Ideas matter greatly to educating marginalized children. Ideas about why they should be educated, as well as ideas about how to educate them effectively. When these ideas lead to changes in government policy and are coupled with specific knowledge about how to implement them, they achieve the greatest impact.

The practice of international development organizations is not always aligned with an effective theory of action for deploying ideas to unleash educational innovation at scale. As the rate of local education innovation accelerates, it becomes ever more urgent that development institutions – if they are to remain relevant – find effective ways to connect ideas about policy and effective practices.

Over the last 65 years, the world has witnessed one of the most remarkable transformations in human history: the universalization of the right to education. This transformation was the result of a social innovation, the Universal Declaration of Human Rights, a charter that in 30 articles spelled out the basic human rights that would lay the foundation for global peace and security, and that committed all nations endorsing the charter to collaborate towards the achievement of those rights as a means to sustainable peace and security.

Much was achieved when advocacy for policy reform was coupled with diffusion of knowledge about how to implement programmes and the practices aligned with them. Less was accomplished when international development institutions remained exclusively focused on policy advocacy or exclusively focused on direct provision of programmes and services. Avoiding the pitfalls of ineffective efforts requires clear thinking about how ideas about what works translate into changes in practice at scale.

The system of United Nations organizations advanced the right to education, to a great extent as a result of the strategic deployment of advocacy for policy changes and ideas about means to achieve them. For example, in the 1950s, the United Nations Educational, Scientific and Cultural Organization (UNESCO) convened regional meetings of Ministers of Education and Finance to make the case for universal primary education. Those meetings served not only to broker political agreements, but also to exchange know-how about achieving those goals. The intellectual repertoire that helped advance the goal of education inclusion comprised double shifts in schools; multi-graded schools to provide a full cycle of primary education in rural areas with one teacher or a few teachers; and schools in clusters to support professional development and share resources. The diffusion of these practices, coupled with the development of capacity in educational planning, school-mapping and other methodologies contributed much to the remarkable expansion in access to education that took place in the developing world in the 1950s and 1960s.

In contrast, when policy frameworks and aspirations, however promising, have not been coupled with knowledge about ‘what works’ in implementing those aspirations, their impact has remained elusive. The Delors Report, ‘Learning the Treasure Within’, a visionary document commissioned by UNESCO that resulted from a thorough global consultation on the goals and purposes of education in the twenty-first century, remains, two decades after it was published, largely aspirational because it was not accompanied by specific knowledge about how to execute on such an expansive set of school purposes.

In response to the challenges of translating policy reform into changes in practice, some development organizations engage in executing these changes themselves, through ‘pilot’ programmes or other implementation mechanisms. The risk in these efforts of engineering innovative programmes is that they may never actually connect with policy, and therefore fail to reach the necessary scale to serve the children who need these services.

From the outset, the international education organizations that emerged as part of the post-world war order to promote development aimed to influence education policy and leverage government resources, not to substitute for those. The goal of designing, identifying and disseminating good education practices has been not only about supporting their replication, but also about deploying them strategically to change the policy and programming status quo so they can, at scale, support greater educational opportunities that lead to equity.

“The risk in efforts of engineering innovative programmes is that they may never actually connect with policy, and therefore fail to reach the necessary scale to serve the children who need these services.”

For instance, beginning in the 1980s, international organizations, including UNICEF, began to advocate for greater learning opportunities in early childhood education. Until then, for most of the world, the right to education was understood to begin at the age of elementary school. The Jomtien Declaration of Education for All, adopted at the World Conference for Education for All, included early childhood education as one of the targets of renewed efforts to promote universal education. These efforts included the documentation and diffusion of programmatic approaches to expand early childhood education, and the capacity development of key stakeholders to implement them.

As a result, many governments expanded early childhood education opportunities. In Mexico, for example, in 2002, Congress modified the education law to guarantee three years of preschool for every child. Similar policies expanding early childhood education have brought clear gains in opportunities for disadvantaged groups. In this case, institutions such as UNICEF or UNESCO have played the role of knowledge brokers – disseminating evidence to show how much development takes place in the early years of life, as well as studies of programmes that work to advance early childhood – and policy advocates. They have made the case for the moral imperative to provide public support for policies and programmes aligned with the broader aspirations in the Universal Declaration of Human Rights, the Convention of the Rights of the Child and similar international legal frameworks.

Advocacy was coupled with the knowledge of good practice as international development institutions played a key role in identifying, documenting and evaluating programmes. UNICEF’s Office of Research, for instance, has produced multiple examples of this kind of work, such as a study of Mexico’s programmes in early childhood education. Such studies helped spur governments’ commitment to expand early childhood education.

Education progress is thus the result not only of advocacy or moral arguments for inclusion, but also of effective use of knowledge about ‘what works’, which can frame the opportunities for educational inclusion into a tractable problem that government leaders can address within institutional resources and political constraints that shape government action. In this sense, educational inclusion has been the result of educational innovation – getting governments to take on a new challenge and to undertake new actions to address it.

Such innovation does not typically take place with governments tackling a problem as if they were the first group of people on the planet to do so, but rather it is inspired, informed and supported by previous or similar attempts elsewhere. Much of the legitimacy of the international development institutions rests on their perceived ability to act as knowledge brokers and on their expertise to manage the process that translates such knowledge into programmes that can achieve the scale desired by governments in sustainable ways.

Educational good practices have been ‘transferred’ from one context to another, and significant progress has been made in recognizing that not all ideas and practices ‘transfer well’. Accordingly, international development institutions have advanced from merely documenting and diffusing ‘good practices’ to creating communities of practice, where policymakers and practitioners can devote time to understand such practices in their context and figure out the necessary adaptations to transfer them.

In diffusing education practices for adoption in diverse cultural settings, a frequent problem is determining which elements were context-dependent and what kind of adaptations should be made in transferring educational innovations across contexts. Many efforts support the transfer of curricular approaches from one setting to another, for example a curriculum to support instruction in literacy, science or math. Given different cultural conditions, institutional settings or levels of teacher capacity, the same curriculum often produces very different results. For instance, outcome-based education in South Africa failed to produce in schools serving disadvantaged children the much expected improvements in instructional quality that education leaders had hoped would help consolidate the political gains of the democratic transition in that country.

To overcome these kinds of challenges with the transfer of educational innovation, some colleagues and I developed an approach to systematically adapt innovations to new contexts. This approach defines transfer as a process of re-design and re-invention, rather than simple adoption, articulating the theory of action of the innovation in such a way that it makes the transfer susceptible to systematic evaluation.

A challenge of the conventional approach to identifying and disseminating good education practices is how to capture the growing range of social innovation that is taking place globally. Global knowledge networks become stale and irrelevant if they can circulate only a limited set of ideas about what works to make education more inclusive. The networks of international development institutions need to become integrated with local and community networks, to be open to the innovation happening locally. This is especially important because the rate of local innovation is accelerating exponentially.

We live at a time of unprecedented opportunities for educational innovation, generated in part by the increasing levels of education of the world’s population and the resulting increase in the number of individuals who have been empowered to be producers of educational innovation. Global educational expansion has increased the number of people with the mindset, ambitions and skills to bring about innovations to tackle ambitious goals. Education entrepreneurs benefit also from new technologies, availability of capital, regulatory frameworks that make it easier for ordinary people – often in small groups – to take on challenges in the past reserved for governments and large international or national organizations. As a result of their efforts, the number of educational innovations is growing exponentially, everywhere. In many places, teachers, students, and ordinary citizens organized in grass-roots non-governmental organizations generate promising innovations to support instructional improvement that enhances the contributions of education to social inclusion.

Some authors have argued that there is great promise in the study of these grass-roots innovations, cases of success against the odds, because identifying approaches that have been proven to work in particular social and geographic settings avoids the challenges of ‘transferring’ ideas developed elsewhere and trying to get them to take root in new settings. By definition, these innovations are already sustainable and adaptive, as they emerged in the very contexts to which they should now be scaled. This concept of ‘positive deviant’ was first developed in nutrition research, as researchers found that even in very poor communities some families had well-nourished children.

The study of positive deviants has been successfully applied to study local innovations in a range of domains, including education. The approach was systematized by Richard Pascale, together with Jerry and Monique Sternin, and later popularized by Malcolm Gladwell in his book Outliers. There is great promise in the study of positive deviants as a way to learn from and capture the rich experience of innovation that takes place in many communities. In many ways, this approach is contrary to the traditional view of development as the result of the transfer of ideas international experts bring, often from abroad, about how to improve the inclusion, efficacy or relevancy of educational institutions.

These two approaches – learning from global experience as well as from grass-roots practices – are complementary rather than antithetical. In an era of globalization, educational innovation is ‘glocal’, rather than purely local or global. For example, in an effort to advance educational inclusion of students with special needs in Brazil, the social entrepreneur Rodrigo Hubner Mendez created an organization to advocate for policy reform and disseminate good practices. This initiative, the Diversa project of the Instituto Rodrigo Mendez, identifies good practices of inclusion in Brazil and other countries, and conducts detailed case studies that are then disseminated through various means, including a comprehensive website as well as periodic meetings of municipal secretaries of education, in which Mendez partners with the Brazilian Ministry of Education to build the capacity of education leaders to replicate or adapt such practices. UNICEF has been a partner of these efforts by a civil society organization to blend the study and dissemination of good local and global practices of inclusion.

There are many similar efforts globally to support educational innovation. For example, Edutopia sustains a rich portal designed to disseminate good practices in education for socio-emotional development, civic education, student-centred education and effective use of technology. Edutopia uses also various social platforms to sustain professional learning communities that can collaborate in finding ways to adapt those ideas to particular contexts, including the use of Twitter to disseminate ideas to inform educational innovation.

In addition to the study and dissemination of good education practices, whether these are global practices or local ‘positive deviants’, and in addition to the different professional learning communities, which can be structured to facilitate an appropriate transfer of those practices across contexts, educational inclusion could be accelerated by the promotion of innovation through improvement networks.

For instance, the Carnegie Foundation for the Advancement of Teaching, under the leadership of Anthony Bryk, is using improvement science – an approach first developed by Don Berwick to improve medical practice in hospitals – to support and study innovation networks focused on specific problems of practice. The Carnegie Foundation supports two improvement networks, the Building a Teaching Effectiveness Network and the Community College Pathways Program. The Community College Pathways Program has focused on resolving the failure of remedial education courses in community colleges, where many disadvantaged students are placed upon entry in community college – and which has shown very poor results in helping students transition to mainstream instruction and eventually graduate from college.

The improvement networks supported by the Carnegie Foundation involved educators working in community colleges in remedial mathematics, building on research on the deficiencies of these courses and on evaluations of experiments to overcome some of these deficiencies. These improvement networks commit to the goal of systematically implementing changes that follow an explicit and shared theory of action, based on the best available evidence. They then measure the results of these efforts, creating a cycle of rapid learning where each implementation is, in effect, an experiment that allows all institutions participating in the network the opportunity to learn from all these concurrent efforts.

To accelerate the process of learning from innovations in educating marginalized children that were designed and identified in different parts of the globe, an improvement network should develop a commitment to sharing this knowledge seamlessly across its members. How to get every node in a network to access the knowledge available to others is indeed a challenge for complex organizations. At the core of this challenge are the traditional silos and forms of organization that impede fluid communication across departments, and the fact that much of the knowledge gained by the participants in the network is tacit, and never formalized into protocols that can be accessed by others. Prior to the availability of the Internet these challenges were intractable for global organizations. This is no longer the case: The challenges to be overcome now have to do with organizational culture, capacity and mindset, but especially with how we think about the way in which ideas about ‘what works’ actually translate into practices in educational institutions as they support all children to learn.