Noam Angrist and colleagues set out a research agenda designed to make implementation itself an object of rigorous study. Their central claim is deceptively simple: the frontier of education research and practice is no longer just discovering what works, but understanding how to make it work repeatedly, equitably and at increasing scale. The authors write that there is growing evidence on what works, and that the task now is to translate that evidence into action and to study that translation systematically, so that effective programs can be delivered again and again to achieve what they describe as effective, equitable and enduring learning for all. In their framing, implementation is not an afterthought to research but a scientific discipline in its own right.

The scale of the problem gives the argument its urgency. Although most children around the globe now enroll in school, a large share of them complete years of schooling without mastering basic literacy and numeracy. Decades of randomized trials and rigorous evaluations have produced genuine breakthroughs: teaching at the right level, structured lesson plans, and cost-effective tutoring programs have all shown strong effects in carefully managed studies. Yet when governments adopt these interventions and roll them out across thousands of schools, effects often shrink dramatically or disappear altogether. The authors point to a diagnostic blind spot that makes this pattern hard to interpret: education research rarely measures how faithfully a program is implemented or how widely it is adopted, so it is often impossible to tell whether disappointing results reflect a flawed program or a flawed rollout.

This distinction between program failure and implementation failure is the conceptual heart of the new agenda. A curriculum reform that succeeds when delivered by a motivated NGO with intensive training and supervision may falter when a ministry of education must train tens of thousands of teachers, manage supply chains, monitor compliance and sustain funding across political cycles. Without data on fidelity, dose, reach and adaptation, evaluators cannot distinguish between a program that genuinely does not scale and a program that was scaled badly. The authors argue that this measurement gap has distorted the field’s accumulated evidence base, leading policymakers to abandon promising interventions prematurely or, conversely, to scale up programs whose real-world performance was never properly understood.

To address this, Angrist and colleagues call for stronger implementation science in global education, drawing explicitly on lessons from health care. Over the past two decades, implementation science has become an established field in global health, where researchers study the barriers and facilitators to delivering interventions such as vaccination campaigns, HIV treatment and maternal health programs. Health researchers have developed validated frameworks for measuring fidelity, adoption, penetration and sustainability, and they routinely build implementation questions into trial design from the outset. Education, the authors contend, has the scientific foundations to do the same: a mature experimental toolkit, growing administrative data systems and an expanding corps of researchers embedded within government systems. What has been missing is a shared framework that treats the journey from evidence to practice as a staged, studyable process.

That framework is the paper’s most concrete contribution. The authors propose a four-stage pathway for turning educational research into effective, equitable practice at scale: efficacy trials, effectiveness trials, policy plans and large-scale implementation. In the first stage, interventions are tested under tightly controlled conditions to establish whether they can work at all. In the second, they are evaluated in real-world settings across diverse populations, schools and implementers to establish whether they do work outside ideal circumstances. The third stage involves translating trial evidence into concrete policy plans that account for cost, political feasibility and delivery systems. The fourth stage is full-scale implementation, where the intervention meets the realities of national education systems, and where, crucially, researchers continue to measure, learn and adapt.

Each stage asks different questions, and the authors emphasize that the later stages have been chronically understudied. Efficacy and effectiveness trials are now relatively common in education research, but the transition from trial to policy and from policy to system-wide delivery remains poorly documented and poorly theorized. The framework insists that scalability is not a binary property of a program but an empirical question that must be tested: does the intervention work across diverse populations and real-world settings, and is it affordable and sustainable when delivered through ordinary government channels? Cost-effectiveness, the authors argue, deserves a central place in this calculus, because an intervention that produces modest learning gains at very low cost may deliver more total benefit at scale than a more powerful but prohibitively expensive alternative.

The equity dimension of the framework is equally deliberate. Programs that succeed on average can mask deep disparities, benefiting children in well-resourced urban schools while leaving behind those in remote, underfunded or crisis-affected settings. By embedding equity into every stage, from trial sampling to implementation monitoring, the agenda aims to ensure that scale-up narrows rather than widens learning gaps. The authors also highlight durability: education systems are subject to political turnover, budget shocks and shifting donor priorities, and interventions that depend on continuous external support rarely survive. Implementation science, in their view, should identify the design features and system conditions that allow gains to endure after the researchers and funders move on.

The authors are candid about the cultural shift their agenda requires. Education economists, psychologists and evaluators have built impressive evidence on what works, but the incentives of academic research reward novel findings from clean trials rather than the slower, messier work of studying delivery. Governments, meanwhile, often lack the time, data and technical capacity to diagnose implementation failures before scaling decisions are made. Bridging this divide means embedding researchers within ministries, building measurement of implementation into routine administrative data, and designing evaluations that answer the questions implementers actually face. The payoff, the authors argue, is a virtuous cycle in which more science makes it into implementation and more implementation is made scientific.

Ultimately, the agenda is a bet that the next large gains in global learning will come not from new interventions but from better delivery of known ones. If the field can systematically study how programs are adopted, adapted, funded and sustained, the growing library of proven interventions could finally translate into learning gains for the hundreds of millions of children who currently sit in classrooms without acquiring basic skills. As the authors put it, implementation science can support the ultimate goal of all children learning. The framework they propose offers the field a roadmap for getting there, one carefully studied stage at a time.

Subject of Research: Implementation science framework for scaling effective education programs from research trials to government delivery

Article Title: A research framework for turning education evidence into practice

Article References: A research framework for turning education evidence into practice. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: implementation science, global education, education policy, learning outcomes, randomized trials, scale-up, education research, teaching at the right level, cost-effectiveness, equity, government school systems, evidence-based practice

The research, published in the Journal of New Approaches in Educational Research, was conducted by a team at the University of Zaragoza that surveyed 1,369 families whose children attend public and state-assisted schools across different Spanish regions. Rather than examining leadership, family involvement, competences and grades in isolation, the researchers built a Structural Equation Model, a sophisticated statistical framework that allows scientists to test how multiple latent constructs, variables that cannot be observed directly, relate to one another simultaneously. The four constructs at the heart of the model were democratic educational leadership, family accompaniment, students’ transversal competences and academic performance as perceived by families.

The leadership construct was deliberately framed in democratic terms. Instead of measuring organisational efficiency or instructional supervision alone, the study captured a relational style of leadership grounded in participation, trust, empathy and transparent communication. Families rated school leaders on indicators such as the ability to communicate confidently, motivate staff, foster a positive climate of coexistence, empathise with the needs of families, teachers and students, and address difficult situations with openness and positivity. This operationalisation reflects a body of theory that distinguishes democratic leadership from transformational, instructional and distributed models, emphasising equity of participation and the active involvement of all stakeholders, including families, in school decision-making.

Family accompaniment, the second pillar of the model, went well beyond conventional notions of parental involvement. It encompassed supporting children’s autonomous learning, attending meetings and tutorials with teachers, encouraging the development of competences, fostering positive attitudes towards learning and promoting study skills at home. Transversal competences, often described as twenty-first-century skills, covered collaboration, autonomy, leadership, technology use, creativity, critical thinking, communication and adaptability. Academic performance, finally, was measured through families’ perceptions of students’ overall achievement, evaluations, marks and progress at school. All indicators were scored on an eleven-point Likert scale running from zero to ten, a format the researchers chose for its sensitivity and its familiarity to families accustomed to standard grading systems.

The technical rigour of the study is considerable. Before testing their theoretical model, the researchers ran confirmatory factor analyses to verify that the measurement structure held together. The combined four-factor model showed strong fit to the data, with a comparative fit index of 0.96, a root mean square error of approximation of 0.05 and a standardised root mean square residual of just 0.03. Internal consistency was high, with Cronbach’s alpha values ranging from 0.84 for family accompaniment to 0.98 for educational leadership. The team also applied robust maximum likelihood estimation with Satorra-Bentler corrections to guard against non-normality in the data, and compared their four-factor model against a single-factor alternative to rule out common-method variance, a statistical artefact that can arise when all data come from the same informants. The multidimensional model fit far better, supporting the validity of the findings.

When the structural model itself was tested, the results were unambiguous. Educational leadership showed a strong positive effect on students’ academic performance, with a standardised coefficient of 0.54, and an even stronger effect on transversal competences, at 0.73, both statistically significant at the highest threshold. Leadership also predicted family accompaniment, with a coefficient of 0.37, confirming that when school leaders communicate openly and build trust, families become more engaged in their children’s education. In turn, family accompaniment predicted both academic performance and competence development, with coefficients of 0.17 and 0.11 respectively. All five hypotheses of the study were supported by the data.

Perhaps the most intriguing finding lay in the mediation analysis. The researchers demonstrated that family accompaniment significantly mediated the relationship between educational leadership and both outcome variables. In practical terms, this means that part of a school leader’s influence on student success travels through the family: leadership practices that lower barriers between schools and homes create formal and informal opportunities for family involvement, which then reinforces motivation, study habits and engagement in children. Although the indirect effects were modest in magnitude, their statistical significance suggests that even small relational improvements, accumulated across years of school-family interaction, can meaningfully shape educational trajectories. This positions democratic leadership not merely as an administrative function but as a relational process that connects classroom practice with family life.

The descriptive data add nuance to this headline result. Families rated leadership most highly for fostering a positive climate and confident communication, but gave lower marks for empathy towards family needs and for the direct, honest transmission of problems to staff, suggesting a gap in the more demanding empathic dimensions of leadership. Family involvement scored highest for attending meetings with teachers and encouraging positive attitudes, while promoting study habits at home lagged behind. Among transversal competences, autonomy, collaboration and technology skills scored well, but leadership, critical thinking and creativity, the higher-order skills most valued in a rapidly changing economy, remained comparatively weak. Older family members perceived leadership and performance less favourably, while university-educated families reported more positive views of their own accompaniment, and perceptions of competence development dipped in larger schools.

The authors are careful about the limits of their evidence. The cross-sectional design captures associations at a single point in time and cannot establish causation, and the convenience sampling strategy, with voluntary participation, may have attracted more engaged families, potentially inflating positive perceptions. All variables were family-reported rather than drawn from objective academic records, and the study was conducted exclusively in the Spanish context, where cultural expectations about school-family relations may differ from those elsewhere. Yet the sample’s size and heterogeneity, the robustness of the statistical procedures and the consistency of the findings with decades of international research on leadership and parental involvement lend the conclusions considerable weight.

The practical implications are hard to ignore. For policymakers, the study argues for leadership training programmes focused on communication, participation and community engagement rather than managerial technique alone. For school leaders, it recommends creating structured opportunities for family participation, improving communication channels and offering concrete guidance to parents on how to support learning at home. And for families, the message is that their accompaniment is not a peripheral supplement to schooling but an active mechanism through which school quality translates into student achievement. In an era when education systems worldwide are searching for levers to build the skills the future demands, this research suggests that the most powerful lever may be the connection itself, the deliberate, empathetic bridge between those who lead schools and those who raise children.

Beyond its immediate findings, the study sits within a broader international conversation about how school quality should be defined. Global frameworks, including the United Nations Sustainable Development Goals, have pushed education systems to look beyond narrow achievement metrics and to consider whether schools cultivate adaptable, critically minded citizens. Transversal competences occupy a central place in this agenda because they are the capacities that transfer across subjects, jobs and life circumstances. By modelling competence development alongside academic performance, the Spanish research team aligned their work with this wider shift in how educational success is conceptualised and measured.

The mediating role of family accompaniment also carries theoretical weight. Leadership research has long debated whether the influence of school leaders on students is direct or indirect, with several influential reviews suggesting that leaders affect learning primarily by shaping conditions around teaching, such as climate, collaboration and community ties. The present findings fit that indirect-effects perspective: leadership appears to matter partly because it opens relational channels that draw families into the educational process. This reframes family engagement not as an optional add-on but as a mechanism that leadership actively activates or neglects.

The descriptive patterns deserve attention from practitioners. The weaker scores for empathy towards family needs and for candid communication about problems hint that schools may excel at routine, pleasant interactions while struggling with the harder relational work of acknowledging difficulties. Similarly, the lag in home-based study habits suggests that families readily attend formal meetings but find it harder to sustain day-to-day learning support, an area where targeted guidance from schools could prove especially valuable.

Future research could build on these results in several directions. Longitudinal designs would help establish whether leadership practices precede family engagement over time, strengthening causal claims that cross-sectional data cannot deliver. Objective measures of achievement, drawn from school records or standardised assessments, would complement family perceptions. Comparative studies across countries and cultures could test whether the leadership-family pathway holds where expectations about parental roles differ. Qualitative work might also illuminate the concrete interactions, a conversation at pickup, a well-timed phone call, through which democratic leadership translates into family action.

For now, the study’s central contribution is its integrated view. Leadership, families, competences and achievement are not separate levers to be pulled independently; they form a connected system in which the quality of relationships between schools and homes helps determine how effectively everything else works.

Subject of Research: The relationship between democratic educational leadership, family accompaniment, transversal competences and academic performance in school students.

Article Title: Educational leadership for improving transversal competences and academic performance: the mediator role of family accompaniment

Article References: Latorre-Cosculluela, C., Mairal-Llebot, M., Vázquez-Toledo, S., & Rivera-Torres, P. (2026). Educational leadership for improving transversal competences and academic performance: the mediator role of family accompaniment. Journal of New Approaches in Educational Research, 15(1), Article 21. https://doi.org/10.1007/s44322-026-00071-2

Image Credits: AI Generated

DOI: 10.1007/s44322-026-00071-2

Keywords: educational leadership, family accompaniment, transversal competences, academic performance, structural equation modelling, democratic leadership, parental involvement, school improvement, 21st-century skills, education research, school-family collaboration, student outcomes