The institutionalisation of evaluation and its potential for isomorphism

Abstract

This article investigates how an imbalance of institutional subsystems can influence the quality of the institutionalisation of evaluation. Building on the concept of coercive isomorphism and the recent country-comparative literature, it theorises that structural asymmetry, where the ‘political system’ administers competitive funding structures, forces practitioners in the ‘social system’ to adapt. These ideas are applied to the policy area of preventing and countering violent extremism (P/CVE), where the literature and initial empirical evidence suggests that governments’ funding power can compel practitioners not only to institutionalise evaluation, but also to adopt standardised, output-focused forms of evaluation. This latter form of coercive isomorphism, driven by the political need for quick accountability, severely diminishes the potential for organisational learning and risks reducing evaluation to a symbolic practice. Nevertheless, the article recognises an adaptive capacity for constructive dialogue and calls for future comparative research looking at different policy areas.

Introduction

The concept of institutionalisation proves analytically useful for determining how deeply embedded the practice of evaluation has become in different countries or policy areas. For over two decades, a small body of literature has proposed various analytical frameworks with indicators that reflect both old and new institutionalist ideas, seeking to make the institutionalisation of evaluation empirically measurable.

The latest of such research distinguishes between three subsystems in which the institutionalisation of evaluation can be observed at national level: a political system, a social system and a system of professionalisation (Stockmann et al., 2020). These systems are distinct spaces of institutionalisation, yet there are also initial signs of entanglement. This article builds on these observations and asks: To what extent can the institutionalisation processes in these subsystems influence each other, and what does this mean for the extent and quality of this institutionalisation overall?

To answer this question, this article draws on another concept from institutional theory, namely isomorphism. After first explaining the concept of institutionalisation and reviewing the literature that has operationalised it in relation to evaluation, the article makes a theoretical argument that builds on the latest comparative research of this literature. It argues that in policy areas where policymakers in the ‘political system’ administer competitive funding structures for services conducted by practitioners in the ‘social system’, two forms of coercive isomorphism within the latter can arise. Due to pressure from the ‘political system’ or competition for public funding from this system, service providers may, on the one hand, increasingly embed evaluation in their own organisational structures or allow their activities to be externally evaluated. On the other hand, they may even adapt the form of these evaluations to the informational needs of their funders. While the former can be considered a positive development towards an evidence-based future, the latter can be problematic in some policy fields. This is the case, for example, in the field of preventing and countering violent extremism (P/CVE), which the article uses to seek initial empirical validation. Finally, the article discusses the implications of these findings and suggests further research avenues.

The institutionalisation of evaluation: Borrowing from sociology

Much like many other concepts in the humanities, that of the institution and the associated process of institutionalisation defy an inter- or even intradisciplinary definition (Colyvas and Powell, 2006: 306). However, sociological definitions are particularly frequently used to render these concepts analytically useful for public policy research. Jepperson (1991: 145), for example, provides such a classical sociological understanding of institution and institutionalisation, describing the former as ‘a social order or pattern that has attained a certain state or property’ in that it is ‘chronically reproduced’. Moreover, for a social pattern to be an institution, ‘departures from the pattern [need to be] counteracted in a regulated fashion’ but their survival must owe ‘to relatively self-activating social processes’. For Jepperson, institutionalisation, on the other hand, denotes ‘the process of such attainment’ (Jepperson, 1991). However, an institution is not only a constantly reproduced social pattern but also transports certain cognitive attitudes about this pattern that go beyond a ‘taken-for-grantedness’. As Vukasovic (2014: 47) describes, institutionalisation is the process ‘through which new, initially ambiguous, unfamiliar and resisted “ways of doing things” become structured, desirable, appropriate, comprehensible, common-place and routinised’. The institution is thus to be distinguished from mere routine actions, which may bear similarity, yet ‘fail to become deeply cognitively embedded’ (Colyvas and Powell, 2006: 312).

Sociologists often study institutions like marriage or elections (Jepperson, 1991: 144), but institutionalisation can apply to a much broader range of behaviours (DiMaggio and Powell, 1991), including evaluation. This activity can be defined in the narrow sense as the systematic assessment of policies, programmes or projects using scientific methods with the purpose of determining their ‘merit, worth and value’ (Vedung, 1997: 3; cf. Mertens, 1998). It is particularly public policy where evaluation has come to be considered one of the social patterns which can become institutionalised by being constantly reproduced, ‘structured, desirable, appropriate, comprehensible, common-place and routinised’ (Vukasovic, 2014: 47). Gaarder and Briceno (2010: 291) define the institutionalisation of evaluation as ‘a process of channelling isolated and spontaneous (programme) evaluation efforts into more formal and systematic approaches, on the presumption that the latter provide a better framework for fully realising the potential of the evaluation practice’. While this perspective presumes agency in shaping institutions, Jepperson (1991: 145) posits that institutions are often reproduced through routine rather than deliberate intervention (cf. Colyvas and Powell, 2006: 306). Yet institutional change can still result from what Vukasovic (2014: 48–49) calls institutional work – deliberate efforts by actors to create, maintain or disrupt institutions (cf. Lawrence et al., 2009). This agency, despite the overall self-reinforcing character of institutions, is key to understanding the ‘genesis of a transformation of institutional arrangements themselves’, one of two scientific enquiries that have developed about institutions, according to Scharpf (2000: 2). The other focuses on the ‘consequences [. . .] institutions may have for actors and actions within their domains’ (Scharpf, 2000).

Measuring the institutionalisation of evaluation empirically: Past and recent avenues

Associated with the former line of scientific inquiry is a tradition of scholarship interested in the institutionalisation of evaluation in comparative perspectives, either looking at different countries or policy areas. This body of literature employs both ‘tangible’ and ‘less tangible’ indicators in line with the ‘new institutionalist’ tradition. It emphasises informal norms, values and ideas that guide action (March and Olsen, 2009; Meyer and Rowan, 1977) in addition to formal rules, laws and structures that are associated with the ‘old institutionalist tradition’ (Lowndes, 2002; Rhodes, 1997).

Following the seminal edited volume ‘Program Evaluation and the Management of Government’ (Rist, 1999), which addressed general principles of programme evaluation and their application within government management and public administration, the first country-comparative study to examine the institutionalisation of evaluation in 19 countries and three international organisations was the ‘International Atlas of Evaluation’ (Furubo et al., 2002), laying important conceptual groundwork for future studies. Though influential, it lacked a consistent conceptual framework, using various terms such as ‘diffusion of evaluation’; ‘history, development and future directions of evaluation’; and ‘evaluation culture maturity’. However, the authors seemed to prefer the latter, proposing nine indicators for measuring the ‘full maturity of the evaluation culture’, among them one related to the spread of evaluation activities in the public sector or another one on the presence of associations or societies for evaluators in the country (Furubo et al., 2002: 7–9; Jacob et al., 2015: 10–12). Due to using an iterative process of scoring the absence or presence of each indicator and aggregating these scores for country comparative purposes, Jacob (2004: 141) later questioned the resulting comparative insights. Further pointing to the terminological flexibility of this research, his follow-up study (Jacob et al., 2015) drew on the same indicators as the Atlas yet specified their phenomenon of interest as ‘evaluation culture and the institutionalisation of evaluation’.

This makes it all the more surprising that at the same time Varone and Jacob (2004) were working on a very similar study independently of Furubo et al. (2002), exploring the ‘degree of organisational and procedural embedding’ of public policy evaluations in 18 democratic regimes and proposing very similar indicators. However, similarly, these authors claim to ultimately analyse the ‘institutionalisation’ of evaluation, defining this in a ‘very limited sociological sense’ as the ‘formal organisation or procedural rule that provides actors with a framework ensuring a certain predictability of their reciprocal behaviour and, consequently, the outcome of collective action’. The institutionalisation of evaluation, they explain, corresponds to a ‘routinisation of the expected, if not mandatory, use of evaluation and can be measured against its effective practice within politico-administrative bodies and, more broadly, within public action networks’ (Furubo et al., 2002, translated from French). In his doctoral dissertation, Jacob (2004: 137–140, translated from French) explains that he and Varone understand the two ‘main modalities of institutionalisation’ to be specific evaluation bodies in government, parliament and the administrative system, especially the Court of Auditors, and the existence of a national epistemic community, embodied though the presence of a national evaluation society, a scientific journal and quality standards. In addition, they consider evaluation clauses, the size of the evaluation community and the number of evaluations conducted to be important indicators. Similar to Furubo et al. (2002) and later Jacob et al. (2015), Varone and Jacob (2004) used a scoring method, yet placed greater emphasis on the more ‘tangible’, formal aspects of institutionalisation.

Alternative approaches have since emerged. Widmer et al. (2009), for example, eschewed scoring entirely in favour of detailed country case studies, guided by broad thematic prompts. Though this provided richer context, the volume focuses more on the state of evaluation and its professionalisation. Despite comparing Germany, Austria and Switzerland, it also contrasts the professionalisation of evaluation in multiple policy areas of the three countries and discusses initial reasons for ‘evaluation pioneers’ and conversely ‘fields with particular development needs for evaluation’ (Widmer and Beywl, 2009: 515, translated from German). Here, it is already apparent that understanding the state of evaluation and its policy area dependency requires an understanding of context beyond political spheres and an epistemic evaluation community. External parties commissioned for policy implementation or evaluation are to be considered as participants in the evaluation market, too (Stockmann and Meyer, 2016; Widmer and Beywl, 2009).

More recently, the academic Centre for Evaluation (CEval) and the department of sociology at Saarland University in Germany developed a systematic analytical framework that is able to account for such complexity for the purpose of cross-country comparison of the institutionalisation of evaluation. The Globe project, the largest comparative project on the institutionalisation of evaluation in national states to date, developed a comprehensive, theoretically anchored analysis grid to structure various country case studies on the state of evaluation in three world regions and in a total of 38 countries (Stockmann et al., 2020, 2022, 2023). Drawing on sociological institutionalism, this model not only integrates formal and informal aspects of institutionalisation but also foregrounds civil society’s role – largely absent in earlier frameworks. Importantly, the model avoids the oversimplification of scoring and instead allows for context-sensitive, qualitative comparison of institutionalisation processes in different countries. The next section will introduce this most recent framework and explain its recognised and, so far, untapped theoretical and empirical potential in understanding difficulties in institutionalisation.

Measuring the extent of institutionalisation comparatively: A recent analysis grid

Out of a wide range of potential subsystems in the social world, Stockmann et al. (2020) selected three as most relevant for institutionalisation of evaluation at the national level. The ‘political system’ focuses on the extent to which evaluation is embedded in structures and processes among political actors. Indicators for institutionalisation in this system are the existence of national legislation or regulatory provisions that mandate evaluations as well as national or sector-specific evaluation policies, the role of parliament or courts of auditors in conducting evaluations and, crucially, the actual use of evaluation as a sign of commitment to evidence-based governance (Meyer et al., 2020: 16). The final synthesis of the comparative results, however, distinguishes the institutionalisation of evaluation from its use (Stockmann et al., 2021: 390).

The ‘social system’ is the second relevant subsystem, highlighting the importance of demand and supply of evaluation beyond political actors. This is grounded in the recognition that evaluation is not only important within political circles but can also be used by civil society in both a collaborative and ‘watchdog’ capacity. Hence, measuring the extent of institutionalisation of evaluation also involves examining the extent to which civil society participates in the evaluation market, whether as providers, advocates or critical users, and the degree to which they push for evaluations to enhance governmental accountability. Another indicator for institutionalisation of evaluation in this ‘social system’ is the awareness of evaluations and the presence of a ‘public dialogue around evaluation results’ (Meyer et al., 2020: 17; Stockmann et al., 2021: 401). By foregrounding civil society as a distinct actor, this social system expands the conventional focus beyond political and administrative mechanisms, marking a novel contribution to the set of indicators previously used to assess the institutionalisation of evaluation. The third and final system is the ‘system of professionalisation’, which, similar to previous operationalisations of institutionalisation, captures to what extent the demand, conduct and use of evaluation are accompanied by competence-building and enhancing measures as well as professionalisation. Indicators for this system include education on, communication about and standards for evaluation (Stockmann et al., 2021: 408).

Using this analysis grid for measuring the institutionalisation of evaluation on a national level comparatively, the research project by Stockmann and his colleagues finds a great heterogeneity of institutionalisation in these three systems, and also great variety regarding different policy areas. However, in all three volumes of the Globe project, the ‘political system’ is identified as the strongest, even though it also shows deficits in institutionalisation (Stockmann, 2018) and the social system as ‘by far the weakest system’ (Stockmann and Meyer, 2022; Stockmann et al., 2021: 406; translated from German). This points to a disbalance in promoting questions about its causes and consequences: Why is the ‘political system’ the strongest and what consequences does this have? These are questions that no longer concern the extent of institutionalisation, which is the focus of the Globe research project, but rather the quality of this institutionalisation.

The authors do explain that the success of institutionalisation has ‘totally different and individual reasons’ per country and that there is no ‘universal development path, which all countries more or less equally follow’ ( Stockmann et al., 2021: 414; translated from German). Nevertheless, they subsequently discuss initial ideas for influencing factors (Stockmann et al., 2021: 416f.). Moreover, the authors explore how the indicators in the three subsystems correlate with the scales of each system. They note that ‘basically, all indicators except civil society demand (where no differences can be observed) have a certain influence on the three scales of each system’ (Stockmann et al., 2021: 414; translated from German). This means that an indicator belonging to the political system, for example, correlates not only with the overall degree of institutionalisation in the political system, but also to a certain extent with that of the social system and the system of professionalisation. This is a strong indication that the political system and the social system do not coexist in complete isolation from one another. However, the system of professionalisation appears to be quite detached. Here, there is a ‘low correlation between professionalisation and the two governance systems’ (Stockmann et al., 2021: 415; translated from German). Hence, the professionalisation of evaluation does not seem to be driven by either the political or the social system.

This hints at an entanglement that raises questions. A logical consequence would be to examine why and how the ‘political’ and the ‘social’ subsystems interact – especially since the former was identified as more dominant than the latter. Moreover, the consequences are relevant, especially how an entanglement may affect the quality of the institutionalisation of evaluation and the type of insights evaluations provide. Given that the institutionalisation of evaluation was found to vary greatly across policy areas, answering such questions requires focusing on a policy area in order to approach these questions in adequate depth. In the following, the article draws on another concept from (neo)institutional theory, that of isomorphism, to attempt to theorise the why, how and consequences of potential entanglement. Afterwards, the article applies these ideas to the policy area of P/CVE.

Institutional isomorphism between the three subsystems

Isomorphism is a neo-institutional concept that broadly describes the adaptation of organisations to the requirements of their environment, resulting in them becoming increasingly similar. Meyer and Rowan (1977) and DiMaggio and Powell (1983) in particular were instrumental in establishing this concept, albeit in slightly different ways (cf. Becker-Ritterspach and Becker-Ritterspach, 2006: 104ff). The latter distinguished between three different types of isomorphism, namely coercive, normative and mimetic isomorphism. Coercive isomorphism describes a process in which organisations adopt models from their environment or from other organisations due to a dependency resulting from external coercion or an internally felt coercion stemming from external context. Normative isomorphism describes an isomorphism that feeds on professionalisation, while mimetic isomorphism is driven by uncertainty and is manifest in organisations adapted to those which they consider role models. The concept of isomorphism offers an institutional theory-based opportunity to analyse the entanglement between the institutionalisation of the three subsystems in its form and consequences.

A context in which isomorphism can arguably occur and which this article primarily addresses is one in which structural conditions give the ‘political system’ the ability to make demands towards certain actors in the ‘social system’. This means that the potential objects of evaluation are not inherent in the ‘political system’, but external to it. Actors in the ‘political system’ do not commission, conduct and make use of an evaluation of a strategy that was formulated in the system, but rather require evaluation from actors in the ‘social system’ that they have funded to deliver a certain service. Such outsourcing, which is common in various policy areas, creates a structural dependency, which might offer at least a part of an explanation for why the ‘political system’ outweighs the ‘social system’ in strength when it comes to the institutionalisation of evaluation. In some policy areas, the ‘political system’ and parts of the ‘social system’ are in an asymmetrical relationship since the former administer competitive funding structures on which the latter often depend. This is similar to an observation Widmer and Beywl (2009: 516) make, namely that outsourcing policy implementation increases the informational needs of administration and thus leads to more evaluation.

However, such an asymmetrical structure of dependency creates, one could argue, the conditions for coercive isomorphism: Funders in the ‘political system’ are in a unique position to make evaluation a prerequisite for public funding, or at least to incentivise it, which leads to evaluation becoming increasingly institutionalised in the ‘social system’ and hence the organisation life of the service deliverers. The relationship of dependency and competition for limited public funding can also create intrinsic motivation, even for organisations that may not currently be part of the funding structure but would like to be in the future. Those who can demonstrate that they can prove the value of their work through evaluation, and thus feed the informational needs of funders, are more likely to secure their existence through state funding. However, it could also be argued that normative isomorphism can occur as an alternative or coexisting phenomenon, namely through intrinsic professionalisation efforts. Identifying the impact of one’s own work can, provided sufficient capacity is available, be entirely intrinsically motivated, for example, in order to further improve it.

If such isomorphism manifests in organisations increasingly behaving in similar ways, being receptive to internal or external evaluation and institutionalising this, there is little cause for concern. On the contrary, in the interests of evidence-based practice, it would be a rather welcome development if evaluation became an integral part of all service providers, provided that the necessary time, financial and human resources are available and that this does not create problematic gaps elsewhere. However, depending on the policy area, it could be viewed rather negatively if this isomorphism took the form of standardisation of evaluation. This could mean that the expectations of the ‘political system’ regarding the form of evaluation and thus the information to be generated are transferred to implementing service providers due to structural dependency. Negative consequences would be, for example, that the services are evaluated but mainly serve symbolic purposes because they do not provide adequate information for meaningful instrumental use. The article will elaborate on these ideas in more detail below in the context of P/CVE, a policy area which, due to its structural asymmetry and high justification pressure, lends itself well for an initial empirical application.

Evaluation in the policy area of P/CVE

The policy area commonly referred to as P/CVE brings together political and administrative actors, practitioners and academics (Walkenhorst, 2019) who develop, fund, implement and conduct research on interventions that share a common goal: to proactively prevent the emergence of violent extremism at an early stage (Benjamin et al., 2021; Sjøen and Jore, 2019), to redirect individuals already involved in violent extremism back towards democratic values (cf. Koehler, 2013) or to support disengagement from violent extremist groups and enable reintegration into society (cf. Gielen, 2020; Marsden, 2017).

The social and security relevance of this policy area puts it under great performance pressure, which, however, can involve diverging expectations. Political decision-makers feel a responsibility to demonstrate that the considerable sums of taxpayers’ money invested in P/CVE are being used effectively (Deutscher Bundestag, 2023: 3; Nehlsen et al., 2020) since ineffective P/CVE programmes and projects waste resources or may even reinforce the drivers of radicalisation they seek to address (Clément et al., 2021: 8; Koehler, 2017: 188). Evaluations promise to fulfil this need for accountability. However, the latter does not only refer exclusively to quality criteria such as effectiveness but also includes normative standards such as fairness and proportionality. Critical voices have long pointed towards the securitisation of social work for P/CVE, or neoliberal logics underlying a ‘local turn’ in P/CVE (Finch and McKendrick, 2019; Heath-Kelly and Shanaah, 2025; Shanaah and Heath-Kelly, 2023: 1725; cf. Sivenbring and Andersson Malmros, 2019), as well as their consequences such as a stigmatisation of especially Muslim communities (Heath-Kelly, 2013; Kundnani and Hayes, 2018). In this context, evaluation also proves a useful tool to examine P/CVE interventions for fairness and proportionality, to uncover adverse side effects and to demonstrate the benefits of approaches that do not rely on risk-based logics and thus avoid securitisation (cf. Marsden, 2017). However, assessing quality criteria such as effectiveness, fairness and proportionality through evaluation is not only in the interests of P/CVE policymakers and critical research, but can also reflect the professionalisation interests of P/CVE practitioners (Malet, 2021).

The policy area of P/CVE thus fulfils the conditions that, according to Widmer and Beywl (2009: 521), should result in high levels of evaluation: pressure to justify, conflicts of values and cost pressure. However, while evaluation is ‘booming’ in other policy areas (Stockmann, 2018; Stockmann and Meyer, 2016), this is far from the case regarding P/CVE (Hirschi and Widmer, 2012; Horgan and Braddock, 2010; Malet, 2021). Although recent national studies in Germany show a slow upturn in self-evaluation (Uhl and Kattein, 2024: 17; Uhl et al., 2022), international comparative studies (Bressan et al., 2024) continue to confirm that evaluation practices in this policy area remain low overall. Shanaah and Heath-Kelly (2023: 1725) problematise that ‘the process of P/CVE adoption and implementation takes place in the context of, and despite, little to no solid empirical proof about the effectiveness of such policies’. Some authors even argue that most P/CVE practice operates without a clear theory of change – that is, without having articulated the assumptions on which their work is based or how they intend to achieve the desired change (e.g. Lewis et al., 2020).

It is not only the great performance pressure but also the structural characteristics of the policy area that make it a well-suited case for exploring possible isomorphisms in the institutionalisation of evaluation between the political and social systems. These theoretical ideas are backed by initial empirical evidence. This is taken from an ongoing research project of the author, involving a qualitative comparative study across Europe and an expert survey on the topic, spanning over 14 countries and involving the views of 47 policymakers, practitioners and academics. References to such views include the ISO country code and the participant number. Moreover, the article draws on two international expert surveys on the topic with 33 P/CVE practitioners and researchers in 14 countries (Bressan et al., 2024) and 30 participants in 12 countries, respectively (Stoffel et al., 2025), as well as further research insights from a German-focused research project (Peace Research Institute Frankfurt, 2024; PrEval-Forschungsverbund, 2025). Where appropriate, additional literature sources are referred to.

The P/CVE field is precisely one such policy area where political actors can make demands for evaluation on actors in the social system, thus creating a context for coercive isomorphism. Usually, there is a political strategy or action plan formulated by the responsible ministries that describes which phenomena of violent extremism should be prevented and specifies the structures for prevention work in the country. There is considerable variation between countries in how these goals are implemented. In many countries, however, a so-called ‘local turn’ has taken place, with municipalities or civil society actors implementing individual P/CVE projects in line with these political priorities – and this is supported by national P/CVE funding programmes. In Germany, for example, there have been several such funding programmes since 1992, currently including the Live Democracy! programme, whose current funding period runs from 2025 to 2032. In the 2024 financial year, it had a budget of €182 million, and towards the end of the last period, in March 2024, around 700 funded projects were counted (Deutscher Bundestag, 2025: 2).

It is also the case that many – though not all – of these funding structures require evidence of effectiveness, including evaluation of funding recipients, or commission evaluations of funded projects. Almost every German P/CVE funding structure has undergone a scientific programme evaluation to date, and individual funded projects have been evaluated. Although not to the same extent, this can also be observed in other countries, for example, in the case of the so-called ‘Versterkingsgelden’ in the Netherlands or the ‘National Impulse Programme’ in Switzerland. International and European survey data also show that evaluation requirements are often imposed on the ‘social system’ by the ‘political system’: the majority of international and European experts surveyed confirm that it is mostly governments who request or initiate P/CVE evaluations. In the first international survey, 75.8 percent of 33 respondents named governments (Bressan et al., 2024: 34), and in the second international survey, the percentage even increased to 83.3 percent of the 30 experts surveyed (Stoffel et al., 2025: 13). In the European survey, just under a third of the 47 experts surveyed stated that political decision-makers always or almost always initiate evaluations. The more detailed responses from European participants, for example, to the question of whether there were administrative requirements for the evaluation of P/CVE measures, confirmed this. Government funding bodies thus occupy an important position of power by establishing formal rules for evaluation (e.g. DKN-01; FIN-02) or making evaluation a condition for funding (e.g. DEU-01; CHE-01; GBR-01). Some experts even directly confirm the imbalance between the ‘political’ and ‘social’ systems by stating that, in the absence of evaluation requirements by the funding body, evaluation would not take place (NLD-01).

Moreover, it should be mentioned that there is hardly any country that has a system of professionalisation with regard to evaluation in the policy area of P/CVE. There are isolated support formats in individual countries, for example, a toolkit for evidence-based work in radicalisation prevention (ESS, 2019) or a Monitoring and Evaluation Centre of Excellence at the International Centre for Counterterrorism in the Netherlands (ICCT, 2024), but the training courses offered to date have mainly been provided transnationally, through European projects such as the Radicalisation Awareness Network, the EU Knowledge Hub on Prevention of Radicalisation or the INDEED project, which has provided an evidence-based evaluation model and associated guidance (Malkki et al., 2023). Even if P/CVE practitioners wanted to professionalise due to intrinsic motivation, they have little support for doing so, which weakens any argument of normative isomorphism. Instead, coercive isomorphism comes to the fore, which will be discussed in the next section. This is also an additional characteristic that makes P/CVE differ from other policy areas: the deficiencies in or lack of a ‘system of professionalisation’ means that there is no normative counterweight to political pressure for evaluation.

Coercive isomorphism in the institutionalisation of evaluation in the policy area of P/CVE

Given the security relevance of P/CVE and the political nature of the policy area, it is not surprising that actors in the ‘political system’ have a strong interest in evaluating publicly funded P/CVE measures. In particular, the government is under pressure to justify its actions – both from society at large and from the opposition. Unambiguous statements about effectiveness or success stories are particularly attractive, as they can be used to compare and contrast P/CVE measures and thus inform policymaking (Clément et al., 2021; Martinaitis et al., 2019: 50). In some countries, ministries even transparently state that they commission evaluations in order to make decisions on the continuation of public funding (Rambøll, 2025: 1).

This leads not only to the expectation that P/CVE measures should be evaluated – and thus harbours the potential for isomorphism through coercion regarding the implementation of evaluation – but also to expectations regarding the form these evaluations should take, which opens the door to isomorphism in the form of standardisation of evaluation methods. In the literature, funders are described as being primarily interested in quantifiable results (Clément et al., 2021; Koynova et al., 2022: 14; Martinaitis et al., 2019: 50), placing P/CVE practitioners under a ‘constant pressure to report quick results’ (Junk, 2021). In general, they are said to favour external and summative evaluations and hence those that are conducted at the end of a project by evaluation institutes or researchers to measure effectiveness (Friedrich et al., 2021: 4; cf. Kupiec et al., 2021). These expectations seem to be increasingly prevailing and determining the form of evaluations in the P/CVE field. In their overview of expert assessments, the authors of the international study note that ‘as key initiators and funders of P/CVE evaluations, governments hold significant power over whether and what type of evaluations are conducted’ (Bressan et al., 2024: 6). The European survey suggests such developments, too. For example, one participant noted that P/CVE practitioners need to ‘conduct short term assessment and evaluation based on the needs of funding bodies and at the cost of the sustainable strategic evaluation plan’ (FIN-03).

Walkenhorst’s observation (2019; translated from German) is particularly poignant and warrants more extensive quotation:

Particularly in the field of religiously motivated extremism, it can be observed that, in the wake of increasing political securitisation tendencies in the sectors of radicalisation prevention and deradicalisation, there has been a shift in focus towards the needs and functional logic of politics and political decision-makers, i.e. the supposed control or legitimisation function of evaluations has come to the fore. This means that evaluations are then primarily tasked with providing accountability: has project X achieved effect Y? This is particularly the case when public funds are involved. Accordingly, there is an increase in the number of evaluations commissioned to trace specific effects (or their absence) within a short period of time, explicitly defining this as the focus of interest.

In line with the theoretical ideas described above, this shift in focus could be described as isomorphism through coercion. Structural dependencies create pressure on recipients of public P/CVE funding not only to implement or endure evaluation, but also to generate specific knowledge in the process. This can even extend beyond funding structures to the entire field of practice, as there are also observations that practitioners are increasingly engaging in evaluations themselves to demonstrate the effectiveness of their work to funders or to stand out in a highly competitive field (KN:IX, 2020: 24; Malet, 2021: 64).

Why is this problematic if it also relates to the form of evaluation? P/CVE is a policy area in which the methodological limitations and possibilities of evaluation, as well as its usefulness for further development and learning, are sometimes controversially discussed. A long-standing evaluation research tradition with developed standards and clear expectations for impact evaluation (e.g. DeGEval, 2016) contrasts with a practice that argues its real-world conditions are incompatible with established ‘gold standards’ like randomised controlled trials (RCTs) (Khalil and Zeuthen, 2016: 32; Koynova et al., 2022: 19). It is considered unethical to withhold preventive interventions from control groups, and the complexity of radicalisation and deradicalisation processes makes it difficult to isolate single causal relationships. Practitioners in the policy area often argue that questions like ‘How many people have been deradicalised’ are too simplistic as impact questions. First, ‘deradicalisation’ is a highly individual process, and determining when someone can be considered ‘deradicalised’ is complex. Second, several impact questions are relevant in a deradicalisation project, including reaching the target group, establishing trust with clients, and determining how well cooperation structures with security authorities function. Despite these challenges, it is not impossible to evaluate P/CVE measures’ outcomes. This is evidenced by the growing body of manuals and tools developed over the last few years, such as those by Dawson et al. (2014) or Malkki et al. (2023).

If coercive isomorphism leads not only to evaluation becoming part of P/CVE practice, but also to certain forms of evaluation that correspond to the knowledge interests of the funders – standardisation in the sense of output-focused evaluations with clear key performance indicators – then this can significantly reduce the added value of evaluations, tie up unnecessary resources that might be lacking elsewhere or, in the worst case, lead to a fundamental aversion to evaluation in the social system.

Evaluations may appear to be institutionalised because they are embedded in organisations and funding structures and are considered desirable or necessary, but these evaluations would only produce a certain type of impact findings, while other understandings of impact would be ignored. This can result in projects being considered ‘unsuccessful’ when they are actually quite effective. Some P/CVE practitioners would argue, for example, that reintegrating only some former violent extremists back into society would be a success of an exit programme, despite a high dropout rate.

If practitioners’ expectations towards evaluation in the ‘social system’ do not match those of funders in the ‘political system’, evaluation can become merely a means to an end and lose its potential to initiate learning and improvement processes. In the international survey, a participant from Australia stated that ‘[b]luntly, most [evaluations] are conducted as a condition of funding. Very few are conducted to genuinely learn and improve’ (Bressan et al., 2024: 34). This is consistent with other authors who observe that funders’ demands for external evaluations may even reduce the likelihood that evaluation results foster organisational learning (Friedrich et al., 2021: 4; Lützinger et al., 2020: 615; Mathiesen and Meier, 2021: 46; Williams and Kleinman, 2014). In the worst case, this can even lead to a strong aversion to evaluation: it reinforces the already existing concern that negative evaluation results could threaten the organisation’s existence or that it could become a political pawn (Raets, 2022: 244; Ullrich et al., 2019: 16; van der Vet and Malkki, 2025: 21).

Outlook

Due to the asymmetrical relationship between P/CVE practitioners and funders, it is unlikely that such isomorphism works in reverse, meaning that the form of evaluation increasingly adapts to the informational needs of the ‘social system’. While those of the ‘political system’ are well-documented, it remains unclear whether their informational needs fundamentally diverge from those of practitioners. The latter might prioritise process-related or formative evaluations over outcome-oriented ones, yet it is also possible that the gap in interests and methodological expectations is less pronounced than that. For example, one European survey participant reported of an ‘intrinsic interest for the implementing organisations to find out about the effectiveness of their efforts to self-improve’ (DEU-03). If this is the case, conflicts over evaluation types could be resolved through early dialogue, focusing on the methodological steps required and possible in the respective P/CVE project. The statement by a P/CVE practitioner in the European survey shows that this necessity for tailoring can also be learned the hard way:

There are reasonable and legitimate concerns about how systematised evaluation can undermine P/CVE practice, e.g. one earlier attempt to adopt consistent evaluation tools across a breadth of interventions damaged CSO-government relationships, absorbed considerable time and resources, and produced little benefit. As such, subsequent evaluation attempts have been more tailored to the different types of interventions being evaluated. (GBR-01)

This example illustrates that even after the initial failure of standardised approaches, there is an adaptive capacity that leads to tailored, negotiated evaluation designs. This is a strong indication that the institutionalisation of evaluation in the P/CVE field does not necessarily lead to deterministic symbolic outcomes but can mature into a valuable tool for evidence-based practice through constructive negotiation between systems – a desirable scenario.

Another way to prevent the negative consequences of coercive isomorphism is to promote normative isomorphism instead. This can be achieved by investing in the ‘system of professionalisation’. As noted above, it is still underdeveloped in the policy area of P/CVE. Yet, strengthening this system would not only increase competence in evaluating P/CVE practice and create more sensitivity to the specificities of this policy area but also provide a shared language for dialogue and expectation management between practitioners and funders. An important milestone would be the establishment of P/CVE evaluation standards. By providing a point of reference, these standards would mark a move away from coercive isomorphism towards professionalised norms for evaluation that different stakeholders adhere to when designing tailored evaluations. However, for the policy area of P/CVE, this still remains a long and challenging road.

Conclusion

This article builds on the imbalance and initial evidence of a possible entanglement of the ‘political system’ and the ‘social system’ in the latest contribution to the cross-country measurement of institutionalisation of evaluation by the Globe research project (e.g. Stockmann et al., 2020). It asked: To what extent can the institutionalisation processes in these subsystems influence each other, and what does this mean for the extent and quality of this institutionalisation overall?

The idea was presented that the structural asymmetry resulting from funding structures in certain policy areas, which turns actors in the ‘political system’ into clients and actors in the ‘social system’ into service providers, creates a context in which coercive isomorphism can flourish. The political system uses its dominant position, based on the dependence of actors in the ‘social system’ on funding structures, to enforce evaluation requirements. This is confirmed in the P/CVE area: here, governments predominantly act as initiators of evaluations, which, on the one hand, leads to the mere institutionalisation of evaluation, but also threatens to extend to its form. By favouring output-focused, summative evaluations and simple key performance indicators, donors are forcing a standardisation of evaluation, a shift that primarily serves the interests of political legitimacy and accountability. In the P/CVE field, the latter greatly reduces the added value for organisational learning, as such evaluations are seen as inadequate for further developing measures with complex effects. Even triggering or reinforcing an aversion to evaluation in the ‘social system’ is possible and poses further risks. It seems that the interaction of the subsystems can influence both the extent of institutionalisation – when pressure from the ‘political system’ leads to greater institutionalisation of evaluation in the ‘social system’ – and the quality of this institutionalisation, although the latter appears mainly negative.

Two possibilities to safeguard against this adverse consequence of coercive isomorphism were discussed. First, fostering more early dialogue between practitioners and funders can align potential divergent informational needs regarding evaluation. Second, evaluation in the policy area may transition towards normative institutionalism by strengthening the ‘system of professionalisation’, especially through the formulation of P/CVE evaluation standards.

In order to further scrutinise the arguments made, it is vital that they are applied to other policy areas. Are the potential opportunities and risks typical of P/CVE, or are they shared by other policy areas that have similar structural starting conditions, such as highly competitive funding structures? One possibility would be to examine similar policy areas such as integration and migration work or other complex preventive activities such as crime or even drug prevention. This look at other policy areas should also be motivated by the question of whether the P/CVE field can still learn something from them and be protected from the onset of harmful developments.