus for substantive engagement through group sessions and offered approaches to behavioural change that were taken up by the participants. Among the two key impact factors identified are, on the one hand, the creation of a third space within the prison system where experiences of appreciation, equality, and democratic coexistence are made possible. On the other hand, the joint establishment of voluntary participation is a necessary foundation for achieving personal engagement with the programme content and fostering individual motivation for change or stabilisation.” (ibid., 119–120)
The impact factors discussed here take up two elements that were already identified as key success factors during the initial evaluations of Violence Prevention Network’s early work in the prison system.
Despite the cross-programme nature of the study, the authors were able to formulate application-oriented recommendations (see ibid., 121-124). This is likely due to the fact that all evaluators had been familiar with the field of practice for many years and were well-acquainted with the realities and circumstances of practical work. Nevertheless, such programme evaluations primarily serve a legitimising and monitoring function vis-à-vis the policy and political levels and wider society. Their formative element therefore lies primarily in the potential impact on future funding periods and programme design – provided that the results are actually taken into account by the decision-makers.
Conclusion
Plenty similar experiences and results could be cited based on the past 20+ years of P/CVE implementation practice at Violence Prevention Network. All of the organisation’s projects and counselling centres have been subjected to different types of evaluations and MEAL processes, and the above text is only intended as an illustration of how evaluation has been approached in Germany thus far – and to shine a light on the benefits and potential shortcomings of certain approaches vis-à-vis practical needs. One thing, however, becomes clear from this illustrative dive into the evaluation history of Violence Prevention Network: Apart from the fact that the evaluation reports consistently issue positive overall assessments of the work itself, it is above all those evaluations that engage with the respective projects in a more in-depth and constructively critical manner, accompanying them throughout the process, and providing formative feedback, that have been able to provide genuine impetus for the further development of practice and the professionalisation of P/CVE and disengagement work. Particularly during the early period in which the organisation’s structures were being established, such insights were essential. They are similarly relevant during times of significant turmoil and fast-paced changes in the political as well as extremist landscapes, such as today.
Over time, needs change. Today, organisations are generally less concerned with determining whether their counselling approaches are rooted in practical and scientific evidence and fundamentally capable of preventing and countering violent extremism. This is something that has been sufficiently demonstrated throughout the past decades. Instead, the focus must lie on the examination of urgent societal challenges (e.g., hybrid extremist phenomena, technological advances) and how they might require established practical approaches to adjust and to innovate to remain efficient and relevant to target groups.
Particularly at a time when extremist and broader societal challenges are subject to ever-faster change, process-oriented and formative evaluation approaches should gain additional traction. Simultaneously, there remains a need for the longitudinal studies that were mentioned at the beginning and highlighted by many of the evaluations cited above. When done correctly, with the appropriate resources, timelines, and practitioner buy-in, this would be a big leap forward for the evaluation of P/CVE. For instance, to ensure efficient processes and realistic expectation management regarding timelines and empirical data access, practitioners should be involved in the methodological development from the start. Building working relationships and trust between practitioners and evaluators has proven to be a critical bottleneck that can determine success or failure of any such endeavours. Furthermore, to increase the practical usefulness of future long-term studies, and thus practitioners’ interest to meaningfully contribute, they should be allowed to contribute their own research questions. Practitioners’ interest to critically self-reflect themselves and their work should not be underestimated and can be a very useful driver for complex, long-term evaluation processes, if taken seriously. And finally, the evaluators in charge should be well acquainted with the work field and its practical, daily realities. This not only helps to avoid frustration and false expectations on both sides but also allows for the research process to be implemented more smoothly and efficiently.
In summary: evaluation and scientific monitoring/accompaniment can exert significant positive influence on practical implementation, as well as on funding practices. The extent to which this potential can be realised, however, strongly depends on whether the type of evaluation commissioned as well as its resources and timeframes appropriately capture current societal challenges, practical realities and needs. Otherwise, evaluations risk becoming a mere formality and (expensive) tick-box-exercise without any real benefit to any of the stakeholders involved.
Author: Maximilian Campos Ruf is a scholar of Islamic and social sciences specialising in terrorism and security studies. He has been working at Violence Prevention Network gGmbH since 2017, currently serving as Director of the Research Department and as Senior International Adviser.
Further insights into the evaluation of Violence Prevention Network’s work (only in German language) can be found at: https://violence-prevention-network.de/ueber-uns/publikationen/evaluationsberichte/
Bibliography
Federal Ministry of the Interior, Building and Community (BMI). 2019. Guidelines for the funding of projects promoting democracy, diversity and the prevention of extremism (funding guidelines “Demokratie leben!”). GMBl 40/2019: 801-803.
Hayes, Jenny. 2013. MAXIME Wedding – Pilot project for the prevention of Islamic extremism in the Berlin district of Wedding (Violence Prevention Network e. V.). Process documentation report. Berlin: IBI – Institute for Education in the Information Society e. V. URL: https://violence-prevention-network.de/wp-content/uploads/2019/07/evaluation-maxime-wedding.pdf (Last accessed: 07/04/2026).
Jakob, Maria, Diegmann, Daniel, Herding, Maruta, Frank, Anja. 2024. Innovations, Impacts, Transfer. Final report on the scientific monitoring of the pilot projects in the thematic area ‘Prevention and Deradicalisation in the Prison System and Probation Service’ within the federal programme ‘Demokratie leben!’ for the funding period 2020–2024. Halle an der Saale: German Youth Institute e. V. URL: https://violence-prevention-network.de/wp-content/uploads/2026/03/Abschlussbericht-Modellprojekte-Strafvollzug_DJI-2020-2024.pdf (Last accessed: 07/04/2026).
Koynova, Svetla, Mönig, Alina, Quent, Matthias, Ohlenforst, Vivienne. 2022. Monitoring, Evaluation and Learning: Experiences and Needs of Practitioners in the Prevention of Right-wing Extremism and Islamism. PRIF Report 7/2022. Frankfurt am Main: Peace Research Institute Frankfurt. URL: https://www.prif.org/fileadmin/Daten/Publikationen/Prif_Reports/2022/prif2207_barrierefrei.pdf (Last accessed: 07/04/2026).
Lukas, Veronika, and Helmut Lukas. 2007. Evaluation of the pilot project ‘Preventive work with young people with right-wing extremist leanings in prisons in the state of Brandenburg’. Final report. Berlin: social.consult. URL: https://violence-prevention-network.de/wp-content/uploads/2018/07/evaluation_verantwortung_ubernehmen_-_abschied_von_hass_und_gewalt_teil_1.pdf (Last accessed: 07/04/2026).
Lukas, Veronika. 2006. Evaluation of the pilot project “Democratic and social action as a key competence for work, training and social integration. Development of a set of measures for social, socio-cultural and political education to address right-wing extremism, xenophobia and violence”. Final report. Berlin: social.consult. URL https://violence-prevention-network.de/wp-content/uploads/2018/07/evaluation_verantwortung_ubernehmen_-_abschied_von_hass_und_gewalt_teil_2.pdf (Last accessed: 07/04/2026).
Lukas, Veronika. 2008a. Evaluation of the pilot project “Preventive work with young people with right-wing extremist leanings in prisons in the state of Brandenburg”. Berlin: social.consult. URL: https://violence-prevention-network.de/wp-content/uploads/2018/07/bericht_2008_wriezen_cottbus.pdf (Last accessed: 07/04/2026).
Lukas, Veronika. 2008b. Evaluation of the pilot project “Preventive work with young people and young adults in prisons in the states of Berlin, Hamburg and Lower Saxony”. Interim report 2008. Berlin: social.consult. URL: https://violence-prevention-network.de/wp-content/uploads/2018/07/bericht_2008_hamburg_hameln_berlin.pdf (Last accessed: 07/04/2026).
Lukas, Helmut. 2012. Study on the legal probation of participants in VPN training courses in the youth prison system. Annotated summary. Berlin: social.consult. URL: https://violence-prevention-network.de/wp-content/uploads/2019/07/vpn-evaluation-lukas-2012.pdf (Last accessed: 07/04/2026).
Moeller, Kurt, and Florian Neuscheler. 2018. Final report on the evaluation of the Hessen Advice Centre – Religious tolerance instead of extremism. Esslingen. URL: https://violence-prevention-network.de/wp-content/uploads/2019/02/Abschlussbericht-Evaluation-Beratungsstelle-Hessen.pdf (Last accessed: 07/04/2026).
Radicalisation Awareness Network (RAN). 2022. The Role of Civil Society Organisations in Exit Work. Brussels: European Commission. URL: https://home-affairs.ec.europa.eu/system/files/2022-07/ran_paper_role_civil_society_organisations_exit_work_052022_en.pdf (Last accessed 07/04/2026).
Schroer-Hippel, Miriam. 2019. Tolerance instead of extremism. The KOMPASS counselling centre (VPN). An evaluation report. Berlin: Berlin State Commission Against Violence. URL: https://violence-prevention-network.de/wp-content/uploads/2021/06/Evaluation-KOMPASS_BFG_66_Heft1.pdf (Last accessed: 07/04/2026).
