Understanding the role of practical knowledge in evidence-based welfare reform—a three-stage model

The paper addresses why and how practical knowledge matters within evidence-based welfare policy. Despite strong governmental commitments to evidence-based policy, implementation often falls short or sparks conflict. Prior literature highlights the need for low policy conflict, a strong policy capacity in administrations, and an adequate supply of credible evidence, while others stress the importance of professional expertise and advocate for evidence-informed practice. The research question guiding the study is: How to understand the role of practical knowledge in evidence-based welfare policies? The study focuses on Denmark’s active labour market policy and public school policy (2010–2022), where practical knowledge is salient but professional power varies. It conceptualizes three knowledge types: political knowledge (judgement and know-how of political actors), scientific-administrative knowledge (systematic analysis and policy capacity to gather and interpret research evidence), and practical knowledge (professionals’ practical wisdom and organizational know-how in implementation). It proposes that the interaction among these forms of knowledge shifts across stages of evidence-based policy, shaping whether evidence is imposed, adapted, or contested. The purpose and importance lie in systematically identifying the changing roles of practical knowledge, showing how public professionals influence evidence uptake through implementation or by re-politicising policy, thereby affecting the success or failure of evidence-based reforms.

The article integrates strands of the evidence-based policy literature often examined separately: political commitment and low conflict as enablers; administrative policy capacity and the availability, relevance, and credibility of evidence; and the role of public professionals’ practical knowledge. Drawing on Head (2008), Jennings and Hall (2012), Howlett (2009, 2015), and others, it links political knowledge and administrative capacity to the adoption of evidence-based policies under low conflict. It situates evidence-based management within a broader NPM agenda, where ministries use evidence to mandate best practices via regulation, benchmarking, and incentives, while acknowledging attribution problems and heterogeneous effects (Paley, 2006; Hall & Jennings, 2008; Weiss et al., 2008). The review underscores that public professionals balance efficiency logics with professional norms, institutional rules, and client needs (Tummers & Bekkers, 2014), which can generate tension between centrally defined evidence and locally grounded expertise. It also highlights professions’ potential collective power to contest policies, reframe debates, and re-politicise evidence use when professional autonomy is threatened (Abbott, 1988; Olson, 1965; MacKillop & Downe, 2023; Boswell, 2009). These insights motivate examining how practical knowledge functions differently across stages of policy processes.

Comparative, explorative case study of two Danish welfare policy domains (active labour market policy and public school policy) from 2010 to 2022. Data sources included: - Document analysis: policy reforms, official ministry documents, research reports (e.g., Amilon et al., 2022; Ministry of Children and Education, 2021; Ministry of Employment, 2023), and ministry knowledge databases/portals and public statements. - Interviews: 17 semi-structured interviews (2022–2023) with ministry officials, stakeholder organizations, and researchers in senior/executive roles, selected to capture diverse perspectives inside and outside ministries. Interviews (~60 minutes) covered availability, relevance, and collection of evidence and its use with local actors. All were audio-recorded, transcribed, and thematically coded in NVivo 12 with a primary focus on political, scientific-administrative, and practical knowledge (with attention to other salient factors).

- Active labour market policy (evidence-based management): The Ministry of Employment built substantial policy capacity in the 2000s–2010s, rating interventions by employment effects and identifying performance gaps. Reforms in the 2010s mandated or incentivized evidence-backed measures (e.g., more caseworker interviews; higher reimbursement for business programs). Compliance was high: total interviews rose 27% (2.2 million in 2011 to 2.8 million in 2019), and average interviews per full-time unemployed increased ~50% (from ~6 to ~9). Business-related programs rose slightly; other activation measures declined. However, aggregate costs did not fall, and gains concentrated among those closest to the labor market; there was limited impact on more disadvantaged groups. In 2022, the government announced plans to shut down job centres and relax central process requirements. - Public school policy (re-politicisation of evidence): Although performance monitoring and testing capacity increased (PISA/PIRLS, national tests), evidence on teaching effects did not map neatly onto performance indicators, and teachers retained legal-professional autonomy. The 2013–2014 national reform (longer school days, 45 minutes of daily physical activity, common learning aims) proceeded amid intense conflict over teachers’ working conditions (lockout and legislated working time), triggering strong union-led resistance and reframing (from a positive holistic school to a negative whole-day school). Implementation was weak; the reform failed to improve student learning, equity, or well-being. - Three-stage model: Stage 1 (central identification of evidence-based policies) occurs under low conflict and with high policy capacity; practical knowledge plays a limited role. Stage 2 (evidence-based management) imposes evidence via rules, benchmarking, incentives; professionals mainly implement, with limited collective resistance. Stage 3 (re-politicisation) features professionals as policy actors leveraging practical knowledge and collective power to contest and reshape evidence use and political agendas. Transition to Stage 3 is more likely where evidence cannot directly govern professional performance and where professions are strong (e.g., teachers).

The findings address the research question by demonstrating that practical knowledge assumes distinct roles across the policy cycle. In labor market policy, professionals’ practical knowledge primarily shaped implementation within an evidence-based management regime, aligning local practice to centrally identified best practices but yielding uneven effects across target groups. In education, strong professional autonomy and collective organization enabled teachers to mobilize practical knowledge to challenge and reframe evidence-based reforms, re-politicising policy and undermining top-down implementation. This dynamic interaction shows that political commitments and administrative capacity do not guarantee sustained evidence-based governance: practical knowledge can both facilitate and impede evidence use, depending on alignment with professional norms, client needs, and institutional autonomy. High-salience conflicts can produce enduring shifts toward more collaborative policymaking to prevent future failures. The model thus clarifies when evidence is likely to be imposed, adapted, or contested, advancing a more nuanced understanding of evidence use in welfare reform.

The study develops a three-stage model explaining how practical knowledge functions within evidence-based welfare reform. Practical knowledge (1) responds to scientific-administrative knowledge during evidence-based management, shaping implementation over extended periods (as in labor policy), and (2) can re-politicise evidence-based policies when powerful professions mobilize collectively (as in education), prompting political reconsideration of reforms. The model contributes a dynamic account of how political, scientific-administrative, and practical knowledge interact across stages of policy, with implications for designing evidence strategies that respect professional autonomy where appropriate and for anticipating conditions under which re-politicisation may occur. It can be applied to other welfare domains to trace stage shifts and to assess whether performance indicators and evidence credibly guide professional practice. Future research could test the model in diverse institutional contexts and examine mechanisms that foster constructive, collaborative evidence use.

The analysis is explorative, based on two cases within the Danish context characterized by comparatively low political conflict and specific professional configurations, which may limit generalizability. Data rely on documentary sources and 17 semi-structured interviews with senior actors; while triangulated, findings reflect these perspectives and available evaluations. The timing and pace of stage transitions may vary across policy areas and countries, and causal attribution of outcomes to evidence-based management versus other factors remains challenging.