Millions Learning: Scaling up Quality Education in Developing Countries Summary Summary Executive the Developing World

MILLIONS LEARNING SCALING UP QUALITY EDUCATION IN DEVELOPING COUNTRIES

Jenny Perlman Robinson and Rebecca Winthrop with Eileen McGivney 1 MILLIONS LEARNING SCALING UP QUALITY EDUCATION IN DEVELOPING COUNTRIES

Jenny Perlman Robinson and Rebecca Winthrop with Eileen McGivney Table of contents

EXECUTIVE SUMMARY 8

FILIPE’S STORY: BRAZIL’S NEW APPROACH TO SCHOOLING IN THE 14 I This report was written by Jenny Perlman Robinson and Rebecca Winthrop with Eileen AMAZON JUNGLE McGivney, along with research support from Jenny Alexander and Priyanka Varma, as well as invaluable assistance over the past few years from the Millions Learning team at the Brookings II ACCELERATED EDUCATION PROGRESS IS URGENTLY NEEDED 20 Institution’s Center for Universal Education, particularly: Lulwah Ayyoub, Mia Blakstad, Zoe Global goals place a premium on learning for all 22 Norris, Priya Shankar, and Samantha Spilka. Where education has failed: The scope and scale of challenges 24 Why focus on learning? 27 Sincere gratitude and appreciation to our Millions Learning consultants who have been equally instrumental in the report: Priscila Cruz, Shushmita Chatterji Dutt, Flavia Goulart, III SCALING SO MILLIONS LEARN: DEFINING A GLOBAL PRIORITY 30 Mayyada Abu Jaber, Haeyeon Jung, Christina Kwauk, Alessia Lefebure, Divya Mansukhani, A focus on scaling 33 Maria May, Jessica Moore, Ainan Nuran, Daniela Petrova, Martin Roeck, Marijke Schouten, Sheikh Islam Tanjeb, Judith-Ann Walker, and Jillian Yoerges. Addressing an evidence gap 33 Defining scaling 35 We are thankful for the guidance and support from our Millions Learning International Methods guiding Millions Learning 39 Advisory Panel Co-Chairs, the Honorable Julia Gillard (The Brookings Institution) and Homi Kharas (The Brookings Institution), and members: Alice Albright (Global Partnership IV FINDINGS: HOW HAS SCALING HAPPENED? 44 for Education), Aaron Benavot (Education for All Global Monitoring Report), Madhav Chavan Scaling success from the margins 46 (Pratham), Larry Cooley (Management Systems International), Claudia Costin (The World Bank Idea adoption and delivery innovation: 50 Group), Elizabeth King (The Brookings Institution), Ruth Levine (The William and Flora Hewlett Two strategies for scaling quality learning from the margins Foundation), Santiago Levy (Inter-American Development Bank), Daniel Low-Beer (World Health Organization), Lu Mai (China Development Research Foundation), Zbigniew Marciniak (University Scaling learning: 14 core ingredients 51 of Warsaw), Shannon May (Bridge International Academies), Dzingai Mutumbuka (Association for the Development of Education in Africa), Vineet Nayar (Sampark Foundation), Tamela Noboa V DESIGN 54 (Discovery Learning Alliance), Liesbet Steer (International Commission on Financing Global 1. Local education needs: Interventions should be designed in response 57 Education Opportunity), and Mark Surman (Mozilla Foundation). to local demand and should ensure the participation of end-users. 2. Cost-effective learning: Cost structures affordable at scale should 64 The Brookings Institution is a nonprofit organization devoted to independent research and be incorporated in the design. policy solutions. Its mission is to conduct high-quality, independent research and, based on that 3. Flexible adaptation: Core elements of effective learning 69 research, to provide innovative, practical recommendations for policymakers and the public. The approaches should be identified and replicated across contexts conclusions and recommendations of any Brookings publication are solely those of its author(s) while adapting the rest to local circumstances. and do not reflect the views of the Institution, its management, or its other scholars. 4. Elevating teachers: Community expertise should be leveraged to 72 support and unburden teachers. Support for this publication and research effort was generously provided by the John D. and Catherine T. MacArthur Foundation and The MasterCard Foundation. The authors also wish to VI DELIVERY 80 acknowledge the broader programmatic support of the William and Flora Hewlett Foundation, the LEGO Foundation, and the Government of Norway. 5. Education alliances: All actors need to work together to achieve a 83 common goal. Brookings recognizes that the value it provides is in its absolute commitment to quality, 6. Learning champions and leaders: As scaling quality learning is 88 independence, and impact. Activities supported by its donors reflect this commitment, and the a political and technical exercise, champions within and outside analysis and recommendations are not determined or influenced by any donation. government and the classroom are crucial. 7. Technological advances: Context-appropriate technologies can 93 LIST OF FIGURES accelerate education progress. Figure 1. A scaling success story: The rapid spread of schooling in 23 8. Windows of opportunity: Effective education approaches are 95 the past 200 years more likely to take root and spread when they align with country Figure 2. The learning crisis: 38% of children not learning basic 25 priorities. literacy and numeracy 9. Better data: Data on learning and scaling play a central role by 98 Figure 3. Reaching universal secondary education: Projected gap 25 motivating informed action at the policy and practice levels. Figure 4. Pathways to scale 36 Figure 5. Case study overview 40 VII FINANCE 102 10. Flexible education financing: Financing should be flexible, 105 Figure 6. Learning Metrics Task Force: Seven domains of learning 42 including to build core operational capacity. Figure 7. Scaling success from the margins 48 11. Long-term education financing:Stable and predictable support is 108 Figure 8. Scaling learning: 14 core ingredients 52 essential. Figure 9. Millions Learning recommendations 131 12. “Middle phase” financing: Financing is required to bridge the 110 critical stage between pilot and broad uptake. LIST OF BOXES Box 1. Pathways for scaling effective approaches 38 VIII ENABLING ENVIRONMENT 114 Box 2. Defining the scope of the study 42 13. Supportive policy environment: Government policy must safeguard 117 Box 3. Lesson Study, Zambia 47 every child’s right to a quality education while remaining open to a Box 4. Sistema de Aprendizaje Tutorial 50 diversity of ideas and actors to contribute to this common aim. Box 5. Bridge International Academies 58 14. A culture of R&D: Ensuring that more children learn requires a 122 strong ethos of experimentation, collecting learning data, and using Box 6. Fundación Escuela Nueva 59 it for continuous improvement. Box 7. BRAC, Non-Formal Primary Education 60 Box 8. Aflatoun International 61 IX TOWARD THE FUTURE: THE NEED FOR INCLUSIVE AND ADAPTIVE 126 Box 9. Room to Read, Literacy Program 62 EDUCATION ECOSYSTEMS IN A CHANGING WORLD Box 10. Pratham, Read India 67 Action 1: Develop a culture of R&D in education 132 Box 11. Educate! 68 Action 2: Share new ideas through a network of Idea Hubs 133 Box 12. Sesame Workshop, Sesame Street 70 Action 3: Activate talent and expertise outside the classroom 134 Box 13. Teach For All 71 Action 4: Fund the middle phase 135 Box 14. INJAZ, Jordan 74 Action 5: Measure and learn what works through better learning and 136 Box 15. Amazonas State Government’s Media Center 76 scaling data Box 16. Worldreader 86 ENDNOTES 138 Box 17. Government action to create space for scaling quality 121 learning BIBLIOGRAPHY 148

ANNEXES Annex 1. Millions Learning case study selection process 155 Annex 2. 14 Millions Learning case studies 158 Annex 3. Acknowledgments 160 8 Millions Learning: Scaling up quality education in developing countries summary summary Executive the developing world. Thestory emerges and how qualityeducationhas scaledin Millions Learning demanded by thelabormarket. problem solving,whichare increasingly such ascriticalthinkingandcollaborative young peopledevelop 21st-century skills, around theworld are struggling tohelp numeracy skills.Additionally, countries country—lack themost basicliteracy and least four years of schoolinadeveloping globe—many of them havingspent at Today, millionchildren 250 around the sector, this gap isnotprojectedtoclose. proceeds asusual intheeducation and developing countries, and if business between educationlevels indeveloped An estimated “100-year gap” persists thinking skills. academic content and higher-order has beenmetwithlittlemastery of core to scale” stories, thisexpansion toooften one of themost widelysuccessful “going of schoolingover thepastyears 150 is Whilethespread committed tosupport. which193Goals, countries have recently United Nations Sustainable Development progress, anditisalsoenshrinedinthe education isatthecenter of anation’s for theirchildren andyouth. Quality with how toscalequalityeducation Around theworld, countries are grappling tellsthestory of where effective newapproaches toimproving children andyouth. With theformer, to deliver educationtomarginalized innovation an educationecosystem, and the spread of newapproaches across two mainways: Scaling from themargins occurs in limited. children whose educationaloptionsare new ways of reaching marginalized a community movement thatdevelops try anewapproach. For others, itinvolves freedom toitsofficials withinadistrict to a flexible central government giving a case-by-case basis.For some, itmeans What constitutes themargins varies on to reach many more children andyouth. grow onthemargins andthenspread and ideasare allowed todevelop and often occurs whennewapproaches that successful scalingof qualitylearning many placesaround theworld. We found learning ishappeningatlarge scalein transformational changeinchildren’s of New Delhi totherainforest inBrazil, What we found isthatfrom theslums India. Uganda andZambia, toJordan and studies, from Brazil andHonduras, to and learning,including14in-depthcase from wide-ranging research onscaling , thedevelopment of newways idea adoption , namely delivery countries. developed andgrown withinandacross alternative educationprograms—have such as distance learningmodelsor for themost marginalized communities— latter, neweducationdelivery approaches adopted by different actors. With the across educationecosystems andbeen to teacherdevelopment—have spread process—from curriculum, to materials, components of theteaching andlearning 4 3 2 1 unburden teachers. Community expertise shouldbeleveraged and tosupport ELEVATING TEACHERS: to localcircumstances. identified and replicated across contexts while adaptingthe rest Core elements of effective learningapproaches shouldbe FLEXIBLE ADAPTATION: the design. Cost structures affordable atscaleshouldbeincorporated in COST-EFFECTIVE LEARNING: and shouldensure of theparticipation end-users. Interventions shouldbedesignedinresponse tolocaldemand LOCAL EDUCATION NEEDS: do thisare: planning for scale from theoutset. Ingredients necessary to Improving learningatscale starts withcommitted leaders DESIGN scaling of qualityeducation. delivering, financing,andenabling include essential elements for designing, in thebroader scalingliterature. They is frequently reinforced from evidence that improve learning.Theirimportance central for scalingeffective approaches learning. Each of theseingredients is the context, contribute toscalingquality different combinationsdependingon We identified 14core ingredients, in

9 10 Millions Learning: Scaling up quality education in developing countries 10 12 11 9 6 8 5 7 amounts. Ingredients neededfor thisare: How resources are allocated matters asmuchabsolute FINANCE informed atthepolicyandpractice levels. action Data onlearningandscalingplayacentral role by motivating DATA:BETTER spread whenthey alignwithcountry priorities. Effective educationapproaches are more likely totake root and WINDOWS OFOPPORTUNITY: progress. Context-appropriate technologiescanaccelerate education TECHNOLOGICAL ADVANCES: crucial. champions withinandoutsidegovernment andtheclassroom are As scalingqualitylearningisapoliticalandtechnicalexercise, LEARNING CHAMPIONSANDLEADERS: All actors needtowork togethertoachieve acommongoal. EDUCATION ALLIANCES: needed for thisare: combination of technical andpolitical actions. Ingredients or delivering atlarge scale isessential. This involves a Attention to theoperational realities of implementing DELIVERY and broad uptake. Financing isrequired tobridgethecriticalstage between pilot “MIDDLE PHASE” FINANCING: isessential.Stable andpredictable support LONG-TERM EDUCATION FINANCING: capacity. Financing shouldbeflexible, includingtobuildcore operational FLEXIBLE EDUCATION FINANCING:

environment where allactors can but they must alsoactively foster an every child’s right toaquality education is ittheirresponsibility todeliver on pivotal role inthisecosystem. Not only improve learning.Governments playa ideas or approaches that most effectively actively helpfacilitate thespread of new experimentation tothrive, andthen provide spacefor innovation and calls for educationecosystems that adaptive educationecosystems. This move toanewnormof inclusive and Scaling qualitylearningrequires a 14 13 continuous improvement. experimentation, learningdata,andusingitfor collecting Ensuring thatmore children learnrequires astrong ethos of A CULTURE OFR&D: and actorstocontribute tothiscommonaim. quality educationwhileremaining opentoadiversity of ideas Government policymust safeguard every child’s right toa SUPPORTIVE POLICY ENVIRONMENT: the scaling process. Ingredients neededfor thisare: policies operate plays acritical role infacilitating orimpeding national to local, theenvironment inwhichprograms or happen inavacuum. Largely guided by governments from As critical asthese three other aspectsare, scaling does not ENABLING ENVIRONMENT

whatever thefuture holds. adapt inarapidly changing world to a nimbleeducationecosystem ready to to date. Moreover, itiskey todeveloping evidence of whathassuccessfully scaled is thebest way tomove forward basedon marginalized children are reached. This bring as well as ensuring that the most leveraging allassetstheseactors must beinclusive andadaptive, sector, and academia.Theseecosystems civil societyorganizations, totheprivate from households,tocommunities, to effectively contribute their expertise—

11 12 Millions Learning: Scaling up quality education in developing countries children andyouth: education ecosystems andultimately helpscale qualitylearningfor millionsof We recommend thefollowing five mainactions to create inclusive andadaptive professionals and technologyspecialists, different types of expertise Fromsupport. nonprofit workers and young graduates tobusiness diverse source actorsoutsideschoolscanbeoneimportant of this assistance. For addressing tougheducationproblems, expertise from who are onthefront linesare overburdened andrequire tangible and energy isneeded.Teachers andothereducationpersonnel communities that are hardest to reach, a creative of injection support To scalequalitylearning inthedeveloping world, including inthe TALENT ANDEXPERTISE OUTSIDETHECLASSROOM ACTIVATE countries andamongregional andglobalactors. global network toallow for experiences andlessonsshared between society—should bediscussed.Thesehubs shouldbelinked through a Approaches ledby allactors—government, educators, business,civil decision-makers tostay uptodatewithrapidly changinginnovations. them. TheIdea Hubs shouldbenimblemechanismsthatallow and sharing effective approaches to improving learning and scaling private sector, should establish Idea Hubs for adapting, identifying, Leaders ingovernments, inpartnership withcivilsocietyandthe NEW IDEASTHROUGHANETWORK OFIDEAHUBS SHARE of R&D ineducation. have the skills and attributes necessary for pushing forward a culture should work togethertocultivateof acohort Learning Leaders who persistent problems. Donors, civilsociety, governments, andbusiness and infrastructure necessarytotrynewapproaches support to is akey step. Governments shouldprovide thepolicyspace, funding, of research anddevelopment (R&D) withintheeducationecosystem approaches tosolvingproblems atscale. Building a strong culture government to civil society to business—must embrace new Leaders across of all parts the education ecosystem—from A CULTURE OFR&D INEDUCATION DEVELOP

seize thismoment intimeand scale theglobalcommunitycollectively can of possibility—thestory story of how Ultimately, Millions Learning THE MIDDLEPHASE FUND body of evidence onhow toscalequalitylearning interventions. approaches tolearningasthey unfold, contributing tobuildinga space toexamine anddocument theprocess of scalingeffective through aReal-time ScalingLab. Such aforum would provide good. Theresearch community shouldimprove dataonscaling Metrics Task Force recommendation on learningdata as aglobal 21 know and do. In particular, new ways of helping teachers assess and move upthrough nationallevel dataonwhatchildren can learning shouldstartattheclassroom level andbeusedby teachers where dataare sporadic andoften of limiteduse. Data onstudent student assessment systems, indeveloping particularly countries Government anddonoragenciesshouldstrengthen national AND SCALING DATA WHAT WORKS LEARNING THROUGHBETTER MEASURE ANDLEARN building scalingcapacity. provide including flexiblefor core costs,support, whichiscrucial for scaling. Additionally,support donoragenciesandfoundations should should develop a more organized ecosystem of education funding to national level. Governments, donoragencies,foundations, andinvestors funding gapbetween newideasorprototypes andimplementation ata of death” toscale. Too often, promising approaches fall toa victim This iscrucialtohelpeffective educationapproaches cross the“valley their respective countries. technology, teachers tosupport andothereducationpersonnel in launch boldAll-In Community initiatives, includingthrough leveraging Governments, civilsociety, andthebusinesscommunity should roles, andhelpreach children whoare falling through thecracks. can bestrategically tappedtoassist educators, elevate themintheir st -century skills will be essential. This is in linewiththeLearning isthe

contribute totheir societies. them toreach theirfullpotential and children andyoung people, enabling quality educationfor alltheworld’s

13 14 Millions Learning: Scaling up quality education in developing countries I schooling intheAmazon jungle Brazil’s new approach to FILIPE’S STORY: 15 16 Millions Learning: Scaling up quality education in developing countries which aimtoachieve high-qualityuniversal Sustainable Development (SDGs), Goals In light of the new United Nations (UN) in Brazil’s Amazon region. educational experiences of young people (SEDUC), isthekey. It istransforming the Amazonas StateSecretariat of Education developed and spearheaded by education and classroom instruction, innovative approach blendingdistance in Brazil, tocontinue theirschooling. The to someof themost remote communities villagesacross Amazonas,2,300 home Filipe and300,000 other students from Media Center hasmadeitpossible for Center teacherinManaus. lessons streamed live from aMedia local government classrooms toreceive surrounding villagestooneof athousand boat along with 19other students from going toschool. Instead, hecommutes by would take weeks totravel, orstop state, hundreds of miles away, which Manaus, the capital city of Amazonas reaching hishighschoolyears: move to have beenfaced withtwo optionsupon the 11thgrade. Adecadeago, hewould students intheircommunity attending another teenagerwere theonlytwo a totalof 78 people. In 2015, Filipe and Brazil. His families, villagehasonly20 along therivers of theAmazon junglein of smallcommunities located thousands Filipe isateenagerliving inoneof the Brazil’s new approach to schoolingintheAmazon jungle Filipe’s story: adaptable educationecosystem thatis a refreshing example of aninclusive and approaches. Brazil’s Amazonas state is countries willneedtoadapttheircurrent and improve quality. To dothis,many grappling withhow to reach every child secondary school,nearlyevery country is and youth from earlychildhoodthrough education for allthe world’s children Amazon rainforest. all students scattered throughout thevast schooling and classesin alldisciplines to very complex challengeof providing Amazonas state tofindasolutionits be fulfilled by the states 2016by forced education asaright for allBrazilians to lawmandatinghighschool 2009 Meanwhile, the federal government’s without being in the national spotlight. policymakers more flexibility to experiment and geographic isolationgave the learning. Amazonas’ peripheral status the student essential for participation that represented them,andlacked among teachers andtheorganizations programs, which were unpopular the failure of traditional distance learning their remote villages.They hadwitnessed glaring gaps insecondaryeducation in SEDUC,whohadavisiontoclose the entrepreneurial group of policymakers Media Center grew from an child’s right toaqualityeducation. partners expertise todeliver onevery harnessing creative ideasandleveraging model, collecting dataonhow studentsmodel, collecting were abletouseevidence andadaptits three years of secondaryschool, they program grew from one grade to a full and continue toimprove SEDUC.As the strengthen themanagement capacity a privateconsulting firm,tohelp sector ministry alsobrought inBain Capital, providers andstudio operators. Thestate partners,such assatelliteservice sector required collaboration withprivate develop aninfrastructure insuch areas student learningwas thegoal,andto in remote areas andbeconducive to in mind.Technology thatwould work ground upwithstudents andteachers ultimately they designeditfrom the between teachers andtechnology— fundamentally reframed therelationship experts inBrazil todevelop amodelthat worked withteachers andpedagogy Along theway, officials from SEDUC administrative functions. and handlesclassroom management and students’guides andsupports learning has a generalist, tutoring teacher who same time. Additionally, eachclassroom students inhundreds of classrooms atthe not onlylecturebutalsointeract with allowing expert teachers in Manaus to broadcast viaabi-directional camera, In theMedia Center model,lectures are ineducationdevelopment. opportunities attention onlocalneeds,constraints, and Rossieli Soares da Silva, whofocused their Amorim andhissuccessor Secretary former secretary of educationGedeão pioneering leaders withinAmazonas state, approach was theentrepreneurial and Media Center’s more locallyrelevant approach todistance educationinto transform thetraditional national What ultimatelymadeitpossible to cases such asMedia Center, where state the globe, by examining adiversity of just how thischangeishappeningacross Millions Learning remotest areas. the learningcrisisfor children ineven the student andteacherdemandcanaddress private collaboration, sector anddriven by effectiveness, enriched by civil societyand experiment, armedwithevidence of by leaders withavisionandabilityto ecosystem. Systems thatare spearheaded an inclusive andadaptive education Media Center illustrates thepower of had ashared vision. sector, that organizations and private the diverse partners, includinginternational able toscaleMedia Center by inviting in strengthenedSEDUC was further and high-quality educationfor theirchildren. movement by rural populations to obtain driven by demandfrom communities—a areas. Much of theexpansion hasbeen technology is appropriate for the remote rather than many specialists, and the far, centers need onegeneralist teacher Brazil, inthatstudents donotneedtotravel scaling inremote areas inAmazonas and components of themodel are suited to high schoolandadulteducation.Thecore education from middleschoolthrough remote communities delivering distance to seven otherstates across Brazil with outside of Manaus. It hasalsobeenadapted of all secondary school students enrolled grade in2007 now reaches 23percent experiment for 10,000 students inthe10th its abilitytoscale. Whatstartedasan As interesting asthemodelitselfhasbeen and theinteractive technology. and teachers responded tothelessons takes adeepdive into 17 18 Millions Learning: Scaling up quality education in developing countries the SDGs willbefar Next,the SDGs from met. learning—and how, at the current pace, detailing the massive gaps in access and accelerate progress inglobaleducation, We startby outliningtheurgent needto students’ learningoutcomes. education approaches thatimprove core ingredients behindscalingeffective this reporttakes a look atsomeof the access toprimaryschoolhasspread, to improve student learning,orhow Rather thanfocus onwhatworks scaled newapproaches toeducation. in partnership witharange of actors, and non-state actorshave pioneered and, a range of actors, scaled new approaches to education. non-state actors have pioneered and,inpartnership with diversity of cases suchasMedia Center, where state and change ishappeningacross theglobe, by examining a Millions Learning takes adeepdive into just how this foster scaling. enabling environment are necessaryto and finally what characteristics of an interventions, how financeplays a role, scale inmind,how todeliver andimplement areas of how todesigninterventions with interventions, organized into four main “core ingredients” for scalinglearning and existing literature, we14 identify we studied. From thesecasestudies for andanalyzingthe14cases selecting live. We alsooutlineourmethodology lead tolasting improvements inpeople’s and deepeneffective approaches that it asarange of pathways thatexpand we focus on scaling and how we define 19 20 Millions Learning: Scaling up quality education in developing countries II URGENTLY NEEDED IS PROGRESS EDUCATION ACCELERATED 21 22 Millions Learning: Scaling up quality education in developing countries followed trends in educationover thepast like anambitiousgoal,those whohave education community this might not seem While tosomeoutside theglobal relevant andeffective learningoutcomes. from pre-primary through secondary, with aimsfornow to2030 universal education driving theglobaleducationagendafrom Thegoal primary schoolenrollment. the educationMDGtoreach universal compared totherather narrow focus of lifelong learning,”exemplifies this when and qualityeducationfor allandpromote 4, to “ensurechallenging. Goal inclusive broader, more ambitious,andmore and targets are contained intheSDGs Development (MDGs), thegoals Goals to theirpredecessors, theMillennium inequality toclimatechange. Compared make progress oneverything from income a new UN global agenda that aims to In 2015, 193 countries adopted theSDGs, on learningfor all Global goalsplace apremium that moment requires arecognition Seizingquality learning opportunities. moment tofocusopportune onscaling scale todateprovides anurgent and all withthemassive failure todosoat on improving learningoutcomesfor education agendathatputsapremium The combinationof anewglobal progress isurgently needed Accelerated education ten 50 yearsten 50 ago. around theworld compared tofive out of primary-school agetodayare inschool countries, nineoutof 10children of disparities still exist across andwithin right toaneducation. demanding, thatitfulfilltheirchildren’s with parents expecting, andsometimes country intheworld hasaschoolsystem Unlike years 50 ago, every todayvirtually to scale” stories todate(see Figure 1). one of themost widelysuccessful “going the globeover thepastyears 150 remains indeed the spread of schooling around expanding access to primary school, and education systems have beenfocused on such as global citizenship. For decades, as well ashigher-order competencies, basics, such as literacy and numeracy, new goalsfocus onbothmastering the levels In is no easy addition, the task. all children achieve sufficient learning few decadesknow thatensuring that demonstrated soclearlyinAmazonas. and adaptive educationecosystems as Scaling qualitylearningneedsinclusive within andoutsidetheformal system. related andco-dependent actors,both fact an ecosystem, whichincludesmany only by governments. Education is in that delivering educationisnotdone

2 Whileenormous 1

Million students in the world 400 800 600 200 CHILDREN OFPRIMARY SCHOOLAGE AREINSCHOOL AROUNDTHEWORLD 0 The rapid spread of schoolinginthepast years 200 1815 A scaling success story: 1835 STUDENTS PRIMARY TODAY 9 OUTOF10 1855 Source: World Development Indicators (2013)

1875 Source: Schofer (2005) SECONDARY STUDENTS 1895 FIGURE 1. 1915

1935 STUDENTS TERTIARY 1955 1975

1995 2007 23 24 Millions Learning: Scaling up quality education in developing countries goal thatwillbehard to achieve is ofAnother aspect the global education gap inmath. to science—andthey might never close the level indeveloped regions when itcomes countries just toreach today’s average than 100years for students in developing it hasbeenestimated itcouldtake more future is grimaswell. Despite limiteddata, and numeracy skills. at orabove second-grade level literacy in 2013, onlyone-third of third graders were division. text in 2014, and just 26 percent could do fifth graders could read asecond-grade single word. percent of secondgraders couldnotread a on theirlanguageandregion, 83to94 is staggering. In in Mali, 2008 depending Figure 2). millionchildren whoare130 inschool(see reading andmathematicsskills—including children around theworld lackbasic estimates thatmore thanone-third of in children enrolling inschool,research countries. Despite thelarge increases today’s low levels of learningindeveloping This 100-year gapcanbeseenclearlyby todayandinto thefuture. both in developed anddeveloping countries— year gap” between educationaloutcomes secondary school,andthere isa“100- early, learningwell, andcompleting been far lesssuccessful onenrolling and compulsory. However, progress has build schoolsandmake educationfree is ascalingmodelfor enrolling children: In low- andmiddle-incomecountries, there The scope andscale of challenges Where education hasfailed: 6 In Kenya, Tanzania, and Uganda 4 For somecountries, thesituation 5 8 In India, lessthanhalfof rural

7 The picture into the Thepictureinto the 3

progress would needtobethree times secondary school.To reach 100percent, to 24-year-olds willhave completedupper only63percentSDGs, of theworld’s 20- 2035, five years past thedeadline for the have completedsecondaryeducation.By Capital, itwillbedecadesuntil allyouth for Demography Human andGlobal projections from theWittgenstein Centre huge challengeaswell. According to secondary educationremains a On theotherendof thespectrum, interventions. are not benefiting from earlychildhood poorest andmost marginalized children for theyoungest children. childhood development programs exist in most countries andfew quality early attend apre-primary educationprogram but still percent less than 50 of children percent since1999 across theworld, pre-primary schoolhasincreased 64 and NorthAmericans. schooling levels of early1900sEuropeans African adultsjust reached average completion. In 2010, South Asian and still 100years behindintermsof schooling school. Theworld’s poorest countries are childhood through theendof secondary every childcompletingschoolfrom early the future. countries—both today andinto in developed anddeveloping between educational outcomes There isa“100-year gap” 11

9 10 Enrollment in In general, the completion. The ODIpapermadecalculations for reaching universal lower secondary schoolattainment anduseddifferent regional definitions, explainingthediscrepancy of schoolandlearnthebasics. children completefour years 400 1 Note: This graph ismotivated by thework undertaken at theOverseas Development Institute (ODI) butcontains theauthors’ calculations for uppersecondary school from ourfigures. Here the projection“goal” ismerely a straight lineanddoesattempt to projectacceleration the actual needed for thetrend to reach 100percent. Reaching universal secondary education:

20- to 24-year-olds who have of38% children not learningbasicliteracy andnumeracy completed upper secondary school million 100% 40% 80% 60% 20% 0% Source: Wittgenstein Centre for Demography Human andGlobal Capital (2015) 1990 The learningcrisis: 1995 Source: EFA Monitoring Global Report (2013-14)

World 2000 2005 Projected gap Africa 2010 FIGURE 3. FIGURE 2. FIGURE 2. 2015 SDG Goal 4.1SDG Goal Target 2020 1 2025 2030 children completefour years 2035 of schoolwithoutlearning

children donotcomplete 100% 39% 63% 130 120 million four years of school. million h basics. the

25 26 Millions Learning: Scaling up quality education in developing countries classroom ornational level. and policymakers ata regularly usedby educators data onlearningthatcan be we often lacksystematic complicated by thefact that These challenges are further that of the 60 poorest countries in the Partnership for Education estimated outcomes. globaldatagap”as “the onlearning by theglobalLearning Metrics Task Force a classroom or national level, referred to used by educators and policymakers at data onlearningthatcanberegularly by thefact thatwe often lacksystematic These challengesare complicated further facing theseinequalities. high-, middle-,andlow-income alike, are children. are out of school versus 9percent of urban children, and40percent of rural children school versus only2percent of therichest percent of thepoorest children are outof rural children. In Nigeria, for example, 71 boys andgirls,between urbanand between therichest and poorest,between Massive inequalityexists inmany countries within countries thatare struggling. that itisthemost marginalized populations developed anddeveloping regions, itisclear In additiontothe100-year gapbetween marginalized whoare falling behind. continues. Again, itisthose whoare most secondary schoolifthecurrent trajectory percent of youth willhave completed regions. In Africa,for example, only39 The challengeisconcentrated inlagging as fast ascurrent rates (see Figure 3). 12 Countries around the world, 13 Arecent study by theGlobal Africa every three years, ortheProgram is administered across Francophone of Education Systems (PASEC), which such astheProgramme for theAnalysis literacy andnumeracy. Theseprograms, ofaspects children’s learning,usually country andprovide snapshots onspecific or programs thatare housedoutsidethe make useof avariety of assessment tools practice. In addition,countries frequently for real-time adjustments topolicyand andareacademic subjects of limiteduse The exams frequently focus onasubset of exams attheendof aschoolingcycle. most countries administer annual national system. place for astudent learningassessment world, only two hadthebasicelements in Education (CUE) atBrookings have found Colleagues attheCenter for Universal an additional$39 billionannually. through secondaryschoolwillrequire quality education from pre-primary goal of ensuring thatallchildren acquire UNESCO estimates that meetingthe challenge inachieving theSDGs. Financing educationisanadditional needed inpolicyandpractice. policymakers toinform whatchangesare that canbeusedby educators and feedback on how students are progressing countries themselves andgives regular learning assessment system thatisrunby replacement for awell-developed student learning levels. However, they are no contribute toaglobalpictureonchildren’s exercises for countries andimportantly assessment for math,canbeuseful Mathematics andScienceStudy(TIMMS) Boston College’s Trends inInternational and Development every three years, or Organisation for Economic Co-operation (PISA), whichisadministered by the for International Student Assessment 14 Across thedeveloping world, 15

Heckman, “skill begetsskill . learning of Nobel Prize-winning economist James control, and perseverance. In the words communication, problem solving,self- includes teamwork and cooperation, frequently called21 beyond, largely withafocus onwhatare primary school,secondaryand foundations for development further in Early childhoodprograms laythe societies andeconomies. grown willhave tremendous on impacts at therate primaryschoolingaccesshas that improve both access and learning workforce. Scalingeducationprograms prosperous economies,andastrong benefits, includinghealthy children, drives many of thesocialand economic learning thatoccurs inschools,which increasingly datashow thatitisthe school hasledtoincredible gains.But Globally, thefocus onaccesstoprimary Why focus onlearning? and enhancedeffectiveness of spending, among donors, aneye toward results They recommend bettercoordination gaps and reach marginalized populations. will make itdifficulttoclose thefunding fragmentation between different funders aid toeducationover recent years andthe the amount needed. In addition, declining education finance, it remains far short of is increasing andisthelargest source of that whiledomestic government spending will have tremendous impacts onsocieties andeconomies. learning attherate schoolingaccess primary hasgrown Scaling education programs thatimprove both access and st -century skills.This begets learning.” for education. of lackof sufficient and effective financing research toaddress thecontinuing barrier children by bringingtogethercritical how to meet theneedsof alltheworld’s established inNovember 2015, istackling EducationGlobal Opportunity, whichwas International Commission on Financing innovative forms of finance. actors, includingharnessingnewand and greater resources across traditional economic growth, theirwork helps to by comparinglearningoutcomeswith growth across countries. For instance, levels—explain differences in economic that differences inskills—not schooling Ludger Woessmann have recently found education economists Eric Hanushek and estimations $8for every $1invested. programs have such highreturns, by some is why earlychildhooddevelopment and lower levels of crime. and secondaryschool,higherwages, including literacy andnumeracy inprimary lead tohigheracademicachievement wages and improves health outcomes, more years of schoolincreases people’s countries have shown thatcompleting and societies.Whilestudies inmany have aprofound oneconomies impact abounds thattheskillslearnedinschool Beyond individualreturns, evidence 17

18 These21 19 st Perhaps this -century skills 16 The 20

27 28 Millions Learning: Scaling up quality education in developing countries and leadprosperous lives, meaning there thinking skills, to contribute to the economy communication, collaboration, andcritical includingliteracy andnumeracyset, plus need tobeequippedwith a robust skill countries. developingglobally usingdatafrom 30 World Bank research confirms this finding skills, however, have grown rapidly. requiring andinterpersonal analytical fallen over thepast halfcentury. Those “routine” skillsinthelabor market has that over time, theshare of jobs requiring Frank Levy, found andRichard Murnane United States,economists David H.Autor, widen thisgap. Using datafrom the improvements totechnologywillonly and theneedsof and thelabormarket, gap between thecapabilitiesof youth the board there isasignificant skills academic studies allfindthatacross surveys, labormarket analyses, and today andinto thefuture. Employer forimportant success inthelabormarket and 21 Learning theright balanceof academic higher. to13.8growth couldbe8.5 percent on international exams, globaleconomic skill level of Finland, often atopscorer if allcountries possessed thecognitive growth. theauthors In estimate fact, that levels of learningcontributed tostunted while Latin America’s comparatively lower to East Asia’s rapid economicexpansion, than years of schooling,thatcontributed by mathandscienceassessments rather It was thehighlevel of skills,measured of economicgrowth from 1960 to2000. completion yet drastically different levels could have similarlyhighlevels of school explain why Latin America and East Asia 22 st -century skillsisincreasingly 24 Thefuture workforce will 23

child intheworld. toreachlearning opportunities every to thedrivers behindscalingquality understanding andgreater attention world. Theseshiftscallfor adeeper navigate anincreasingly connected digital literacy and critical thinking to people willneedahost of skills,including the environment and health, young urbanization willpose newchallengesto and SouthAsia. Population growth and youth, insub-Saharan particularly Africa and bettereducationfor children and from developing regions, callsfor more increasing share of theworkforce come with a demographic shift that will see an populationgrowth, coupled vital. Global societies willonlybecomemore andmore can bringtobothindividualsand In thefuture, thebenefitsthateducation of natural disasters. education are more resilient intheface that communities withhigherlevels of sustainability. growth andimprove environmental run cansignificantly slow population have fewer children, whichinthelong school. to getthemvaccinated, andsendthem and alsohave well-nourished children, are more likely tobehealthy themselves levels of education. isduetomothers2009 havinghigher globallybetween 1970mortality and has shown thathalfof the declineinchild the environment.For example, research can have greatonhealthand impacts improved levels of educationandlearning Equally aseconomic benefits, important systems toclose theglobalskillsgap. is anurgent needfor strong education 26 Additionally, educatedmothers 27 Research hasalsoshown 28 25

Educated parents 29 SCALING SO MILLIONS LEARN: III Defining a global priority 32 Millions Learning: Scaling up quality education in developing countries A crucialquestion for theglobal Defining aglobalpriority Scaling somillionslearn: education’s scalingstory. Our insights Millions Learning education. not comprehensive, of aspect quality of competencies,acritical,although learning outcomesacross adiversity to theterminthisreport asimproving of experiences andoutcomes, we refer education canencompassawiderange different contexts. While the term quality learning models,whichcanbeadaptedin principles of effective teachingand process behindspreading theessential brands orfixed models,but rather the scaling isnotfocused onexpanding learning have scaled.Our interest in approaches toimproving children’s the process behindhow effective countries. We seektobetterunderstand has scaledinlow- andmiddle-income quality educationfor children andyouth Our central question iswhere andhow lives. This is the focus of lead tolasting improvements intheir to high-qualitylearningexperiences that ensure that more children have access the adoptionofishow to theSDGs, education community, especiallysince income countries. for children andyouth hasscaled inlow- andmiddle- Our central question iswhere andhow quality education seeks tocontribute to Millions Learning . ecosystems are essential for effective that inclusive andadaptive education and outcomes.Ultimately we find improve children’s learningexperiences successfully scaledapproaches that pathways andtodiffering degrees, (see Annex 2)thathave, alongdifferent review andby careful study of 14cases are informed by anin-depthliterature quality education. comprehensive, aspectof a critical, althoughnot diversity of competencies, learning outcomes across a in thisreport asimproving outcomes, we refer to theterm wide range of experiences and education can encompass a While theterm quality across thedeveloping world. to scalequalityeducationmore quickly around whatneedstobedoneinsight, discussion, basedonnewevidence and of approaches togoscale. Thegoal Millions Learning istocatalyze a researchers such asKatharine Conn up success.” “stop ignoringwhatwe know andscale called on the education community to widely citededucationresearcher, has some time. For example, John Hattie, a for improving children’s learningfor has beenstudying effective approaches Fortunately, theeducationcommunity Addressing anevidence gap and youth, althoughfrequently notall forlearning opportunities children is amixof formal andnonformal Within thisconstellation there invariably youth country orcontext. inaparticular engaged ineducatingchildren and and privateactorsthatare sector all constellation of government, civilsociety, ecosystem, meaning the broad Here we focus on the education children’sthat support andyouth’s learning. lasting influenceoneducationecosystems effective practicesorinnovations have are interested intheprocess by which policies orpractices? In thisway we across thesystem changingtheir taken up, ultimatelyleadingtoactors children’s learningbecomeincreasingly demonstrably effective inimproving How do approaches that have been We focus of scaling. ontheaction up across theeducationecosystem. these approaches spread andare taken or nongovernmental actors,andhow learning, whetherledby government or innovations toimproving children’s Our entry point iseffective approaches A focus onscaling 31 Hattie, alongwithother children. research onimproving learningfor all expand inaneffort toadd existing which effective interventions grow and have concentrated ontheprocess by on at-scale educationreform andhence worknot seektoduplicatetheimportant intersections.important However, we do of scalingandat-scale reforms have series. and Successful Reformers inEducation and Development’s Strong Performers Organisation for Economic Co-operation Better Education Results (SABER), andthe World Bank’s Systems Approach for Development Progress initiative, the the Overseas Development Institute’s scale educationreforms, for example, complements ongoingwork onat- This focus ontheprocess of scaling education. exercise theirright toaquality can orresources,those with littleopportunity of ensuring thatall children, especially government thatbears theresponsibility ecosystem of actors,itisultimatelythe on teachingandlearning. Whilethere student-centered interventions thatfocus of of similarities,such astheimportance analyses have demonstrated anumber developing world, thesevarious meta- learning. approaches to improving student numerous studies effective toidentify and Patrick McEwan, hasreviewed are strategically connected. 30 Clearly, these perspectives 32 Across thedeveloped and 29 In this 33 34 Millions Learning: Scaling up quality education in developing countries International, andothers. ManagementExpandNet, Systems by Brookings Institution colleagues, of PoorPeople Bring Development Solutions to Millions publication the process of scaling.Thisincludesthe there isagrowing evidence baseon development outside of education, In theresearch oninternational is unfortunately quitelimited.” promising educationinitiatives inAfrica empirical research onscalingup stated that“accessible systematic Molapi Sebataneand AmbassadorE. Martial Dembélé, Professor Joel Samoff, the evidence related toAfrica,Professor not how toscalewhatworks. examines questions of whatworks and focused onimproving learningprimarily literature for developing countries that is Additionally,Goals. theeducation the UN’s past Millennium Development this asithasbeenthecentral focus of global efforts have coalescedaround more children into school. Most recently all played arole inhelpinggetmillions programs, andforeign have aidsupport level. Government policies, innovative at the primaryto education, particularly stories hasbeenof expanding access countries, oneof the greatest scaling may bebecauseacross developing on improving learningoutcomes.This expanding accesstoschoolandnot education doesexist, itisfrequently on Where evidence onscalingand shown to make adifference. effective approaches thathave been an equallypressing needistoscale across different skills,ages, and contexts, explore what works to improve learning is undoubtedlyaneedtocontinue to Getting to Scale: How to andotherongoingwork 35 Unfortunately, 33 34 Reviewing

Johannes F. Linnstate in their review of nonresident seniorfellow at Brookings As Professor Arntraud and Hartmann results atscaleandthatthey will not. guarantees thatthey have notachieved assistance of lessthantwo years virtually projects, ameanduration of external In a review of international development countries toimprove learningoutcomes. they face developing insupporting going toscaleasthebiggest challenge investors inglobaleducationidentified multilateral, foundation, andcorporate For example, a2013 survey of 36bilateral, development them. actorsthatsupport is alsoaconcernfor theinternational to scaleeffective approaches tolearning government ornongovernmental actors developing world. But how tosupport halls of ministries of educationacross the investment are beingdebatedinthe what specialprograms meritlong-term effective, how toallocate resources, and groups. Questions of whatpoliciesare concern for governments andcivilsociety outcomes for children isanabiding quality learningexperiences and How toscale education that delivers included afocus oneducation. and scaling found that only 16 of 158 studies oninternational development scaling literature. One review of existing education isnotwell covered inthis need to besuccessful inlife. helps themlearntheskillsthey access to aneducation that access to education butalso how to scale not onlychildren’s emerging bodyof evidence on We seekto buildonthe 36

38 37

from theinfluential work of Cynthia program, practice, oridea.Borrowing approaches mayconsist of apolicy, improvements inpeople’s lives. effective approaches that lead tolasting of pathways thatexpand anddeepen the term“scaling” torepresent arange definitions inthescalingliterature, we use institutionalizing practices.Building off such asexpanding tonewcountries or spread, explosion—and outcomes, of processes—expansion, replication, many different ways, describinga variety international development literature in “scaling” hasbeenusedthroughout the across sectorsandactors.Rather, theterm accepted definition of scalingthatisused study theproblem. There isnot a single these interventions, andhow researchers and implemented, how donors fund how programs and policies are designed policymakers develop scaling strategies, How scalingisdefinedinfluences how Defining scaling children’s accesstoeducationbutalso of evidence onhow toscalenotonly We seek to build on the emerging body understand education’s scaling story. Hence, up.” focus onlearningand especially scaling usually to thedetriment of anadequate industry andthedevelopment business, focus oninnovation isendemicintheaid scaling ininternational development, “the 39

Millions Learning that leadto lasting improvements inpeople’s lives. pathways that expand anddeepeneffective approaches We usethe term “scaling” to represent arange of seekstobetter 40 These the sametime. context, most likely should,bepursued at pathways can, and depending on the mutually exclusive. One ormore of these multiple pathways toscale, theseare not of theactorsinvolved. Whilethere are resource constraints, andcapabilities the intervention, politicalenvironment, are taken dependingonthegoalsof effective approaches,pathways certain private thatisengagedinscaling sector it isgovernments, civilsociety, orthe scaling, describedinBox 1.Whether terms of four different pathways for We have found itusefultothinkin improvement. andsustainsupport alasting educational the depth of change necessary to coverage whilesimultaneously ensuring ofthe act “scaling” isaboutexpanding University, isthat whatisimportant Coburn, professor atNorthwestern been usedtoinform policy andpractice. respective sector’s scalingstory and haveThese reports contributed totheir Institute’s sister publication, the International Food Policy Research Millions Saved Development’sGlobal seminalreports, such as healthwiththeCenter for by thescalingstories from othersectors, in life. Inwe have thiseffort, beeninspired learn theskillsthey needtobesuccessful access to an education that helps them 41

, and agriculture with Millions Fed . 35 36 Millions Learning: Scaling up quality education in developing countries FOR SCALING UP DIFFERENT PATHWAYS IMPACT FOUR

Pathways to scale

FIGURE 4. FUNCTIONAL ORGANIZATIONAL VERTICAL HORIZONTAL More activities capacity Strengthening and places More people Policy adoption

37 38 Millions Learning: Scaling up quality education in developing countries education, andmicrofinance, among other areas of community development. has expanded itswork to includehealth interventions, nonformal andformal help refugees returning from the country’s liberation war. Over time, the NGO BRACNGO started in Bangladesh in 1972 as a small rehabilitation project to Pertains to theexpansion of thetypeof activities orareas of The engagement. FUNCTIONAL SCALING: inclusion andeconomic citizenship education for children andyouth worldwide. and strengthen existing efforts of individualentities committed to financial children andyouth. The ideabehindthisglobalnetwork isto poolresources broader globalsocialmovement to strengthen thefinancial capabilities of International spinningoff Child and Youth Finance International to createa of other institutions or the creation of a new entity. An example is Aflatoun efficiency, andsustainabilityactivities. of This canalsoincludethe involvement Pertains to increasing organizational strength to improve effectiveness, ORGANIZATIONAL SCALING: based curriculumfor thesecondary schoolentrepreneurship subject. incorporated thiseffective approach into national policy by rolling out askills- after close collaboration withEducate!, theMinistry of Education inUganda skills across five districts in Uganda, targeting secondaryschools. In 2012, and curriculaonentrepreneurship, leadership, andworkforce readiness a school-basedinitiative in2009, buildingaskills-based teaching approach nongovernmental organization (NGO) Educate! isanexample. It started as Pertains to initiatives moving from local to nationwide The engagement. VERTICAL SCALING: Street reaches millionsof children inmore countries than150 around theworld. children for schoolby combining entertainment with education. Today, Sesame Sesame Workshop, started intheUnited States in1969 withthegoalof preparing district, andtoday ithasbeenexpanded across all10provinces of thecountry. government, for example, piloted anapproach to teacher development inone geographic area to another to reach more peopleandcommunities. The Zambian Pertains to breadth of coverage of anintervention, theexpansion from one HORIZONTAL SCALING: Box 1. Pathways for scaling effective approaches Policy adoption More peopleandplaces More activities Strengthening capacity 42 a measurable improvement of learning cases were thatdemonstrated selected more than100potential casesreviewed, been scaledtovarying degrees. Out of where qualitylearningfor children has in-depthanalysis ofconducted 14cases Using acasestudy method,we scope of theproject. 2 provides adetailedsummary of the tools todemonstrate improvements. Box data from existing learningassessment multiple competenciesandrelied on has looked atstudent learningacross secondary education. Our analysis earlychildhoodthroughparticularly on education for children and youth, secondary research. We have focused we have primaryand conducted low- andmiddle-incomecountries, for children andyouth has scaled in of where andhow qualitylearning To helpanswer ourcentral question Methods guiding and practices around girls’ education (functional). magazine-style talk shows inallthree countries, whichare changingattitudes of Kenya, Ghana, andNigeria (horizontal). In addition, DLAproduced popular (organizational), DLAexpanded to 1,000 additional schoolsacross new regions Department for International Development (DfID)’s Girls’ Education Challenge in apublicprivate partnership withDiscovery Communications andU.K. centers”“learning inalmost schoolsin16countries. 500 In 2015-2016, media to improve student learningandteacher effectiveness, had established in 2012, Discovery Learning Alliance (DLA),anonprofit organization using As discussed,these pathways often happenincombination. For example, PURSUING MULTIPLE PATHWAYS Millions Learning in geography andpopulationsreached. secondary programs, aswell asdiversity of interventions from early childhood to a range of scalingpathways andtypes balanced against the desire to identify that hasyet tobeexplored. Thiswas all of expanding, anewangleoraspect course correctionmadeintheprocess contentious issue addressed, somebold an interesting story totell—some experience, whichoften included these casestoexamine theirscaling being debated.Rather, we selected some of theapproachesin fact, are or approaches toimproving learning; meant torepresent thebest models documented cases.Thecasesare not well-known cases,aswell asbetter- endeavored tobringlight local,less criteria). Wecomplete selection have contexts andlevels (see Annex 1for income country across adiversity of to differing degrees inalow- ormiddle- 39 40

Millions Learning: Scaling up quality education in developing countries AMAZONAS STATE GOVERNMENT’S MEDIACENTER Over 5millionstudents Alternative andlower primary secondary schoolprogram 16 countries, starting inColombia FUNDACIÓN NUEVA ESCUELA Approximately 156millionchildren development, Early grade reading &mathematics Early childhooddevelopment, Cognitive in theUnited States More than150countries, starting SESAME WORKSHOP, SESAME STREET or 12% of Ugandan secondaryschools 120,000 students schools, partner in350 development, Teacher training Secondary education, Entrepreneurial Rwanda, Uganda EDUCATE! 300,000 students to date Alternative secondaryschoolprogram Brazil, Colombia, Ecuador, Honduras, Nicaragua SISTEMA DEAPRENDIZAJE TUTORIAL (SAT) reading digital books permonth Over 5.6 millionpeopleto date &1.1 millionpeople Literacy andsecondaryschool inprimary 69 countries, starting inGhana WORLDREADER 1.8 millionstudents &46,000 teachers to date Teaching training andsecondaryschools inprimary Zambia (all 10provinces) STUDY,LESSON ZAMBIA teachers& 60lecturing to date 300,000 tutoring students, teachers, 2,200 Formal secondaryschooldistance learningprogram Amazonas State, Brazil Case study overview

Sesame Street Workshop, Sesame FIGURE 5.

Fundación Tutorial (SAT) Aprendizaje Escuela Sistema de Nueva

Government’s Media Center Amazonas State State

Worldreader

Educate! Lesson Study, For All Teach Zambia

International Academies Bridge Jordan INJAZ,

Read India Pratham, Program Literacy Room to Read,

Non-Formal Education Primary Primary BRAC, International Aflatoun

. over 6millionstudents indirectly viastate or district government partnerships . Over 350,000 students directly inrural areas from 2014-2015 & Early childhooddevelopment, (low-cost Primary private school) . . specifically) &10million students (Room to Read overall) to date education inearlychildhood,primary, andsecondary schools BRAC, NON-FORMAL PRIMARY EDUCATION Bangladesh, Cambodia, India,Laos, Nepal, SouthAfrica, 110,000 students and2,000 teachers (Literacy Program 1.1 millionstudents teachers &52,323 andalumnito date Alternative pathways to teaching inearlychildhood, 1.2 millionstudents1.2 &over 23,000 volunteers to date ROOM TO READ, LITERACY PROGRAM Financial literacy, Work readiness, Entrepreneurial 670,000 students inBangladesh&over 900,000 Entrepreneurial development, Socialandfinancial Over 100,000 students & 8,000 teachers to date BRIDGE INTERNATIONAL ACADEMIES students inother four NFPEcountries eachyear Almost 4millionstudents &34,000 teachers, 39 countries, starting intheUnited States facilitators, andpeereducators eachyear SriLanka, Tanzania, Vietnam, Zambia Remedial education schools inprimary Early grade reading, Teacher training AFLATOUN INTERNATIONAL Alternative schoolprogram primary Afghanistan, Bangladesh, Pakistan, development insecondaryschools primary, andsecondaryschools PRATHAM, READINDIA 116 countries, starting inIndia Jordan (all 12governorates) Philippines, South Sudan Philippines, SouthSudan Kenya, Nigeria,Uganda India (across 23states) TEACH FOR ALL and United Kingdom INJAZ, JORDAN

41 42 Millions Learning: Scaling up quality education in developing countries intentional learningoccurs. be programs orpolicies thattake place informal ornonformal settings where learning, we have focused onexamining education interventions. These can While many factors, such as improved nutrition, contribute to children’s EDUCATION INTERVENTIONS: Source: Learning Metrics Task Force Seven domainsof learning Figure 6. Learning Metrics Task Force: relevant, asseeninFigure 6below. and youth. The task force identifies seven domains of learningasparticularly and other actors onthetopic of whatcompetencies are important for allchildren Metrics Task Force, which has engagedabroad range of policymakers, academics, emotional learning,and critical thinking. We are guidedby theglobalLearning across multiple competencies, includingliteracy, numeracy, science, socialand as decreases indropout rates andincreases inprogression rates. strong indication of improvements inlearning,alongwithproxy indicators, such or other methods that are clearly defined to demonstrate progress inlearning; or international, regional, ornational assessments; householdsurveys; evaluations; settings andcircumstances andisnot limited to academic institutions. We rely on skills, and attitudes that ultimately enrich their lives. Learning isdefined asaprocess whereby peopledevelop a range of knowledge, LEARNING: Box Defining 2. thescope of the study well-being emotional Physical Physical Social &

& thearts Culture

POST-PRIMARY communication CHILDHOOD 44 PRIMARY Literacy & EARLY As such,learningcan take place indiverse & cognition Learning approaches 43 Learning can be measured

mathematics Numeracy & technology Science & complete list of advisors andaffiliations). and health(see front cover interior for outside of education, such as technology and academia,including from sectors philanthropy, business,nonprofits, up of experts whorepresent government, by anInternational Advisory Panel, made thinking. We have beengenerously guided economics andpsychology, anddesign science, productdevelopment, behavioral development andscaling,implementation including literature oninternational surveyed various literatures onscaling, depth interviews, andfieldvisits. We also from amix of desk-based research, in- To examine thecases,we analyzed data national income (GNI) percapita. countries asdefined by the whichclassifies World Bank, economies basedon gross For ourpurposes, we are focusing onlower, low-middle, orupper-middle income LOW- ANDMIDDLE-INCOME COUNTRIES: • • internationally accepted definitions of the following: age criterion inthestudy butthree maineducational levels instead, usingthe We have focused broadly onchildren andyouth. We have not usedastrict CHILDREN ANDYOUTH: and ethnic minorities, where access to quality learninghasbeenparticularlylimited. crisis-affected areas, girlsfrom rural areas, children and young peoplewithdisabilities, interested inmarginalized groups, suchaschildren andyouth livinginextreme poverty, • PRIMARY EDUCATION: school entry, includingearlychildhoodeducation orpre-primary. EARLY CHILDHOOD: vocational training andsecond-chance learningprograms. formal lower anduppersecondarynonformal education, suchas SECONDARY EDUCATION: schoolandnonformalprimary programs, suchasaccelerated learning. Approximate age range is zero to 8. through Birth Approximate agerange is5to 15. Includes formal 46 Approximate age range is 10 to 19. Includes Within thiscountry context, we are particularly and STIR Education. RTI International, Schools of Tomorrow, Results Institute, Rocketship Education, Education Reform Sector Program, Rapid Naandi Foundation,M-PESA, Punjab Camfed, Discovery Learning Alliance, These included: findings in the report. from other cases that illustrate specific anin-depthreview,conducted we drew In additiontothe14caseswhere we academics, andbusinessleaders. policymakers, funders, educators, series of roundtables, withpractitioners, extensive consultations, through a We have alsogreatly benefitedfrom 45

43 44 Millions Learning: Scaling up quality education in developing countries IV How hasscaling happened? FINDINGS:

45 46 Millions Learning: Scaling up quality education in developing countries in new, chaoticspaces where innovations how toimprove learning,they often occur where we doseemore radical overhauls to Therefore, perhaps itisnotsurprising that outcomes. “at-scale” problem such aspoorlearning are talkingaboutaddressing analready be thatwhenitcomestoeducation,we tackling at-scale educationreform could scratch. One reason for thedifficulty of an existing system thantoscale from earlier, itcanbemore difficultto reform systems. Often,becauseof choicesmade reforms insidemainstream education under theradar—rather thanlarge-scale began at the margins—in some instances We exa Scaling success from themargins to thistransformational change. Scaling approaches that are effective are central adapt policiesorpracticestotake up that allow spacefor newapproaches and In places around theworld. change in learning is happening in many to therainforest inBrazil, transformational communities. From theslumsof New Delhi even insomeof themost marginalized access to quality education is possible— Millions Learning How hasscaling happened? Findings: clusive and adaptive education ecosystems mined casesthatfor themost part 47

findsthat scalingchildren’s

what canwe learnfrom them? Whatfollows How didthe14casesexamined and doit, education actorstowork together. italsorequiresfrequently these thannot, with creative solutionsto roadblocks. More actors, topivot andadaptcomeup along withcivilsociety, business,andother quality learningrequires governments, recognized asthefather of diffusion of Everett Rogers, a sociologist who is found intheoriesof scalinginnovation. to thecenter isonethathas long been This notionof moving from theperiphery educational options. marginalized children whohadlimited that developed new ways of reaching For others itwas acommunity movement within adistrict totryanewapproach. government givingfreedom toitsofficials case. For some itwas aflexible central What constituted themargins varied by spread andwas taken upmore broadly. started onthemargins before theidea theeffective newapproachesthan not, can flourish. We found thatmore often scaling ininternational development. make totheexisting bodyof evidence on cases reveal andthecontributions they is theeducationscalingstory thatthese this conceptinhis2002 bestseller journalist popularized Malcolm Gladwell innovative ideasandtechniques.” successful applicationandexpansion of “highlights key attributesthatfacilitate discussion of scalingby notingthatit the relevance of Rogers’ modeltothe system.” over timeamongthemembers of asocial communicated channels thoughcertain processas “the inwhichaninnovation is innovation theory, describeditin1962 innovations can flourish. new, chaotic spaces where learning, they often occur in overhauls to how to improve where we doseemore radical Perhaps itisnot surprisingthat instructed study to creative thinking. to approach, aninquiry allowing for students’ learningto transform from Furthermore, teachers switched from traditional chalk andtalk methods activities and develop critical thinking, presentation, and teamwork skills. Study, Zambian students have more opportunities to conduct hands-on date. Internal andexternal evaluations demonstrate thatwithLesson 1.8 million students and 46,000 teachers in all10 Zambian provinces to improving of the supervision school managers. Lesson Study has reached strengthen schoolsystems by encouraging teamwork among teachers and Cooperation Agency, started Lesson Study in2005. The approach aimsto The government of Zambia, inpartnership withtheJapan International and assessing lessons. Lesson Study hasspread to more countries. than50 knowledge andskillsto improve teaching through planning,demonstrating, collaborative learning, whereby and secondary teachers primary share Lesson Study isaJapanese-originated practice of peer-to-peer Box 3.Lesson Study, Zambia 48 Hartmann and Linn explain 49 The to thecenter. innovation canmove from theperiphery illustrates the pathway along which an Nesta, aUK-based international NGO, One usefulframework,developed by early adopters as“crossing thechasm.” moment of securingthe13.5percent of Author Geoffrey Moore describesthe to achieving mass-market success. are the“early adopters” asbeingcritical the 13.5percent of thepopulationwho service, oridea.Heproduct, focuses on innovation andtheadoptioncurve for a discussesthelawof diffusion action, of describing how great leaders inspire For example, in authorSimon Sinek, moving from themargins tothecenter. provide insight into further this notion of an idea or intervention.authorsOther there is critical mass acceptance of The Tipping Point asthemoment when 50

47 48 Millions Learning: Scaling up quality education in developing countries Experimentation under theradar can happen Scaling success from themargins SECTOR PUBLIC BEGINS ONTHEMARGINSBEFORE BEING

A NEW IDEAORAPPROACH OFTEN TAKEN UPMORE BROADLY ECOSYSTEM EDUCATION SOCIETY FIGURE 7. CIVIL SECTOR PRIVATE MARGINS SCALING QUALITY LEARNING FROM STRATEGIES FOR THE MARGINS INNOVATION ADOPTION DELIVERY IDEA TWO TWO 2 1

often for marginalized communities Spreading new approaches across Innovating ineducation delivery EDUCATION ECOSYSTEM an education ecosystem ECOSYSTEM EDUCATION

49 50 Millions Learning: Scaling up quality education in developing countries learning for teachers totheZambian practice of peer-to-peer, collaborative (JICA), adapted a Japanese-originated International Cooperation Agency Zambia, inpartnership withtheJapan is anexample where thegovernment of adopted by different actors. Lesson Study across educationecosystems andbeen to teacherdevelopment—have spread process—from curriculum,tomaterials, components of theteachingandlearning effective newapproaches toimproving With theformer, ideaadoption, delivery for marginalized communities. ecosystem and by innovating in education new approaches across aneducation within developing countries: by spreading has occurred alongtwo primarypathways Scaling quality learning from the margins learning from themargins Two strategies for scaling quality Idea adoption anddelivery innovation: improves learningoutcomes, women’s empowerment, andcivicresponsibility. degrees. A number of peer-reviewed studies indicate that SAT in Honduras graduates can continue onto college orget jobs thatrequire secondary and recognized by governments in Colombia and Honduras, and as aresult, and has reached an estimated 300,000 students to date. SAT is accredited Brazil, andEcuador (Guatemala alsohosted anSAT program until 2005) late 1970s. Today, itisimplemented inColombia, Honduras, Nicaragua, Application and Teaching of theScience (FUNDAEC) developed SAT inthe context of agricultural innovation. The ColombianFoundation NGO for the into teaching methods, such as learning mathematics and science in the marginalized communities. SAT integrates relevant andpractice theory methodology to promote rural education andcommunity development in The program’s trained teachers, ortutors, utilize by a“learn doing” alternative secondaryeducation program for rural youth inLatin America. Sistema deAprendizaje Tutorial (Tutorial Learning System), orSAT, isan Box 4.Sistema deAprendizaje Tutorial education needsandrealities of rural youth. from thenationalgovernment, tomeetthe that isdelivered by anNGO, withsupport alternative secondaryeducationprogram a specificallytailored andtargeted de Aprendizaje Tutorial (SAT) inHonduras, across countries. Anexample isSistema been developed and grown within and the most marginalized communities have new educationdelivery approaches for With thelatter, delivery innovation, context. broader scalingliterature. cases andthatare across supported the the scalingof qualitylearninginthe14 the core ingredients thathave enabled patterns andthemesemerge. Theseare and policiesdiscussed,somecommon Looking across therange of programs just theopposite outcomeinanother. failed in a given setting and resulted in why scalingof areform succeeded or found thatcontext often explained Reviews of scaling educationinAfrica can bereplicated across countries. size-fits-all model or blueprint to scale depends onthecontext. Clearly, noone- The right combinationof ingredients We have identified 14core ingredients thatare organized around four mainareas: Scaling learning:14core ingredients process. operate plays acriticalrole in facilitating orimpedingthescaling national tolocal,theecosystem inwhichprograms orpolicies happen inavacuum. Largely guidedby governments from As criticalasthesethree are, otheraspects scalingdoesnot ENABLING ENVIRONMENT absolute amounts. matters isoften asmuch abouthow they are allocatedasabout While resources matterwhenitcomestoscaling,what certainly FINANCE involves acombinationof technicalandpoliticalactions. of implementing ordelivering atlarge-scale isessential. This As asdesignis,attention important totheoperational realities DELIVERY planning for scalefrom theoutset. Improving learningatscalestartswithcommittedleaders DESIGN 51

of children’s learning. on thecontext, canpromote thescaling dependingin various combinationsthat, we seearange of ingredients deployed impeding thescalingprocess. economy considerations facilitating or specific organizations andpolitical driversimportant of scaling are the Similarly, Linnfindsthatthemost 52 Indeed, 51 52 Millions Learning: Scaling up quality education in developing countries 14 10 11 9 Scaling learning:

E 12 N

A 14 core ingredients

8 B

A F N D N

C G E

E FIGURE 8. L

E I 7

E V

R E I

G O

N 1

N T

5 2 3 4 13 ENVIRONMENT ENABLING DELIVERY FINANCE DESIGN Flexible education Local education Education alliances financing 5 needs 10 1 Learning champions and leaders Cost-effective 6 Supportive policy environment learning Long-term education 2 Technological 13 advances financing 7 Flexible adaptation 11 opportunity Windows of 3 8 “Middle phase” Better data financing Elevating A culture teachers of R&D of 12 14 4 9

53 V DESIGN 1. Local education needs: Introduction Interventions should be designed in response to local demand and should ensure the participation of end-users.

One essential element of designing with a way that was locally relevant—think Cost-effective Elevating scale in mind is to develop programs computers or textbooks gathering dust DESIGN learning teachers and policies that students, parents, or in cabinets. Even in the case studies teachers actually want—not just what examined for this report, there are governments, implementers, or donors examples where programs did not initially 1 2 3 4 think they need. While this sounds take into account a community’s needs obvious, there are countless examples of and needed to make changes along the well-meaning and thoughtful education way. In particular, the cases underscored Local education Flexible adaptation interventions that are not taken up simply the importance of ensuring participation

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions because there was not a real demand from the community from the start and needs for them or they were not designed in building in local accountability.

One essential element of designing with scale in mind is The guidance from the scaling literature any program or policy must tap into local to develop programs and policies that students, parents, strongly suggests that going to scale education needs as identified by any and must be designed for from the outset.53 all of students, teachers, and parents. or teachers actually want—not just what governments, This does not imply limitless scale, but However, responding to local demand implementers, or donors think they need. rather, a clear vision of what the endgame is not enough. Interventions must also is and a theory of change about the ensure that costs are feasible at scale best way to get there.54 For example, as well as identify the core components an endgame could be government that are integral to the success of the adoption of a practice, and a strategy learning intervention, while adapting Ensuring participation of local to get there could be through advising other pieces based on the local context. the Ministry of Education on a curriculum Lastly, elevating teachers and leveraging community reform. It is worth noting that not every community expertise have been an effective intervention should scale; some important approach to scaling in low- Beginning in the 1970s in response to to the needs of the poor.55 As Jeffrey are effective by the very nature of their resource environments. criticisms of what were seen as top- Bradach, partner and co-founder at small size. But if scale is the objective, down approaches in development, The Bridgespan Group, writes, “success then experience in education and other participatory approaches to local of transformative scale often hinges sectors shows that it should be carefully development increasingly gained on involvement of local communities planned from the start. Going to scale must be attention and importance. The underlying in formulation and implementation of designed for from the outset. premise was that enabling communities solutions.”56 Participatory approaches The case studies reviewed and additional and citizens to define priorities and not only ensure more appropriately evidence examined point to the following This does not imply limitless engage in the design of a program designed interventions but also generate actions that are central to designing for scale, but rather, a clear vision or policy would ensure a process that the buy-in and create champions needed scale from the outset. First and foremost, of what the endgame is. was more inclusive and responsive to support and sustain scale.

56 57 According to the Monitor Group, “People for business in Kenya without uniforms in living at the base of the pyramid should be an attempt to keep the school affordable seen as customers and not beneficiaries; for families living under the poverty line. It Box 6. Fundación Escuela Nueva they will spend their money, or switch quickly learned, however, that communities livelihoods, or invest valuable time, only if wanted uniforms because they bestowed a Fundación Escuela Nueva, a nongovernmental organization behind the they calculate the transaction to be worth sense of identity and pride in the students, Escuela Nueva school model, works to improve the quality of rural primary their while.”57 For example, even though regardless of the additional expense. and lower secondary schools. The model emphasizes student, teacher, students in SAT’s alternative secondary school administrator, and community participation in its active, cooperative, schools, known as “centers” in Honduras, Escuela Nueva, started by the not-for-profit and personalized learning model, by which students teach each other scored higher on tests than students in organization Fundación Escuela Nueva, is and themselves, while teachers and administrators act as facilitators and traditional schools, some parents were an alternative primary school model that coaches, respectively. Established in Colombia in 1987, Escuela Nueva grew initially resistant to sending their children, began in rural areas of Colombia in the out of a movement to transform the quality of education in remote, public as they believed alternative education to mid-1970s and today has reached 5 million primary schools that dates back to the 1970s. Since then, Escuela Nueva has be an inferior form of schooling.58 Over students in 16 countries, from Honduras been adopted as a national Colombian policy and has scaled to 16 countries, time, once families could see the results to Vietnam to India.59 It effectively tapped reaching more than 5 million students. According to an independent study, of the program, including university into the demand among parents in students in grades three and five in Escuela Nueva schools in Colombia admission exam scores, they became marginalized communities to have their have achieved higher language and mathematics scores, with significantly supportive and the stigma disappeared. children enrolled in schools that foster lower dropout and repetition rates, in rural schools with the program than in

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions Bridge International Academies, a forlearning and development. It did this by traditional schools without the program. profit company that owns and operates emphasizing participation—of students, low-cost private schools, originally opened teachers, school administrators, and the

broader community. In the Fundación family feedback and preferences, aided Box 5. Bridge International Academies Escuela Nueva approach, the curriculum in part by innovative mechanisms, such is designed to enable students to pace as a 24-hour customer care hotline that Bridge International Academies is a for-profit private school chain founded themselves and to foster collaboration receives more than 2,000 calls a day. with the aim of providing students with high-quality pre-primary and primary through “cooperative learning” techniques. Bridge and Fundación Escuela Nueva education at a low cost. Each school is designed to ensure that children master Fundamental to Escuela Nueva’s model demonstrate that, regardless of the the national curriculum’s content while developing social and collaborative are its learning guides, which enable financing structure, any effective large- 21st-century skills. Local teachers are recruited and then trained on classroom self-paced, self-directed learning in a scale education program must tap into management, academic coaching, and lesson delivery. Each subject’s teacher structural pedagogical way. Teachers local demand and needs. guide is digitally published to a tablet, which is also used to collect teacher and students validate these learning Other examples from Asia to Africa behavior and pupil assessment data, driving continuous improvement in the guides for relevance and usability. The to Central America also highlight how program’s design. Bridge started in Kenya in 2008 and since then has scaled to school calendar is adjusted to take into initiatives that are designed to be relevant Uganda and Nigeria. To date, Bridge has reached more than 100,000 students account that children’s learning was being to the educational needs and interests of and employed 8,000 teachers in over 450 nursery and primary schools. Bridge disrupted because of seasonal demands low-income communities can scale—and partners with governments and NGOs to strengthen education systems across to work in the field. how they fail to do so when they have the world, including a recent partnership to operate public partnership schools Bridge International Academies designs not adequately engaged or understood in Liberia. According to a 2013 Early Grade Reading Assessment/Early Grade its program based on vast amounts of community demand. Math Assessment in Kenya, Bridge students gained the equivalent of 252 more data collected via various methods from days of reading education and 288 more days of mathematics education than community members, students, parents, One of BRAC’s initial projects in the early their peers in neighboring public schools over the course of 26 months. and teachers. Constant revisions are made 1970s, adult literacy classes in a remote based on student results and teacher and rural area in Bangladesh, was considered

58 59 to be a “dismal failure.”60 It built village something similar for their children. Thus, centers in northeast Bangladesh, where BRAC’s Non-Formal Primary Education villagers could come in the evening to (NFPE) program was launched, and it Box 8. Aflatoun International take free literacy and numeracy classes. has become the world’s largest private, BRAC candidly shared why after 18 secular school system,61 accounting for Aflatoun International aims to build children’s personal, interpersonal, financial, months, only 5 percent of the original roughly 5 percent of all primary school and entrepreneurial skills through social and financial education programs in 5,000 villagers still came to the centers: enrollment in Bangladesh, with formal formal early childhood, primary, and secondary school settings, as well as in because the idea originated from BRAC recognition by the government. nonformal settings. Taking a social franchise approach, Aflatoun International senior management and not the villagers. works through local partners that train teachers and trainers to educate BRAC had assumed that villagers would BRAC learned this lesson once again children, both those who are in school and those who are out of school, about want to come together to learn to read when designing its Empowerment and their rights, financial savings, and how to start enterprises. Established in India in and write, but after a long day’s work, Livelihood for Adolescents (ELA) program, 1991, Aflatoun is engaged in 116 countries and reaches almost 4 million children people wanted to learn things that which is active in five sub-Saharan and 34,000 teachers, facilitators, and peer educators each year. Recently, were more immediately usable in their African countries. In South Sudan, the Aflatoun and its partners have been working with governments in 28 countries daily lives. BRAC conducted a survey program sought to serve young women to integrate social and financial education into national curricula. Findings to identify the villagers’ actual concerns 15 years and older, in conformity with the from a systematic review and meta-analysis of 21 randomized controlled trials and revised the classes to teach more World Bank’s Adolescent Girls Initiative, on financial education interventions targeting children and youth found that relevant subjects, such as animal an economic opportunities program that Aflatoun’s effect on financial behavior, or savings and resource allocation habits,

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions husbandry, health, nutrition, and child provided funding for ELA in that country. was more than double the average effect across the 26 interventions studied. care. The reintroduced classes became However, BRAC found that in order to so popular that parents asked for achieve its objectives, including reducing

child marriage and promoting sexual settings in 116 countries, has partnerships and reproductive health and rights, ELA in three Francophone African countries Box 7. BRAC, Non-Formal Primary Education needed to reach girls earlier, at ages 11 to with national teacher unions whose 14, and revised its program accordingly. interest is to improve the relevance of the social and financial education curriculum BRAC’s Non-Formal Primary Education (NFPE) program provides children who Often overlooked is the role and interests and increase the training available to dropped out of or never enrolled in primary school with an opportunity to learn. that teachers themselves have in teachers in those countries. To date, this It condenses the general, five-year primary school curriculum into four years, improving the educational systems where has resulted in curriculum integration in allowing students, upon completion, to take national secondary school exams they work. Aflatoun International, a social Togo and Niger and signed agreements and transition into public or private secondary schools. The program is taught by and financial education program in early to move forward with the process at the local teachers, recruited from the community (who usually stay with each cohort childhood, primary and secondary school secondary level in Cameroon. of children for the entire four years), and focuses on cognitive subjects, such as Bengali, English, mathematics, science, and social science, and noncognitive subjects, such as arts and communication. BRAC was launched in Bangladesh in the early 1980s, and over time, its NFPE program has expanded to Afghanistan, Pakistan, Philippines, and South Sudan. Today, BRAC NFPE reaches 670,000 Strengthening accountability to respond students in Bangladesh and over 900,000 students in BRAC’s other four NFPE countries each year. The organization works closely with the Bangladeshi to local education needs government, and many consider it to be the largest private school network in the world. BRAC’s dropout rate is below 5 percent, and its pass rate routinely Improving accountability between As Linn argues, “The longer the chain surpasses government school rates in Bangladesh. education providers and clients (students, of accountability between design and families, and teachers) is another important ultimate beneficiaries, the greater the strategy for tapping into local demand. likelihood that interests will diverge.”62 The

60 61 World Bank highlights this relationship in or in-kind donations such as land use, Camfed’s model, a secondary education are under way within the public system its triangle of accountability. Most public construction materials, or books. This up- program for rural marginalized girls living to shorten the route of accountability sector production uses the “long route” front commitment generates a different in five sub-Saharan African countries, has between the state and citizen, such as the of accountability, whereby citizens or set of expectations for project outcomes, accountability at its core. Its programs Punjab government’s Education Sector clients must rely on political processes creates a real sense of partnership, and begin with a focus on the individual girl Reform Program (PESRP) in Pakistan, an to leverage their voices and influence increases the likelihood that projects will in rural Africa as its client—looking at the initiative supported in part by assistance the state’s compact with providers rather not only be implemented on time, but world from her perspective and identifying from the United Kingdom.65 The program than the ability to hold service providers that communities are more inclined to the barriers that must be overcome was created to complement the Punjab directly accountable.63 sustain project activities themselves for to ensure she can receive a quality Education Reform Roadmap, with the the longer term. It opens the door for and empowering education. Camfed intention of improving the province’s Many cases reviewed used a community a strong community interest in learning approaches community members as quality of primary and secondary co-investment model as a strategy the details of programmatic activities, “shareholders,” ensuring that all systems education. To shorten accountability for gauging community interest and providing input and feedback about the and processes are accountable to them. between the government and its citizens, promoting ownership and accountability. efficacy of overall program designs in the While many critics said that this approach PESRP established approximately As part of any new project activity, community and accountability for the could not scale because of its focus on the 56,000 school councils to promote Room to Read, an NGO focused on use of resources. Room to Read staff in individual girl, Camfed found that it was community participation and parental improving literacy and gender equality Tanzania reported community leaders actually that specific angle that allowed involvement in their designated Punjabi in education, requires a community losing their jobs in elections because the project to scale and to do so at pace, public schools. Through this process, local

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions contribution to the overall effort of they had not been accountable enough achieving sixfold expansion in the number citizens are empowered to monitor school approximately 15 percent of total costs. for the community funds pledged toward of girls receiving support in Malawi, performance, while school councils are This can be in the form of cash, labor, school-building construction.64 for example, from 2,000 to 12,000, granted annual budgets to respond to within three months. This approach to community needs.66 scale is premised on its governance model, refined over the past 23 years, Scaling effective learning practices or that places accountability to girls at its approaches requires being responsive core and dovetails with local education to local education needs, in particular Box 9. Room to Read, Literacy Program systems. This has enabled it to scale a those of students, teachers, and parents. needs-based financing mechanism that It also requires accountability not only to Room to Read is a nongovernmental organization focused on improving mobilizes extensive local infrastructure the government and external partners literacy and gender equality in education. Its Literacy Program uses a co- and assets to address girls’ school-going but most importantly to these “end- investment approach to work in collaboration with local communities, partner requirements, including provision of school users.” Local community participation organizations, and governments to ensure that schools have a structured fees and other education costs alongside and ownership of a program or policy are library with books in the children’s local language and that teachers and learning assistance and psychosocial also essential to sustain an intervention librarians are trained in evidence-based reading and writing instructional support. Camfed scaled its model from for the long term. While being responsive methods. Literacy coaches are placed in classrooms alongside teachers to reaching tens of thousands of girls to and inclusive are necessary conditions for provide instructional support. Together, these intervention components hundreds of thousands of girls within just scaling an effective learning intervention, aim to develop literacy skills and a habit of reading among primary school a few months, supporting them through a alone they are insufficient. At times, more children. Since it was founded in Nepal in 2000, Room to Read has expanded full cycle of junior secondary school. information is needed to demonstrate to nine other countries across Asia and Africa and overall, benefited 10 million and persuade communities of the benefit children across 17,500 communities. More specifically, its Literacy Program This is relevant even beyond non-state of education. It also requires some of the has reached 110,000 students and 2,000 teachers to date. Compared with provision of education. Experiments following key ingredients. an evaluation of 70 educational interventions in low- to upper-middle-income countries (of which Room to Read was not a part), Room to Read’s impact is nearly 10 times that of the other interventions’ average impact.

62 63 and what research has been carried out provide unit costs of interventions are 2. Cost-effective learning: has been hampered by the scarcity of usually unavailable for a given context. This relevant cost data reported in the public leads to underestimating or overestimating Cost structures affordable at scale should domain.”71 A review of scaling in nutrition an intervention, which can have a negative be incorporated in the design. found that detailed costing studies that impact on efforts to scale.72 Ensuring the right costs—for whoever is taking it up, the government or parents—is another crucial component of Understanding what is relevant and write, “scaling requires both addition and successfully scaled interventions. desired by students and parents is crucial subtraction . . . the problem of more is to designing an effective intervention that also a problem of less.”69 harnesses community demand. However, if the costs of the intervention are too high A surprisingly large number of pilots Balancing cost and quality for a government to adopt or for an actor focused on improving children’s learning to replicate at scale, then no matter how pay scant attention to costs. In a review Few would argue with the logic of finding example, the organization went through aligned it is to community needs, it will of cost-effectiveness analysis of education a cost structure that allows scaling, a controversial change in its literacy not scale. Ensuring the right costs—for and health interventions in developing including scaling that reaches the most instruction approach. Previously, it had Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions whoever is taking it up, the government countries, McEwan discusses how impact vulnerable young people, which is often a used a large number of manipulatives in or parents—is another crucial component evaluations cannot always inform resource more costly endeavor.73 But the question its classroom activities. These included of successfully scaled interventions. allocation decisions unless the costs of is how this can be achieved without flash cards, literacy wheels, dice with interventions are considered alongside sacrificing quality. words on them, small series of six- to 10- Too often, interventions are designed at their effects. He goes on to argue that “cost- page decodable texts, and individual the outset solely for effectiveness and effectiveness analysis is a straightforward First and foremost, it requires student writing books. Teachers reportedly not for the efficiencies that are required but underutilized tool for determining understanding from the outset the longer- enjoyed these resources because they for scale. This leaves “the bridge to scale which of two or more interventions term cost implications of sustainably were fun and broke up the monotony of too far to cross.”67 As Ian Thorpe writes provides a (non-pecuniary) unit of effect at scaling—and delivering at scale—based the school day. However, the diversity of in his blog Knowledge Management on least cost.” He finds that “across multiple on sound cost projections.74 This requires materials was expensive to produce, and a Dollar a Day, many pilots start with sectors, education and health projects identifying the incremental cost of all preparing activities and switching among “a kind of loving attention and specific are, by far, among the least likely to report resources (i.e., personnel, facilities, them took too much of the school period. starting conditions that couldn’t easily results of a CBA [cost-benefit analysis] at materials) incurred by all stakeholders Room to Read has therefore consolidated be replicated.”68 If you begin with the project appraisal.”70 In a review of scaling (i.e., schools, government, householders) its classroom materials into a single, luxury, gold-plated model, it becomes in development more broadly, Hartmann for the development, implementation, nondisposable student book per grade very difficult to determine which pieces to and Linn stated that “research on the cost and maintenance of an intervention as per country. It includes the same engaging remove—either because psychologically implications of scaling has been limited and after it scales.75 At the same time, as activities, but they are relatively less people come to expect them, or Laurence Chandy, a fellow at Brookings, expensive to produce on a per-child basis. pragmatically it is difficult to determine and colleagues write, there are real These books, which can be used over which elements are responsible for driving challenges in developing cost projections, multiple years, are much closer to a price improvements. Even if a program is highly Too often, interventions are as scaling involves changing cost curves, point that can be absorbed by Ministry of effective, if it relies on too many resources designed at the outset solely for altering beneficiary behavior and the Education budgets over time. to reach too few children, the approach effectiveness and not for the policy environment.76 it uses holds little promise of spreading efficiencies that are required for This same logic explains why Fundación and being taken up by others. As Robert Room to Read considers cost-effectiveness Escuela Nueva has, despite multiple offers Sutton and Huggy Rao, professors at the scale. This leaves “the bridge to as one of the core design features in by technology companies, very cautiously Institute of Design at Stanford University, scale too far to cross.” its program development. In 2014, for analyzed the value add and implications

64 65 of integrating tablets or mobiles for part of a larger donor-sponsored report community resources, many of the cases and avoided creating parallel structures teachers or students into classrooms on BRAC’s NFPE program, a team from a also creatively leveraged community by trying to rejuvenate and optimize where the organization works.77 Its point well-respected national accounting and members’ expertise to both unburden the existing but underutilized positions is that a pedagogical transformation market research firm undertook a small teacher, help enrich students’ learning and structures within the government. at the classroom level needs to happen survey to compare the private and public experience, and ultimately enable a Read India’s teacher-led summer camps first for technology to meaningfully and cost of rural government primary schools lower cost structure. In many instances, energized existing cadres of Cluster effectively impact learning. In addition, and BRAC NFPE centers in Bangladesh. this double or triple win helped reduce Resource Center Coordinators (CRCCs)81 the cost-effectiveness of the model has The study concluded that the public cost expenses while also bringing a level of to oversee teaching and learning activities. to be guaranteed. Adding even relatively per enrolled student was on par with energy, commitment, and authenticity to Traditionally, CRCCs were considered inexpensive technology could increase the the formal and nonformal BRAC schools the program. fit for only routine administrative and cost per child and make it harder to scale but that the high dropout rate from regulatory tasks. However, Pratham’s across contexts, including communities formal schools resulted in public and Creatively using attempt to institutionalize Read India with limited financial and human resources. private costs per third grade completer government infrastructure: through CRCCs revitalized their roles. more than four times the cost per NFPE Maximizing the effectiveness of positions Pratham trained CRCCs for four days and Detailed cost projections are partly how completer.80 and buildings within the government can then had them practice in the field for 20 Bridge International Academies could help keep costs low, for both government days. After that, Pratham trained teachers expand as rapidly as it has. It prioritizes Cases reviewed demonstrate ways that and nongovernment actors. For example, in the new methodology of grouping and the quality education it expects at a costs were kept low without sacrificing Pratham’s Read India, a remedial teaching at the right level. In this way,

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions particular price point. For example, quality. This includes through leveraging education program that helps children in Read India’s teacher training activities Bridge expects to provide r esults that technology, community engagement, grades three to five learn reading, writing, were less radical, more cost-efficient, and are as good as or better than those of existing government structure, and and basic math, intentionally kept costs low more easily replicated. neighboring public schools at a price that experimenting with cost-recovery measures. families living on less than $2 per day can afford. According to Bridge, the average Leveraging technology: household income of its pupils is $136.22 In many of the examples reviewed, per month, with an average of four or five technology played an important role in members per household, meaning that reaching a scalable cost by generating Box 10. Pratham, Read India 4.4 percent of a family’s average income efficiencies, by automating work and is spent on each child’s education.78 optimizing the use of resources as Pratham’s Read India initiative works to ensure that children in grades three Maintaining this average price of $6 per programs and policies scaled. Bridge to five acquire the ability to read simple text and do basic arithmetic. This month has driven the team to be ruthless International Academies has leveraged approach involves identifying children’s current learning levels, regardless about driving down costs.79 Any proposed technology through a smartphone with a of age or grade, and grouping them by level for instruction, an approach addition to the current model is translated customized Enterprise Resource Planning referred to as “teaching at the right level.” Pratham’s Read India program into the number of families that can no (ERP) to monitor all payments in and out of started in 2007 and today is implemented in 23 Indian states, having directly longer be served. the school—school fees, teacher and staff affected more than 350,000 primary school students directly in rural areas salaries, and so on—reducing financial from 2014 to 2015 and 6 million students indirectly through its partnerships Policymakers say that it is important transaction costs and allowing it to need with state or district governments. A series of randomized evaluations to not only have costs per unit but also only one nonteaching staff member at conducted by J-PAL of Pratham programs found significant gains in learning cost comparisons to the alternative each school. The role of technology will outcomes when children are grouped by level rather than by grade. Pratham to help channel resources to the most be discussed further below. also facilitates the Annual Status of Education Report (ASER), a nationwide effective intervention. At first glance, an citizen-led household survey that assesses basic reading and arithmetic intervention might appear expensive, Leveraging community levels of children in over 550 of India’s rural districts. ASER has been carried but compared with the alternative or to engagement and resources: out every year starting in 2005, and since 2008, began functioning as an the savings it will generate, it might be As discussed further below on elevating autonomous unit within the Pratham network. a great value for the money. In 1992, as the role of teachers and leveraging

66 67 Questions on cost recovery 3. Flexible adaptation: A number of the cases in which NGOs launched a technical assistance arm called Core elements of effective learning approaches should be did the scaling have been experimenting Room to Read Accelerator to share and with cost recovery efforts, with varying train other organizations and government identified and replicated across contexts while adapting degrees of success. These organizations partners that are interested in implementing the rest to local circumstances. rely heavily on philanthropic support and similar delivery models. Room to Read are interested in diversifying their funding Accelerator takes best practices from its base. For example, with Educate!, a core work, codifies them, and systematizes leadership and entrepreneurship training the approach, charging partners a fee program in secondary schools in Uganda, sufficient to recover its costs. The idea is Ultimately, even if a pilot effectively In the literature, striking this balance is its original financial plan included charging for this unit to not just be self-sustaining but meets educational needs with a most often associated with understanding 84 partnering schools along a sliding scale and to allow Room to Read to scale and serve scalable cost, it can face obstacles when what is essential to achieving impact. increasing the amount over time, starting at even more children in areas that it would not expanding because the design is either In discussing the spread and adoption $200 per year per school and eventually otherwise be able to reach through direct too rigid or too flexible. In the cases and of innovation, Anna Davies, professor at charging $600. This fee did improve implementation.82 literature examined, there appeared to Trinity College Dublin, outlines that the school buy-in, but Educate! found that be an optimal point between wholesale diffusion process should reflect replication replication and costly customization. The of the “core content” of an idea, rather

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions the time and effort involved in collecting Whether costs were kept low from payment was not worth the amount it was leveraging existing infrastructure, creatively key appeared to be flexible adaptation than exactly replicate every aspect of 85 charging. Therefore, it recently decided to tapping community resources, or utilizing of the model. This requires what Nesta the original. In Nesta’s interviews with 83 keep a nominal fee for partner schools to appropriate technologies, it appears that refers to as “identifying the core” — social innovators, this meant establishing demonstrate commitment but not as a cost there are a number of lessons to learn essential aspects of the model that must “what’s fixed and what’s flexible—in recovery mechanism for the program. Today, from low-resource environments. In these be maintained as it scales. The reviewed relation to the model, scaling routes, 86 more than 80 percent of its revenue comes contexts, actors did not have a choice but to case studies established an important goals and aims.” Often, the core is a from foundations. Room to Read recently keep costs low as they scaled. balance between those elements that practice, mission, or approach rather than are nonnegotiable and replicated across spreading a specific education model. contexts and the other aspects that are more flexible and should be adapted For SAT, the underlying philosophy of to each location. The challenge is developing a generation of socially Box 11. Educate! striking the right balance between local minded young people who can serve as adaptation and fidelity to the original engines of sustainable development in Educate! is a nongovernmental organization that addresses the mismatch between model. their communities is the core approach to secondary education and employment opportunities in sub-Saharan Africa, as well as a tradition of rote memorization and theoretical teaching methods, with the ultimate goal of teaching African youth to solve poverty for themselves and their communities. It provides students with skills training in leadership, entrepreneurship, and workforce readiness, along with mentorship to start real businesses at school, The reviewed case studies established an important and the model itself is delivered through trained teachers and youth mentors. balance between those elements that are nonnegotiable Educate! began in Uganda in 2009 and now reaches 120,000 students in 350 and replicated across contexts and the other aspects that schools, or 12 percent of Ugandan secondary schools, having advised the reform of Uganda’s upper secondary entrepreneurship curriculum and incorporated its model are more flexible and should be adapted to each location. into Rwanda’s national secondary school curriculum. Midline results from Educate!’s The challenge is striking the right balance between local randomized controlled trials indicate that its secondary students’ incomes are adaptation and fidelity to the original model. doubled after graduating secondary school and that they are 64 percent more likely to start a business and 123 percent more likely to initiate a community project.

68 69 developing its program and organization. cash transactions. This was also the case The Teach For All network comprises 39 with the scaling of Lesson Study across Box 12. Sesame Workshop, Sesame Street independent partner organizations that Zambia, a peer-to-peer teacher training share a mission and commit to certain practice. While it was important that Sesame Workshop’s Sesame Street uses television, radio, videos, websites, principles but have full autonomy in this Japanese-adapted practice be seen books, and social media to educate preschool-aged children on literacy determining how best to achieve them. as part of Zambia’s national teacher and math, emotional well-being, health and wellness, and respect and professional development reform, what understanding. The show began in the United States in 1969 and has since Of course, striking this balance between was key to its success was the existing expanded to more than 150 countries, reaching approximately 156 million universal replication and local adaptation countrywide network of Teacher Resource children and serving as the single largest informal educator of children in the can be hard, even when being deliberate. Centers with officers tasked with guiding world. Studies demonstrate that Sesame Street is an effective learning tool Sometimes, the elements that are and monitoring these new practices. for children with a positive impact across countries. A meta-analysis of its assumed to be core are not necessarily educational impact in low- and middle-income countries found an impact that what are behind a program’s success. The cases reviewed addressed the is comparable to that of other early childhood interventions, with scale being In an example outside the education tension between localization and scale its key distinguishing factor. sector, what was ultimately found to be by identifying core elements that were responsible for the exponential spread integral to an intervention’s success— of M-PESA, the mobile money platform be it an underlying principle, a piece of in Kenya, was not the mobile technology, technology, or an existing distribution

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions but the distribution network of local network. At the same time, there was every school in every country. In addition and many steps are taken to minimize the vendors who served as M-PESA agents flexibility to adapt the model or approach to sharing a common value, the curriculum, formal and informal costs of attendance. to register new customers and facilitate to the local context as needed. or 80 texts developed by the parent Parents are engaged regularly. organization in Colombia, Foundation for the Application and Teaching of the Sesame Workshop, the nonprofit behind the Science (FUNDAEC), is a core component children’s television show Sesame Street, Box 13. Teach For All that is shared across the countries. provides a framework for its international Beyond the curriculum, much of SAT’s Sesame Street co-productions that share Teach For All is an international network of local, independent organizations implementation is determined by the local a universal style, target age group, and that recruit and train recent top-performing graduates and professionals context. This includes when and where to whole-child approach for core learning to commit two years to teaching in their country’s underserved schools and meet, the number of hours per week to goals for that age group, but leave it to communities, with the goal of developing a pipeline of future education meet, and the coordination structure in local production teams to develop the leaders. Teach For All was co-founded by Teach For America and the United each country. series based on the needs of children Kingdom’s Teach First in 2007 and has since become a global network of 39 and the specific education goals of each country partners. The network has reached 1.1 million students and 52,323 While BRAC schools incorporate a country. teachers and alumni to date and has had an impact on both individual country’s national educational curriculum students and overall education systems. Studies from partner countries into its program, the classrooms look and Teach For All, an international network demonstrate learning gains, such as a Mathematica Policy Research study feel remarkably similar across the various of local, independent organizations that that found students taught by Teach For America math fellows demonstrated countries, as diverse as Afghanistan, recruit and train recent top-performing an additional 2.6 months of learning over the course of the year compared South Sudan, and Philippines. BRAC’s graduates and professionals to commit with students taught by novice and veteran teachers. At a systems level, Teach core components include hiring female two years to teaching in their countries’ For All has contributed to broader education reform and change through its teachers from the local community and underserved schools and communities, fellows and alumni, as 50 to 80 percent of alumni from most partners stay full supporting them with regular training. attributes much of its success to its time in the education sector. Low-income children, girls, and other approach of ensuring that each network marginalized youth are given priority, partner operates with full ownership for

70 71 teachers in 2015.93 Projections estimate for professional development and 4. Elevating teachers: that to provide every child with a primary preparation.95 education by 2030, 25.8 million teachers Community expertise should be leveraged to support will need to be recruited.94 A number of the cases reviewed for this and unburden teachers. study designed effective approaches that To address this shortage, many schools could scale quality learning by elevating have adopted a double shift system, the role of the teacher, leveraging expertise requiring teachers to work 12 hours that exists in the community, and providing per day, teaching two separate groups multiple pathways to bringing motivated Teachers are perhaps the most important in school.87 In a study of 15,000 teachers of kids back-to-back, leaving no time people into the teaching profession. actors in children’s educational across Latin America, learning outcomes experiences, especially for marginalized of students were less influenced by being children. Evidence from numerous studies in a good or bad school than they were shows that the quality of teachers by being taught by a strong or weak Activating community expertise and significantly influences children’s learning teacher.88 technology to unburden teachers

From India to Uganda, elevating to prepare and teach. On average,

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions Overburdened teachers teachers’ roles and attracting other since 2015, a cadre of 3,170 individual resources to support teachers, volunteers have delivered 2,500 sessions Teachers’ jobs are very complex, and hurdles. Many teachers have classes with whether through technology or each year, forming the largest network of around the world, teachers are being asked more than 40 students, surging up to through expertise that is outside of committed volunteers in Jordan today. A to do a wide range of tasks not directly 100 in some countries after adoption of the school in the community, has been byproduct of this collaboration has been related to their core job of instructing free universal primary education.90 Many a strategy used across multiple cases advancing a culture of volunteerism in children. For example, a typical day for a teachers are doing the best they can amid we studied. Room to Read, Educate!, Jordan. Beyond volunteering to teach, teacher includes instructing during school difficult environments. In some places, Read India, and INJAZ, a Jordanian private sector companies support INJAZ hours, preparing lesson plans, grading teaching positions are used by politicians nonprofit organization that provides by adopting schools, sharing information student progress, managing classroom as a form of political patronage and are entrepreneurship and work readiness and data, and providing employment and behavior, tracking and reporting data, assigned not to those who are motivated training for secondary and post- internship opportunities to graduates. communicating with parents, providing and trained to be teachers, but to those secondary students, are all examples of extra assistance to students falling behind, supporters for whom politicians need to rethinking who in the community can add Similarly, Educate! in Uganda has mentoring or providing social support for curry favor.91 This puts a heavier burden to students’ education without requiring leveraged community members to teach children with difficult family circumstances, on those teachers who are motivated to teachers to take on more responsibilities. Uganda’s entrepreneurship curriculum meeting with school personnel, buying try to compensate for their peers who Across the cases it varies whether these alongside teachers in schools, relying school supplies . . . and the list goes on. are not. In some post-conflict contexts, community members are volunteers or on entrepreneurs and employees from Some argue that education, and by teachers go months without salaries.92 employees, but in each case they are local businesses. To deliver the lessons, extension teachers, is being asked to solve In Liberia, a national survey of teachers not only adding support to teachers but Educate! primarily hires alumni of its a wide range of social problems.89 This after the civil war showed that most also building passion in communities to program who are young entrepreneurs view, that education is a panacea solution teachers—in some areas 90 percent—held prioritize education. themselves. Its model allows teachers to many social ills, places a very large, and multiple jobs (i.e., farming and tutoring in to be supporters of the subject, but as perhaps unfair, burden of responsibility on addition to teaching) to make ends meet. For example, INJAZ partners with private with INJAZ, relieves teachers of having teachers. In the developing world especially, there sector companies and trains its employees to get up to speed on new topics. This simply are not enough teachers. Recent who volunteer to teach the lessons. This strategy is coupled with teacher training In addition, in many developing countries estimates suggest a global teacher helps teachers by not burdening them programs, which are tied to long-term teachers face seemingly insurmountable shortage of 2.7 million primary school with yet another subject they have government education reform goals.

72 73 poor learning,” something Pratham’s CEO how technology is leveraged unburdens said no advocacy campaign could have them from the normally overwhelming Box 14. INJAZ, Jordan accomplished.96 list of tasks they face. For example, in Amazonas state in Brazil, Media Center, INJAZ, an independent Jordanian nonprofit organization, specializes in youth This is not to say that relying on a distance learning formal secondary empowerment programs and leverages the public, private, and social sectors community members and volunteers education program, is designed for two to help bridge the skills gap between the secondary school system and the is without challenges. While some types of teachers: a specialist lecturing changing needs of the labor market. INJAZ programs, delivered through a community members were paid and others teacher who communicates via satellite network of more than 23,000 trained volunteers, provide youth from seventh volunteered, their training was a crucial from a central studio, and a generalist grade to university level and post-graduation with content and activities element of the program’s success. Quality tutoring teacher who facilitates learning that improve their financial literacy, ethical leadership, teamwork, creative control and delivery standardization in the classroom. Lecturing teachers are thinking, communication, and interpersonal skills. For older youth in colleges can be difficult in these models. INJAZ, highly trained and research their subjects and universities, youth centers, and vocational training institutes, the majority for example, faced this issue, to which it of expertise, develop content for the of the programs offered provide direct links to real-world opportunities responded by implementing mandatory lectures, and reach thousands of students and support for them to gain work experience or otherwise develop their volunteer trainings at the beginning of at a time, whereas the tutoring teachers professional and entrepreneurial skillset. Since its inception in 1999, INJAZ each semester before volunteers can teach work directly in the classrooms to guide has scaled across all 12 governorates and has reached 1.2 million youth to a class. Read India also struggled with student learning, support classroom date. The program has been integrated into Jordan’s official school schedule, quality control and actually scaled back interactions, monitor student activities,

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions and, according to an internal study, INJAZ graduates had an unemployment after its first phase to work on its training. and perform administrative duties, without rate of 19 percent compared to the national rate of 32 percent. Today, whether it is community volunteers being responsible for instruction, content or teachers implementing the Read India delivery, or deep knowledge of any one program, Pratham requires a heavy dose of subject. training. For example, in Pratham trainings, Other programs reach out to a school’s Pratham has used a number of strategies participants leave the training location to parents and community members to help to implement its Read India model for go into nearby schools and “practice” the In cases from both Brazil and unburden teachers. Room to Read’s Girls’ teaching at the right level. Across all of them, techniques they have learned in the training Kenya, the way teachers’ Education Program, for example, trains government officials have played important sessions. In many cases, especially when roles are configured and educated, empowered women from the roles from sanctioning the experimentation Pratham is working with the government, how technology is leveraged community to be “social mobilizers” to with this approach to actively participating officials have to “practice” the Pratham work directly with the more than 30,000 or putting resources behind it. In some method in their schools every day for 15 to unburdens them from the girls in the program as their mentors, cases, community volunteers are trained to 20 days to build mastery of the approach. normally overwhelming list of counselors, and advocates. In addition teach children literacy skills in summer and tasks they face. to serving as excellent role models, these outside schools. In other cases, government With the right design and flexible women provide critical personalized teachers, Pratham staff, and community adaptation as programs scale, leveraging emotional guidance and life skills training volunteers work with students during the community members’ expertise is in many In other cases, technology has been that many teachers do not have the school day. This strategy is effective because cases an effective strategy for helping leveraged to split the traditional capacity to provide every girl in the it leverages community resources to help improve student learning, along with responsibilities of a teacher into those classroom, plus out-of-school engagement support teachers and to reach and teach supporting more traditional teachers inside creating lessons versus those teaching and home visits that can be unrealistic for students who are falling behind. As in the and outside the classroom. Governments lessons. For example, a core component teachers to undertake. This added support other cases, the community engagement that are flexible and help sanction, support, of Bridge International Academies’ model system improves girls’ school participation model had a dual benefit: elevating or sustain this approach are crucial in the is that teachers receive a daily teacher and life skills, and it has resulted in other and unburdening teachers to provide scaling process. guide with lessons via tablets, which is positive externalities, such as empowering remedial tutoring along with instruction, essentially a scripted lesson plan for them women and developing more female role and engaging 10,000 volunteers and In cases from both Brazil and Kenya, the to follow. By centrally developing all the models in the community. bringing them into the “battlefield against way teachers’ roles are configured and teacher and learner materials, teachers

74 75 personnel and using technology to take student progress and administrative on administrative tasks usually handled data-intensive tasks—versus those that Box 15. Amazonas State Government’s Media Center by teachers. In the future, Smith envisions cannot, such as classroom facilitation a further optimizing of the teacher’s role and student interaction. Interestingly, this Brazil’s Amazonas state government’s Media Center initiative is a locally where teachers specialize according to dynamic was almost the opposite when developed, formal secondary school model seeking to address the disparity their strengths in instruction, technology groups leveraged parent and community in education access between urban and rural areas. It employs digital provides real-time analysis of student workers rather than technology to satellite technology to deliver live lessons from “lecturing” teachers at the learning progress, and parents engage unburden teachers. From Pratham to Media Center studio in the capital, Manaus, to up to 1,000 classrooms more deeply with schools.99 Room to Read, it was the extra person- across Amazonas state, with “tutoring” teachers located in each classroom to-person support, interaction, and with anywhere from five to 25 students. This initiative allows for bidirectional These examples show how technology attention to students and their needs interactivity, meaning students can stream the teacher’s lecture at the studio enables scaling when it comes to that community members took on to and present information back, thereby appearing to all other classrooms and teachers. From Media Center to help alleviate teachers. Ultimately, all of to the lecturing teacher in the studio. Established in Amazonas state in Brazil Bridge International Academies, the cases examined focused on quality in 2007, Media Center’s 60 lecturing teachers and 2,200 tutoring teachers some approaches unbundle teacher interaction and instruction as central to have reached 300,000 students across 2,300 communities—approximately responsibilities into those that technology children’s learning experience even as 25 percent of secondary school students outside of Manaus—to date. The can enable them to scale—namely they often thought quite differently about Media Center model has been adapted to seven other states in Brazil to lecturing, lesson design, and tracking how this should be achieved.

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions serve difficult-to-reach populations. Since its establishment, lower to upper secondary school progression rates have increased, dropout rates have nearly halved between 2008 and 2011, and children’s learning in Amazonas state has steadily improved as reflected on the Brazilian Education Quality Index. Diversifying pathways to teaching

Given the different ways in which the shortages, hiring para-teachers from role of the teacher is configured across within communities has helped address are relieved of the burden of creating “unbundling” the role of the teacher.98 This the cases, it makes sense that there are the teacher shortage in remote areas, content and lesson plans and are able approach is beginning to be employed different pathways to becoming a teacher. and having teachers who are of the same to spend more time focusing on their by a range of programs, including one The traditional formula of completing demographic background helps close students’ progress. One Bridge teacher in inside the United States. For example, higher education, specializing during the “social distance” between teachers Nairobi expressed that the scripts gave Preston Smith, co-founder and CEO of pre-service teacher training, and then and students and ultimately improves her confidence and allowed her to focus on Rocketship Education, a network of U.S. teaching in the classroom is supplemented student learning.100 students who needed additional support.97 public charter schools for low-income by other routes. The approach varies in This approach also frees up teachers students, argues that elevating the role the different cases—inside and outside For example, SAT in Latin America does from a range of administrative tasks. For of teachers by unburdening them and communities, in and out of schools— not employ the term “teacher” but uses example, the tablets allow centralized data allowing them to focus solely on instruction but they exemplify the need for diverse “tutor” to signify a less hierarchical collection and analysis, facilitating central is an important factor in the program’s pathways into the teaching profession, conception of how the learning process monitoring of many aspects of teaching success. Effectively teaching children especially for underserved areas. unfolds. Tutors, who are often secondary and learning that would traditionally be in marginalized communities takes, he school graduates from the community the responsibility of teachers to collect and says, “a great deal of rigor.” The more The case studies are backed up by where they will ultimately be teaching, monitor. time teachers can focus on teaching and literature that finds that hiring teachers undergo intensive training in SAT content instruction, and the less time they spend from local communities helps close the and pedagogy, with a heavy focus on Unburdening teachers from the multiple on tasks not related to teaching, the cultural and linguistic gap between student-centered learning. Tutors have demands placed upon them, and finding better off both teachers and students will teachers and students, thereby improving neither the typical teacher profile nor the new ways to deliver some of the tasks be. Rocketship does so through several learning outcomes and enrollment. In typical teacher role. Tutors are meant to previously assigned to teachers, is known as strategies, including deploying other areas experiencing regular teacher learn along with their students, and they

76 77 stay with one cohort during their six years In activating resources from from one that emphasizes what teachers innovate in their classrooms, supporting of lower and upper secondary school. The local community members or lack (known as a deficit approach) to them and connecting them to a network of job of the tutor is to facilitate students’ recognizing the assets teachers bring other innovative teachers to collaborate. active engagement in SAT lessons and technology—either directly or and giving them the space to build on STIR has found that good leadership and actively develop their own skills as they through opening up additional these assets. Teachers at school are able recognition from local communities are go by participating in 10 days of training pathways to becoming a to produce their own ideas to improve more important than salary or training for every three months. These tutors are teacher—ensuring that their lessons and are empowered to be motivating teachers. Preliminary results actively involved in the communities and the main agents, with support from peers, have shown promise—in some Ugandan often in the students’ lives as well, building teachers are supported and in their own training. schools, teachers were four times as close relationships and spending time respected is an important likely to arrive on time after completing with students and their families outside element of lifting some of the A new approach initiated in India the STIR program.101 of lessons. In Honduras, the government burden from their shoulders. reinforces the importance of this has adapted to the community needs respect. STIR Education, which focuses In conclusion, activating resources from and supported this new pathway for on improving teaching, has worked outside of the school and developing becoming a teacher by agreeing to pay with more than 12,000 teachers in new pathways to do so sustainably can the tutors’ salaries. partner organizations operate, providing India and Uganda, identifying teacher play an important role in unburdening a mix of in-kind or financial support to “changemakers” who are inspired to teachers and helping to improve learning. Teach For All takes a different approach. help develop and sustain this additional

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions Instead of leveraging members of the pathway to becoming a teacher. Initially community, its network is built on the this approach may help fill a teacher idea that countries should be channeling shortage in an underserved community, much more of their top talent toward but the organization’s ultimate goal is to improving education for vulnerable build a movement of talented individuals children. Partners in the Teach For All who become lifelong educators, network diversify the pathways into innovators, and advocates for expanding teaching by recruiting and training opportunities for marginalized children. outstanding university graduates and Across their network partners, more young professionals, many of whom than 60 percent of their alumni stay did not plan or study to be educators, within education and many others work to teach in marginalized communities. on issues related to education and low- Again, the focus is much more heavily on income communities from sectors such as in-service rather than pre-service training policy, medicine, and law. and the “fellows” in the Teach For All organizations around the world receive In activating resources from local the bulk of their training after being community members or technology— deployed to a school and classroom. A either directly or through opening up global network of Teach For All fellows additional pathways to becoming a and alumni supports sharing of lessons teacher—ensuring that teachers are learned, and training opportunities are supported and respected is an important available throughout the two years that element of lifting some of the burden from fellows are in the program. their shoulders. The experience of the Zambian government in Lesson Study is Governments are almost always crucial a powerful reminder of the importance of partners in this effort and across the 39 respect. In the initiative, the government countries where Teach For All independent has flipped its teacher training approach

78 79 Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions VI DELIVERY

80 81 5. Education alliances: Introduction All actors need to work together to achieve a common goal.

“Social change depends, in other words, on alliances between what could be called the ‘bees’ and the ‘trees.’ The bees are the small organizations, individuals and groups who have the new ideas, and are mobile, quick and able to cross-pollinate. The trees are the big organizations—governments, companies or big NGOs—which are poor at creativity but generally good at implementation, and which have the resilience, roots and scale to make things happen. Both need each other, and most social change comes from alliances between the two, just as most change within organisations depends on alliances between leaders and groups well down the formal hierarchy.” Learning champions Windows of — Geoff Mulgan, Nesta103 Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions DELIVERY and leaders opportunity

Although partnerships are notoriously youth—not just those in the right place at 5 6 7 8 9 difficult and some evidence shows they the right time—are learning well. are not always effective, across the 14 cases we reviewed, they were essential Partnerships contribute to scaling Education Technological Better data in bringing all the right skills sets to quality learning by pooling resources alliances advances the table. The enormous challenge of and ensuring an appropriate division of ensuring quality learning for all in the labor.104 The role of partners may include 21st century requires bringing all actors and assets to bear—both the bees, such as social innovators that can experiment The enormous challenge of Designing for scale is half the battle. At resources and skills of various partners and nimbly cross-pollinate, and the ensuring quality learning for trees, such as government agencies, the end of the day, however, whether or for large-scale gains in learning; all in the 21st century requires not an education initiative successfully cultivating “learning” leaders and which are essential for any education and sustainably scales depends as much champions both within and outside of the effort, including reaching those most bringing all actors and assets on how it is implemented. As Chandy and government and classroom; seizing key disadvantaged, to spread nationally. It is to bear—both the bees, such colleagues write, “Delivery is what makes moments more open to spreading and rare for any one actor to encompass all as social innovators that can the skills required to scale sustainably and getting to scale not merely difficult but adopting new ideas to improve learning; experiment and nimbly cross- complex.”102 As a result, they identify utilizing appropriate technologies to equitably. Even when governments are the difficulties of scaling as a process drive efficiencies and overcome context- leading the way, as they are in some of the pollinate, and the trees, such challenge. Through a review of cases specific barriers to learning; and using cases we reviewed, they often reach out as government agencies, which and literature, our analysis found that a range of data to continuously drive to civil society, communities, or the private are essential for any education sector to bring in key expertise they are delivering at scale requires a combination improvements in programs and policies effort, including reaching of both technical and political strategies. as well as to motivate and sustain action lacking. Certainly, however, governments These include leveraging the diverse in support of scaling quality learning. are essential in ensuring all children and those most disadvantaged.

82 83 bringing financing, providing technical While not exhaustive, these elements are that creates an emotional connection Best results come when all assistance, or generating political interdependent and include: establish a with its target audience of preschool- elements of the teaching- support. Cases and literature reviewed clear, shared goal; align incentives toward aged children. Sesame Workshop’s long- identified some of the key elements that meeting this goal; ensure accountability term success has depended on a broad learning ecosystem prioritize contribute to partnerships working to to the learners; and build trust among base of partners, including funding common learning goals and align enable a scalable approach to delivery. partners.105 agencies, local production partners, themselves to achieve them. broadcasters, government ministries, the education and academic communities, Cases and literature reviewed identified some of the local NGOs, and other private and be television, radio, print, or other media, key elements that contribute to partnerships working public partnerships. Sesame Workshop’s and researchers represent the voice of to enable a scalable approach to delivery. While not content creation model is a collaborative the child and provide information about exhaustive, these elements are interdependent and process among producers, educational the program’s effectiveness. While local content specialists, and researchers. needs and contexts can differ around include: establish a clear, shared goal; align incentives Producers and writers are responsible the world, Sesame Street partners share toward meeting this goal; ensure accountability to the for educational content and the creative the objective of using media to meet the learners; and build trust among partners. elements of the production, whether it critical needs of young children.

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions Establish a clear, shared goal Align incentives

Interventions that have effectivelygovernment systems is that the best results Effective partnerships for scale require Demonstrating intermediate results can leveraged partnerships for scale have come when all elements of the teaching- aligning incentives so that time, skills, also help keep actors engaged so that focused on addressing a binding constraint learning ecosystem prioritize common knowledge, and efforts of multiple they see the benefits to partnering. that is present in multiple contexts, such learning goals and align themselves to individuals are channeled in ways that as a lack of high-quality early childhood achieve them. It is essential that learning produce jointly valued outcomes.106 A For example, in 2003, the Naandi educational opportunities, a shortage goals are clearly articulated in a way that focus on shared outcomes is a good Foundation, in partnership with the of qualified teachers, or a mismatch teachers can understand and that they are way to align incentives. Hartmann government of Andhra Pradesh, between labor demand and supply. achievable. If training, academic support, and Linn identify incentives as a key to established a midday meal program in Successful interventions also approach materials, teaching-learning methods, and drive behavior of actors and institutions response to the Supreme Court of India’s these challenges, many of which are measurement are all aligned to support toward scaling. They discuss how in the decree that all regional governments incredibly complex, with simple ideas each other, then the chances of success absence of a profit motive in education, must provide public school children that are understood and resonate with are high. other incentives are needed to substitute with daily lunches. The program was a broad audience—students, parents, for market forces. These can include initially designed to provide nutritious teachers, government officials, and any Sesame Street, the world’s largest rewards, competitions, evaluations, and and hygienic food to 150,000 children other stakeholders. In fact, the greater informal early childhood educator, began pressure through political processes.107 in Hyderabad, free of cost. By the end and more complex the challenge, such in 1969 with a well-founded, simple idea Worldreader’s approach to scale, which of 2013, as a result of additional state as addressing a country’s poor learning to address the inequity of children’s provides digital books on low-cost government partnerships, the program outcomes, the more important it is to have school readiness, particularly among e-readers and mobile phones, has been was feeding 1.1 million children across a clear definition of the problem being disadvantaged children, by combining to identify “non-exclusive and growth- 10,453 schools in Andhra Pradesh, targeted, shared outcomes to achieve, and entertainment with education. More than oriented partners” where interests align.108 Chhattisgarh, Madhya Pradesh, a strategy to address it. 40 years later, Sesame Street reaches This includes publishers interested in Odisha, and Rajasthan. Through these more than 156 million children in over expanding their market share with partnerships between the Naandi One of the main learnings from Pratham’s 150 countries by adhering to this same digital content and cellphone carriers Foundation and regional governments, 15 years of experience in partnering with principle of developing engaging content interested in increasing customers’ usage. Naandi was able to function with

84 85 government subsidies in the form of free private sector organizations, such as the risk taking, and long-term commitment accountability for girls’ welfare; these are grain, operational funding, and kitchen Global Alliance for Improved Nutrition, required to scale. Rather, accountability girls’ entitlements, for which communities space, while regional governments, in Britannia Industries Ltd., and Faith Foods, systems for scaling quality learning are responsible. In turn, Camfed ensures turn, were able to fulfill their commitment as well as social sector organizations, such should include feedback loops that that communities have full access to to the Supreme Court’s ruling that as PATH, to add nutritional diversity to allow for rapid learning and adaptation. program data, including on how well “cooked midday meals” are a necessary midday meals.109 Ultimately, these public- girls and schools are doing, so that component of children’s education. In private partnerships helped the Naandi Camfed, a secondary education program they can be responsive and informed in addition to public sector partnerships, Foundation develop a very efficient for rural marginalized girls living in five their approach to tackling the obstacles the Naandi Foundation partnered with process for addressing malnutrition. sub-Saharan African countries, flips the girls face. This approach to governance premise of bottom-up accountability to and accountability has been critical one of bottom-up decision-making and to scaling Camfed’s program and top-down accountability. Communities impact, and it challenges the common are empowered with control over perception that community participation Box 16. Worldreader resources and determine how these and efficient, accountable management can best be deployed to tackle local are incompatible in the transition from Worldreader provides culturally and linguistically relevant digital books to obstacles to girls’ education—obstacles small single-community initiatives to the developing world through low-cost e-readers and mobile phones. It aims that they identify and understand large-scale, multi-community or multi- to improve literacy skills and instill a love of reading by providing children firsthand. This reinforces a sense of country programs.

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions and their families with immediate access to teaching, learning, and reading materials. The organization uses an integrative approach that combines context-appropriate technology, access to more than 31,000 books in 43 languages, teacher support, and community engagement. Since its first Build trust programs launched in Ghana in 2010, the organization has reached in excess of 5.6 million people, and as a result of the program, 1.1 million people in 69 At the core of these partnerships is over program implementation and the countries are reading digital books every month. Data from an evaluation trust. This has proved to be especially freedom to advocate for curriculum of Worldreader’s early grade reading program in Ghana demonstrate important during the startup phase. In the integration at the national level. Trust improvements in oral reading fluency and reading comprehension, as well as case studies reviewed, initial investments can be one of the most important assets the development of positive reading habits, among student users. in the idea were crucial—and these were for a partnership, particularly in contexts generally based on the investor’s trust where there are few guarantees, formal in the founder. As one early investor contract law is rarely enforced, and in Bridge International Academies infrastructure is limited. shared, “we bet on the jockeys, not the horse.” This was the case with Pratham While government remains the primary Ensure accountability to learners getting started more than 20 years duty bearer of ensuring the right to ago with an initial investment from the a quality education for all, improving Successful alliances for scaling quality learning, it is not just accountability former chairman of the Industrial Credit learning at scale requires the combined learning ensure that each partner is held as to whether a program or policy is and Investment Corporation of India, efforts of many actors to leverage to account for mutually agreed upon achieving good learning outcomes, but DfID’s early investment in Asociación diverse resources, capacities, and skills— results. According to Hartmann and also whether it is creating the conditions Bayán in Honduras in the 1990s, and particularly related to financing, service Linn, a functioning accountability system for effective scaling of successful early and long-term support of INJAZ delivery, and knowledge generation. includes three elements: availability interventions.110 At the same time, such in Jordan by the United States Agency This requires the interaction of state and and use of information, mechanisms for an accountability system must not lead for International Development (USAID). non-state partners at various levels and monitoring performance, and existence to a rigid compliance system of tracking Aflatoun International defines its social across various sectors, such as education, of adequate incentives for compliance. milestones against inflexible models that franchise as a “trust-based” network health, nutrition, workforce development, When it comes to scaling quality might undermine the experimentation, that gives partners local autonomy and economic growth.

86 87 6. Learning champions Political champions In many of the cases reviewed, a co-creation of an initiative, rather than and leaders: government champion at a national or bringing them a fully baked, evaluated local level was found to be the linchpin model to roll out across the country. STIR As scaling quality learning is a political and technical exercise, behind experimentation and greater Education, which is catalyzing a teacher- champions within and outside government and the classroom participation in policymaking. In the case led movement to reclaim teaching and of INJAZ, the entrepreneurial training improve students’ learning, found success are crucial. program in secondary schools that has in working with governments in Uganda been integrated across all 12 directorates and India to co-create the initiative rather in Jordan, Her Majesty Queen Rania Al- than presenting an already developed Abdullah provided early support that model for adoption. was instrumental in providing INJAZ with Scaling quality learning requires to create or lead an initiative, champions credibility and access to a wider network. Political champions need not reside at the champions and leaders at all levels— with political will and capital to scale Additionally, government support, national or state level. Sometimes, it is more inside and outside the classroom and and sustain an idea, and local leaders specifically INJAZ’s partnership with effective to focus on policymakers who education ecosystem. Addressing who might not be as well known but Jordan’s Ministry of Education, has been are closest to the problem—officials who learning challenges at scale requires whose everyday efforts are contributing

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions crucial for scaling the school program, given see the challenges firsthand, understand dedicated individuals—visionary leaders to children learning. the ministry’s role in institutionalizing and the necessary action to address the issue, accrediting programs in schools. In 2011, and can be held accountable if actions the Ministry of Planning and International are not taken. Over the course of its Cooperation became another key partner 20-year history, Pratham has cultivated Visionary leaders in helping INJAZ expand into Jordanian government supporters at all levels with colleges and universities. mixed results. It has found that while time- A common feature in scaling studies coalition-building skills, and their ability consuming, it is often easier to cultivate across disciplines is the critical role that to articulate a clear vision and motivate In Mexico, the secretary of education in champions at the local level, as those visionary leaders played in driving the others.112 They also invested significantly 111 the state of Puebla, Luis Maldonado, and closest to the problem are generally more scaling process. These leaders came in leadership development within their Patricia Vázquez, who at the time was the likely to understand the nature of the from all sectors—government, civil society, respective organizations and systems. director of technology in the Department problem and be supportive of the actions and private sector—and served as the While some are born leaders, others of Education in Puebla, made an required. founders or leaders of their organizations are cultivated through their experiences important difference in helping Teach For or policy programs. They included people and practices. This, for example, is the All’s partner Enseña por México (EPM) get The cases reviewed often benefited such as Amorim and Soares da Silva, the idea behind Teach For All’s partners started. Prior to both leaders’ involvement, from continuity or stability in leadership. two secretaries of education in Amazonas placing individuals in classrooms, so the organization was struggling to take off Media Center in Amazonas benefited state who created Amazonas State that through the experience of working and with their support, CEO Erik Ramirez- from relatively stable and continuous Government’s Media Center; Madhav in communities and collaborating with Ruiz was able to scale EP’s work. Today, political leadership. The current secretary Chavan, co-founder of Pratham; and parents, schools, and students, they gain EPM works in nine states across the of education, Rossieli Soares da Silva, Joan Ganz Cooney, founder of Sesame a deep understanding of the problems country.113 assumed office in 2012, but he had been Workshop. As Hartmann and Linn outline, and the potential solutions and then go working with the previous secretary, common features of these leaders on to become education leaders inside Cultivating champions often involves Gedeão Amorim, for five years before that. are their persistence, networking, and and outside schools. engaging government officials in the Research by McKinsey and Company

88 89 If there is clear evidence of impact, people come to other leaders cannot be in denial. It has the experience with Schools of Tomorrow, expect it and it becomes much more difficult to reverse. also played an important role in offering a government initiative that focuses policymakers credible alternatives to on reducing school dropout rates and show that the issue can be addressed. improving learning in disadvantaged identified leadership continuity as essential three approaches have demonstrated Pratham has been effective at providing schools in Rio de Janeiro’s favelas. While in not only catalyzing a reform but in some degree of effectiveness. The first decision-makers with a tested and the effort was initiated under the former sustaining it. It found that the median tenure is to work across political party lines. proven menu of options for implementing secretary of education in Rio de Janeiro, of education leaders of improving systems is One of the core tenets of Teach For All its “teaching at the right level” approach. Claudia Costin, it has lasted into the seven years. This is in stark contrast to the is that any country partner must be a term of her successor. Arguably, one of norm. For example, the average tenure for non-state entity working in partnership If there is clear evidence of impact, the factors behind this longevity is the superintendents of urban school districts with the public and private sectors. As a people come to expect it and it becomes program’s strong evidence of impact.119 in the United States is three years; for result, the partner organizations are less much more difficult to reverse. This was education secretaries in England, it is two vulnerable to leadership changes. This years; in France, it is two years.114 has been the case with Teach First, which represents 20 percent of new teachers in When continuity is lacking, the challenge disadvantaged schools across England Champions within the classrooms becomes to sustain the reform beyond and Wales,116 after growing almost 25 a champion’s time in office. In the case of percent per year during its first 11 years.117 In addition to government support, new program, policy, or practice for it to be

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions SAT, despite its success with cultivating Part of this success is attributed to its teacher buy-in is critical for any large- scaled successfully. In fact, case studies local champions, Asociación Bayán finds deliberate approach to work across party scale intervention to succeed. Whether reveal that sometimes it was preferable it is an ongoing, time-consuming process lines. As a result, it was the only education teachers directly deliver the program that they were not responsible for it. But to sustain this support, particularly in initiative supported by all three political or support others to do so, experiences it was critical that teachers understood political systems that have high turnover of parties during the 2010 UK general demonstrate that involving teachers as and were behind the reform. Educate! administrations and/or personnel. Pratham election.118 SAT has also insulated itself key agents of change is vital in scaling experimented with teachers delivering its has also had the experience of programs from changes in power and potential interventions. In the case of Lesson Study co-curricular entrepreneurship, leadership, unraveling when champions have left political patronage by retaining control in Zambia, teachers were not merely and workforce readiness program in office, causing the associated government of tutor hiring, placement, dismissal, and taught a new teaching technique during secondary schools but encountered partnerships to disappear. The reforms that management. an in-service workshop; they were also some resistance as teachers viewed it tend to survive are those that are firmly empowered to identify what was needed, as an additional task that was not part entrenched. Because citizens often feel A second approach is to build innovation or collaboratively develop a lesson plan, and of their job description or assessment. strongly entitled to the reforms, it becomes reform leaders and champions throughout then practice delivering it and discussing As previously discussed, Educate! found politically impossible to retract them. This the system so that continued support is the experience with peers.120 This is also greater success (in terms of impact) by is another reason for strong community not dependent on one champion. This was STIR’s approach, where teachers are training mentors, who were graduates of participation in change processes. part of the motivation behind Fundación encouraged and supported to lead the program, and young entrepreneurs Escuela Nueva’s founding. In the regions positive changes in their classrooms, to deliver the entrepreneurial curriculum Beyond human relationships, where Fundación Escuela Nueva has through sharing and putting into action and modeling new teaching methods organizations can take steps to help managed to bring partners together their own ideas for improvement through that teachers could eventually integrate expand and sustain reforms. In its review to integrally implement the model, “micro-innovations.”121 They then are into the classroom. It also developed of 17 large-scale health interventions, educational outcomes have flourished. supported to build their own teacher teacher associations to provide teachers Millions Saved found that “mobilizing networks and become “change agents” with practical in-service training to political leadership and champions takes The third approach relies on data to in pushing for broader systematic change develop the core interactive teaching a little luck and a lot of preparation.”115 catalyze and sustain the political will within their schools and districts. skills and ultimately be better positioned needed for scaling. Often, evidence has to effectively adopt national education Drawing from our review of 14 education been used to demonstrate the severity This does not mean, however, that teachers reforms. Similarly, Pratham works to ensure cases and the literature around the issue, of the problem so that policymakers and must be responsible for the delivery of any that head teachers and teachers are

90 91 supportive of its remedial program before in which trained volunteers, rather than bringing it to schools, even in instances teachers, deliver it. 7. Technological advances: Context-appropriate technologies can accelerate Paying attention to those education progress. who stand to lose

As critical as it is to dedicate time to disappeared when the government To date, the vast majority of developing Technology brought cost cultivating champions, it can be equally implemented the intervention across the country governments’ engagement saving and efficiency- important to understand those who country.125 with technology in education at scale stand to lose as a result of an effective has been putting computers into generating processes. approach scaling. Literature finds that Getting consensus around a new and classrooms. However, this was not the programs may have failed at scale relatively unproven model is difficult, situation with the cases reviewed. Where because they did not pay sufficient particularly if it requires reallocation technology was used, it was to overcome by digitizing a variety of textbooks, attention to institutional incentives, of funding resources, adjustments in a context-specific barrier, such as poor storybooks, and reference materials that vested interests, and how those who human resources, and other politically infrastructure or a lack of materials could be accessed through e-readers

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions stand to lose out would react.122 Some difficult activities such as curtailing some or trained teachers. Media Center and cellphones. of the latter may not be apparent until services and replacing them with others. employed multi-point videoconference after the pilot stage if the intervention Those who benefit from additional capabilities to overcome long distances In some cases, as discussed earlier, is showing success at scaling.123 This resources are likely to be excited by the (and few roads) between communities and technology brought cost saving and can quickly undermine any gains made. prospect, while those who stand to lose secondary schools. Bridge International efficiency-generating processes, such Daron Acemoglu, an economist at MIT, through reallocation will predictably feel Academies addressed the shortage of as automating payment through dubbed this the “seesaw effect”—making otherwise. It is important to think about trained teachers by recruiting high school cellphones and automating data a change without redistributing power what it takes to generate partnerships graduates from local communities and collection and analysis systems. Bridge or the equilibrium of power usually leads and overcome opposition where people supporting them through not only training International Academies’ teacher guides to a counteracting force so strong that stand to lose.126 but comprehensive teacher guides have been important in freeing up the change is unlikely to have significant developed by expert subject matter time for teachers to focus on teaching impact.124 This was experienced with Often, there is reluctance to redraw lines, teachers and delivered via tablets. While and individual engagement, rather the scaling of a contract teacher close down programs, or replace existing seasoned teachers would for obvious than on lesson planning, attendance intervention in Kenya. While an NGO-led resources. Jaideep Prabhu, a Cambridge reasons feel constrained by scripted tracking, administrative paperwork, pilot in western Kenya found the hiring professor who has written extensively lessons, Bridge argues that the new and other activities that detract from of contract teachers to be effective in about innovation, calls these kinds of teachers in their schools are reassured active engagement with students in the raising students’ tests scores, its impact changes a “willingness to cannibalize” by the guidance and it is the best way classroom. Similarly, by automating and and considers them a key characteristic to ensure all their students learn. At its centralizing all of the back-office work, of innovative systems.127 Many cases we inception, Worldreader aimed to address managers can focus on monitoring and reviewed were strategic about launching the lack of appropriate books and other supporting teachers and students, rather As critical as it is to dedicate an intervention where it would likely not reading materials in developing countries than time-consuming administrative work. time to cultivating champions, threaten vested interests or upset power it can be equally important to balances. Media Center, for example, understand those who stand to did not start with its distance-learning program in the capital of Amazonas, but Where technology was used, it was to overcome a context- lose as a result of an effective rather, in the middle of the jungle where specific barrier, such as poor infrastructure or a lack of approach scaling. it was not threatening to existing actors. materials or trained teachers.

92 93 Educate! has identified mobile money videos and activities they should use with learning through more real-time data and implementers’ time. At the end of the as one of the biggest cost- and time- their class each week; preloaded feature collection and processing via tablets day, the cases underscored that it was not saving measures as it has scaled. The staff phones and pico projectors to bring content and mobile phones. All of this ultimately about a specific piece of technology but working across Uganda no longer have to classrooms in low-income communities; helped to increase transparency and rather about how technology enabled to travel back and forth to Kampala to and interactive voice response systems to accountability and to free up educators’ society to do something better. receive their biweekly paycheck or money enable users to access radio episodes on for their programs. That saves time, which their mobile phones. frees them up to focus on their day-to-day work. It also saves money that would have One potential risk of leveraging gone for travel, gas, and lodging. technology for scaling is that it often targets and benefits those who already As technological advances occur, many of have access to the Internet and therefore the cases reviewed underscored the central may perpetuate inequalities that exist 8. Windows of opportunity: importance of human involvement in the in access to technology. For example, process of integrating technology within in many countries, men tend to control Effective education approaches are more likely to take their operations. In the case of Pratham, the household phones. However, with root and spread when they align with country priorities. having people interact with, analyze, and deliberate, targeted action, technology interpret the data is important for learning can be leveraged to help overcome

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions from data. Pratham CEO Rukmini Banerji historical inequities. Worldreader found described the human interface as the glue that, on average, women spent six times between back-office technology and front- as much time reading on mobile phones Scaling requires a certain aptitude for in rural areas. In some cases, such as in office visualization. Unfortunately, far too as men. The organization is examining opportunism—not bound by rigid strategy, Colombia, Nicaragua, and Honduras, many interventions have fallen into the the reasons further, but hypotheses but flexible enough to take advantage the government was open to alternative trap of choosing a technology first and include that the mobility and privacy of of windows of opportunity. Programs models to help address the problem of then looking for an educational problem digital reading—and perhaps technology that were successfully integrated into teacher shortages in rural areas. SAT was to solve with it, rather than the other way in general—drive adolescent women and the national education system often ready to step in and provide its program. around. Teacher training, power supply, girls to read more. identified opportunities for a win-win Today, Central American governments are and systems for maintenance have not proposition. Case studies demonstrated motivated by the challenge of stemming received equal attention. Many of the cases While there is great potential in the ways how innovation more readily takes root the tide of rural to urban migration. At reviewed underscored how technology has technology can accelerate progress in and spreads when it responds directly various strategic points from the 1970s to be appropriate for the context and the learning at scale, the cases reviewed to challenges facing the state and aligns to present day, SAT has been poised users, and that “high tech” was not always identified more modest, albeit important, with existing government priorities and to leverage recognition by national the best solution, especially in the contexts ways that technology was contributing policies. As Chris Dede, professor at governments that space exists for both of low resources or low literacy. Sesame to scaling. This included so called front- Harvard’s Graduate School of Education, traditional and alternative programs. Workshop’s projects, for example, use office technology that interfaced with and colleagues write, “If an innovation is technology that run the gamut from high- leaners and teachers, such as interactive designed to fit a larger reform agenda, that tech (i.e., tablets and smartphones) to low- videos in Media Center, e-readers and innovation will almost certainly be more tech (i.e., print materials, radio, and radio mobile apps with Worldreader, and sustainable and therefore scalable.”128 over basic mobile phones). In addition television with Sesame Street. The Scaling requires a certain to creating content that directly targets cases also identified how technology SAT has benefited from, and in some aptitude for opportunism— children, Sesame Workshop also develops provided critical back-office support cases strategically leveraged, key not bound by rigid strategy, resources to support caregivers using that helped to improve operations, such external moments to advance its mission. but flexible enough to take technology. In India, for example, it has as Bridge International Academies and For example, in the 1990s, many Central used mobiles phones as a tool to remind Educate! generating cost savings by American governments felt political advantage of windows of teachers about the Galli Galli Sim Sim automating payments and more rapid pressure to provide secondary schooling opportunity.

94 95 Media Center, the state-led distance As Educate! expands into additional Although this request fell outside of the a public school operator in Liberia— learning secondary school program in countries beyond Uganda, it has made regular programming, staff were very starting with a pilot of 50 schools. Amazonas, launched at an opportune the strategic decision to target countries happy to use Room to Read’s resources What made this possible for Liberia time in Brazil, as the country was in the with active skills-based reforms that and distribution networks to support was that the cost basis of the delivery process of enacting a law guaranteeing align to its model. For example, Educate! the recovery efforts. The organization of service of the Bridge intervention at universal secondary education. By decided to enter Rwanda in 2015, when mobilized the movement of 500,000 scale is comparable to the ministry’s per 2016, states would be responsible for the country was undertaking curriculum books to affected communities. capita spending for primary education. ensuring that all children have access reforms to its secondary education, Consequently, Bridge can deliver its to three years of secondary school—an with a stated goal of a competency- Through the global financial crisis, model and not reduce teacher salaries enormous achievement for a middle- based curriculum. Educate! had clearly Aflatoun International found itself as from public sector levels. A budget income country like Brazil, where identified gaps in both the Uganda and one of the few organizations providing limited by a parent-funded model, having more than 2 million children are not Rwanda school systems, where the skills financial education in primary and demonstrated learning impact at scale, attending secondary school.129 It is also being taught did not match the labor secondary schools around the world is now serving government needs and is an enormous task for under-resourced market demand. It had also pushed through its network of NGO partners. It being integrated into the public sector. states such as Amazonas, where only six for reforms to teaching methods that has been able to leverage the interest of of the 62 municipalities are connected by moved from rote learning to practice- national actors, including central banks, This example demonstrates a road to the capital. For the thousands of based education and from more ministries of finance, and financial government’s willingness to experiment young people living in small communities theoretical to skills-based education. institutions, to support 28 education and allow different partners to

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions scattered along the Amazon jungle, it That positioned Educate! to provide the departments on integrating financial participate in the delivery of education could take days or even weeks to travel education ministries with an evaluated education into their education systems. while retaining ultimate responsibility to secondary school by boat. Ensuring and effective solution to an urgent for achieving good quality outcomes. that all children stay in school for all problem that the country needed to Governments can also serve as “pull” However, this is not to underestimate 12 years would have been difficult to address, rather than trying to create a forces, conducting due diligence to the challenges such a partnership needs achieve without innovative models such parallel structure. find partners that meet their needs or to overcome in order to succeed. It is yet as Media Center. that can carry out a reform agenda. to be seen to what extent Bridge will Crises have also provided moments more In Liberia, following reforms in the civil need to deviate from its original model Bridge International Academies had not amendable to adapting or reforming service and Ministry of Health that to adhere to government demands, and originally planned to open in Nigeria in existing systems. Room to Read was enabled the government to better if in turn this will affect the quality or its first phase of international expansion, one of the first organizations that the engage with citizens’ needs and innovation of the schools it is operating. but it responded to a DfID request for government of Nepal turned to for support drive service delivery, President Ellen In addition, the pilot may uncover some bids to improve learning outcomes in after the devastating earthquakes in Johnson-Sirleaf sought improvements evidence about the impact of Bridge’s the private market for education serving 2015. In this unprecedented situation in the delivery of education.130 This led model for the poorest children who more than 1 million children in Lagos. that destroyed more than 700,000 to review of education interventions in would be unable to access any school Bridge now works in Lagos as part of homes and 47,000 classrooms around Africa, formation of a public-private that required a fee to be paid by parents. a government-sanctioned program, the country, the thought was that Room partnership technical committee, and the result of a multiyear relationship to Read could distribute storybooks ultimately, the Partnership Schools What the case studies underscore is that between DfID and the Lagos State to affected communities in ways that for Liberia program, where education change depends to a great degree on Ministry of Education. would create some joy and learning. service providers are contracted to the social and political atmosphere. The operate public schools on behalf of the organizations that were most successful government, financed by government seized those moments where there were Case studies demonstrated how innovation more readily and free to children’s families. Bridge supportive, positive attitudes toward International Academies was asked change. These cases successfully framed takes root and spreads when it responds directly to to host an inspection by the Liberian their intervention as an effective means challenges facing the state and aligns with existing government, and after several months of achieving the education priorities of government priorities and policies. of engagement, agreed to become the country.

96 97 in the form of rigorous evaluations—was It is also part of the premise of a 9. Better data: translated into an emotional imperative movement to understand how well to act. Educate! found that some of its children are learning to read in their first Data on learning and scaling play a central role by motivating earliest champions knew students who years of primary school. RTI International, informed action at the policy and practice levels. had gone through the program and were with support from USAID, developed the impressed with its impact. Policymakers Early Grade Reading Assessment (EGRA) identified that this secondhand evidence, in 2006, motivated by the concern that which was then backed up by more limited reading skills in the first years of rigorous quantitative data, contributed to school were holding many children back The case studies and literature underscore shaping policy and practice. Across the the incorporation of Educate!’s curriculum from successfully proceeding through the central importance of evidence in case studies, data play at least three into Uganda’s national education system. school. Because RTI made the EGRA tool scaling efforts. Various forms of data— important roles in the scaling process: open source and freely available, its use from experimental, scientifically designed motivating action to address the problem; Asociación Bayán’s leadership, the spread rapidly across the developing evaluations, to qualitative data, to student shaping the design and implementation of organization behind SAT in Honduras, world, and today, the tool has been testimonials—all play an important role in the response; and sustaining the response. was savvy and brought along local adapted and used by governments, schools, politicians and donors to visit schools civil society organizations, and the private (known as centers) since the program’s sector in over 40 countries. Data from the inception. It invited the DfID to participate wide range of contexts in which EGRA has

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions Motivate action in an early evaluation so that DfID been used, which showed the surprisingly could see the impact firsthand. Similarly, low levels of reading proficiency within In each of the case studies, data have national policy in Colombia and scaling to Asociación Bayán spent considerable many countries, provided a wake-up call to been instrumental in cultivating key other countries.132 time inviting local politicians to personally many education actors, from teachers to champions by arming decision-makers see the changes in students. Once local ministers to global policy leaders. with the information needed to make Case studies also revealed that various policymakers were onboard, Asociación difficult decisions about where to invest forms of evidence have been effective Bayán worked tirelessly to turn political The case studies and literature, particularly finite resources. As Adele Cassola, PhD in influencing policymakers to scale support into binding legal frameworks in behavioral economics and psychology, student at Columbia University, and Jody certain interventions. Impact evaluations and agreements. show the importance of quick wins to Heymann, Dean of the Karin Fielding certainly can play a role in building a demonstrate that change is possible. This School of Public Health, describe, evidence- knowledge base of “what is working here,” While even the most rigorous and is also the theory behind Rapid Results, based decision making is about the best but experience shows that the spread accessible data will not necessarily be an approach that focuses on mobilizing use of scientifically based knowledge and integration of interventions may acted on, they certainly can play an communities to overcome specific in a realistic time frame with limited require firsthand knowledge of change important role in changing behavior. challenges through leveraging their own resources.131 One of the best weapons in a community, which can be particularly Experience shows that the first step in assets. The idea is based on the premise against the status quo is actionable data. important for policymakers. This is motivating action is often to show the that what is lacking in development is not Worldreader’s use of e-book sales data supported by Rogers’ seminal work, in severity and urgency of the problem, specific information, money, or technology has been key in demonstrating the viability which he writes, “Most individuals evaluate demonstrate that change is possible, and but the motivation and confidence to use of digital publishing to African publishers, an innovation, not on the basis of scientific ensure that capacity exists to take action. available resources. Rapid Results Institute by proving the existence of a robust local research by experts, but through the This is the premise behind Lesson Study widely implements this approach with market and a nascent international one. subjective evaluations of near-peers who in approaching teachers as key change great success across developing countries, Vicky Colbert, co-author of the Escuela have adopted the innovation.”133 A pattern agents. Teachers work collaboratively including with corporations in Ethiopia Nueva pedagogical model and founder across the case studies was that once key to identify challenges in their classroom, to increase HIV testing of employees, in of the nonprofit Fundación Escuela decision-makers saw an intervention’s develop lesson plans to address the Nicaragua to improve dairy farm’s milk Nueva, claims that a strong research results firsthand, the program or policy challenges, and practice delivering the quality, and in Rwanda to double the and evidence base was a major reason was expanded. There was typically a lesson with other teachers observing and number of attended births—each one in Escuela Nueva was successful in affecting pivotal moment when information—often offering constructive feedback. less than 100 days.134

98 99 its teacher tablets in schools. Bridge is Educate! invests considerably in building Shape design and delivery then able to very quickly, efficiently, and tools that go beyond tracking participation effectively roll out any changes across to measuring impact. It dedicates 12 to 13 A common feature of interventions that decisions. In this way, it has managed to its more than 450 schools, as it controls percent of its budget toward monitoring have successfully scaled is that they rarely multiply impact in the face of competing the entire supply chain—from school and evaluation. This includes tracking 20 135 follow a linear path of research to action. needs and scarce resources. Over the construction to curriculum design to performance indicators on a weekly and Rather, they undergo a more circular past 20 years, Pratham has combined teacher training. Room to Read, meanwhile, trimesterly basis through SMS messaging process of experimentation, learning, and lessons from rigorous evidence with uses data to evolve and improve its and smartphones. Educate! built its own course corrections. This requires ongoing field-level experience to formulate programs, as it considers transparency tool to measure leadership, creativity, self- testing, evaluation, and revision of models— and inform strategies and programs. central to its success. When findings efficacy, and savings behavior, since it could beyond proving the initial efficacy of a model To track the progress of more than a showed that the national curriculum in not find any existing tools to measure these as it continues to expand. Historically, there million children a year, Pratham uses Vietnam was already achieving strong “soft” skills appropriate for the sub-Saharan has been much greater attention to and simple tools and methods to measure outcomes in foundational reading skills African context. This instrument is called support of the proof of the concept phase, and monitor learning gains made by among students, Room to Read decided the Secondary Skills Assessment Tool and is with too few resources and attention to each group in learning camps through to focus its Literacy Program resources open source for other organizations to use. It scaling plans after the initial intervention is the year. These data are uploaded on a on growing its network of learning is continuously refined based on adjustments 136 proved. As a result, it should not come as portal from the field and are available environments and publishing quality made to the program. Data are collected a surprise that many pilots do not survive for all team members to view at any reading materials. It was therefore able in real time, allowing Educate! to monitor 137 Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions beyond their beginning phase. time. According to Banerji, “The biggest to save financial and human resources in performance and make any necessary 139 need for data in our system is for us.” this area and divert them into other more changes across all schools. This rapid From Sesame Workshop’s start, research This learning has been incorporated into pressing work in Vietnam and elsewhere program monitoring helps to maintain quality played a prominent role in its content a suite of evidence-based options for in the Room to Read network. control, as Educate! continues to expand. creation model. Programming is a implementing Read India that is offered continuous process that begins with the to state governments to adopt, as has assessment of need and includes ongoing been taken up to various degrees with the formative research and, where possible, states of Bihar, Maharashtra, and Uttar Sustain scale summative evaluations of the program’s Pradesh. impact. All of this, in turn, feeds back into The role of data extends beyond As the former secretary of education in thinking about the needs for the next season Bridge International Academies is persuading decision-makers to invest Rio de Janeiro, Costin explained, while of programming. Every season of Sesame somewhat of a unique case in its ability scarce resources into scaling an discussing Schools of Tomorrow, “In Latin Street is an experiment. The integration of to continuously strengthen its programs. It initiative—data can also have an impact America, there are lots of laws that are research into the production process and collects and mines an enormous amount on an initiative’s sustainability. Once data written but that do not happen, things the spirit of experimenting have resulted in of data in real time through the use of demonstrate improvements, it is difficult to that cannot be enforced. Through my programs that are examined critically, where reverse progress. In their seminal scaling experience as a public policy specialist, I teams test new ideas and make important work, Hartmann and Linn argue that well- see the best way to ensure continuity of course corrections as necessary. As founder A common feature of designed evaluations can build political any program is to really bring results, as Joan Ganz Cooney has said, “Without interventions that have support even if political parties change.140 then it becomes something desirable.”141 research, there would be no Sesame.”138 successfully scaled is that they A key driver behind Pratham’s success rarely follow a linear path of The role of data extends beyond persuading decision- resides in its emphasis on experimentation research to action. Rather, they makers to invest scarce resources into scaling an and learning. This includes an openness undergo a more circular process and honesty about where things are not initiative—data can also have an impact on an initiative’s working. Pratham has used this evidence- of experimentation, learning, sustainability. Once data demonstrate improvements, it is based approach for making critical and course corrections. difficult to reverse progress.

100 101 VII FINANCE 10. Flexible education financing: Introduction Financing should be flexible, including to build core operational capacity.

Case studies reveal that flexible of Zambia invited the JICA to provide funding was instrumental in allowing for technical support in adapting and experimentation, learning, and growth: rolling out Lesson Study across the governments and donors that focused on country, it has provided approximately agreed-upon results to achieve but were 90 percent of the financing over the flexible and had trust in the implementer past 10 years for the in-service, peer- about the best way to achieve them. to-peer teacher training practice. These This included investments in activities cases offer powerful examples of where that were most needed to support the governments provided predictable and

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions scaling process at various points, which flexible financing that facilitated the Long-term education frequently required developing core expansion of quality learning initiatives. operational capacity. FINANCE financing External donors can also play an If the overall goal is to have an effective important role in the scaling process, learning intervention go to scale, particularly in investing in earlier stages 10 11 12 financing from government is key. of experimentation before an idea Domestic governments remain the largest or approach is proven. Foundations, source of basic education financing, and individuals, and corporate donors Flexible education “Middle phase” the percentage of national budgets generally have a greater appetite financing financing dedicated to education has continued for risk taking than governments that to steadily increase since 2002 in many face pressure to demonstrate more low- and low-middle-income countries. immediate results to their constituents. The cases examined highlight a range Despite this comparative advantage of ways that governments are providing and best intentions, donors at times Designing and delivering at scale of the source, financing for scale needs to financing that is key to scaling. In the can undermine the scaling process with require resources—and the more you take a long-term approach, invest in core case of SAT, the Honduran government short-term or more restrictive funding. grow, the more resources you need. To organizational capacity, and activate is paying for the largest recurrent cost The average development assistance complicate matters, when it comes to middle-phase funding. of the alternative secondary education “project” lasts less than two years from scaling quality learning, how financing program run by the NGO Asociación start to completion142 and does not allow is structured and allocated seems to be Bayán—tutors’, or teachers’, salaries. at least as important as total financing When it comes to scaling quality Pratham was able to reach 5 million available. The cases demonstrate, and learning, how financing is children across India last year with its Flexible funding was teaching at the right level approach by the literature supports, the notion that structured and allocated seems instrumental in allowing for stability and flexibility of financing are working through government-funded necessary for scale to occur in ways that to be at least as important as schools where teachers were trained experimentation, learning, contribute to lasting change. Regardless total financing available. in the approach. While the government and growth.

104 105 for much time for the type of investments external, nongovernmental sources, we that were more “tangible, measurable, describe, too often donors are inhibiting required for building systems and capacity. focus in the following sections on the role attributable and controlled.”145 This is not scale by supporting only one-off initiatives. Ultimate results being targeted, such as that donors can play in better facilitating to argue against investments in direct “Instead of supporting an organization’s improved learning, can take years. the scaling process and how external support, but to argue that investments in overall mission, funders often prefer to funders and governments can work institution building are needed as well. provide grants to programs that target Therefore, given the need for much together more effectively to help quality a particular issue over a limited period of improved financing approaches from learning interventions spread. Network secretariats often play an time. Corporate foundations, in particular, important role in scaling, including by often allocate capital to efforts that align helping to share lessons, maintain quality, with their own institutional goals but and ensure fidelity to a model as it expands not necessarily with the broad goals of Financing for core operational support and is adapted to new contexts. Networks the nonprofits they fund.” This prevents are also critical for the coalition- and field- organizations from growing and further According to Chris West, former director manage them. For example, as it continued building often required for large-scale developing their capacity.146 of Shell Foundation, creating scalable and to expand, Worldreader required an reforms. sustainable enterprises requires building investment in product and software In an encouraging sign, Grantmakers for “core capacity, operating systems, and development that was particularly difficult However, in many of the case studies Effective Organizations President and robust infrastructure.”143 In focusing on to get funded. This is not to argue for reviewed, the secretariat was often the CEO Kathleen Enright argues in favor why some businesses scale successfully limitless or even proportional growth of most difficult part to get funded. Teach of investing in capacity building that can 147 Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions and many others do not, Verne Harnish, an organization or institution as it scales, For All is an example where the global help organizations boost their impact. author and founder of Entrepreneurs’ but to support the core systems needed to organization provides critical support Over the past eight years, the group Organization, identifies three challenges scale impact. to each of the 39 independent partner has built tools and resources to support in scale: inability to grow enough leaders, organizations to adapt and implement grant makers to do so. In 2013, GuideStar, failure to address increased competitive The rise of rating agencies and attempts to the approach in their respective countries. Charity Navigator, and the Better Business pressure, and lack of systems and structures boil down an organization’s effectiveness This support includes providing a global Bureau Wise Giving Alliance joined forces to handle the complexities of growth.144 based on a single metric—a percentage platform for sharing of ideas and to encourage donors to stop the practice going to “overhead” expenses—can supporting partners in adapting the ideas of bluntly evaluating nonprofits based Building core operational capacity and intentionally or unintentionally reward to their local contexts. Initially, despite on what they refer to as an “overhead ensuring a smooth transition between those who do not invest in building systems the central role that it plays, Teach For myth”—“the false conception that financial innovation and scale require flexible and infrastructure. Donor restrictions often All found it challenging to demonstrate ratios are the sole indicator of nonprofit financial support. Unfortunately, this type limit the amount that can go to overhead, the value of its support and to generate performance,” and instead focus on of flexible financing for core system support and politicians might prefer investments funding for a global approach to results. They have published two open is more often the exception than the norm. in buildings, books, computers, and other education. Fortunately, as the Teach letters to U.S.-based donors, challenging Arguably, some of the cases reviewed were tangible inputs and outputs that can be For All network has grown over the past them to change their practices.148 held back from expanding or spreading seen by their constituents, rather than more eight years, it has become easier to make more quickly because of limitations with intangible systems, such as monitoring and the case that the organization’s role in Recently, the Ford Foundation announced core support. Because scaling at its evaluation systems, that are required for providing support to partners is critical. an important transformation of its funding essence is about people and process, scaling. A McKinsey and Company study strategy to provide grantees with greater it is ironic that often these are the most on philanthropy in India found that while As Alice Gugelev and Andrew Stern access to general funds. More than 40 difficult components to fund. Managers 90 percent of donors and experts believe at the Global Development Incubator percent of all Ford Foundation funds will of some cases felt pressure in fundraising that investing in institutional strengthening to keep their overhead low, especially as and policy change is important for Network secretariats often play an important role in the organization grew and needed the overall transformation, 60 to 65 percent support more than ever to develop systems of interventions supported were in the scaling, including by helping to share lessons, maintain to monitor larger programs efficiently and category of providing direct support to quality, and ensure fidelity to a model as it expands and is build human resources and capacity to beneficiaries, which included outcomes adapted to new contexts.

106 107 be allocated to general operating support, According to Pratham’s leaders, the most development cooperation agencies are Successful efforts at scaling twice what was previously available. helpful donors were those who were open- “overly ambitious and not adapted to the learning interventions often Harkening back to the foundation’s minded in terms of what Pratham did on pace of reform in the partner country.”154 The contributions to the civil rights movement, the ground but were systematic in holding case studies revealed that reform processes benefited from long-term Ford Foundation President Darren Walker its leaders accountable to what they said are slow. This contrasts with the average financial and technical believes that this commitment will enable they were going to achieve. These donors donor engagement of 613 days from start to support, as well as supportive 155 the foundation to have a greater impact had spent time with Pratham in the field completion. policymakers and donors who on such issues as inequality, which will and were along for the journey, keen to require social movements to tackle see results influence governments. Donor Successful efforts at scaling learning understood the realities of entrenched social problems. “We’re going agencies and foundations, such as the interventions often benefited from long-term what it would take to achieve to move away from bending our grantees Hewlett Foundation, played a catalytic role financial and technical support, as well as the effort’s desired impact. to fit into our boxes and do a better job by encouraging Pratham just enough and at supportive policymakers and donors who of listening and learning.”149 Ruth Levine, just the right time to help Read India follow understood the realities of what it would director of the global development and its charted path, but they did not suffocate take to achieve the effort’s desired impact. and financial education programs to population program at the William and Pratham by questioning its every move. According to INJAZ, the USAID long-term demonstrate the viability of its material Flora Hewlett Foundation, advocates financial commitment is what truly allowed within the education system and to kick- for grantees and funders to be more Particularly with interventions originated the skills building program to become start the development of relevant content realistic about what it takes to start a by non-state actors, donors can play a established and grow. This commitment also for the national context. It estimates that

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions new program or organization, which may critical role in providing flexible support came with USAID’s know-how and auditing the average time between the start of require building a “zero year” into budgets that helps to build core operational systems, from which INJAZ learned and such a program and work on curriculum and timelines—that crucial initial year of capacity needed to scale effective ideas began building its own systems.156 integration to commence to be almost five development and planning.150 and approaches. years. The practice of Lesson Study, a peer-to- peer in-service teacher learning method, In the Quality Education in Developing did not scale across Zambia overnight. It Countries initiative, the Bill and Melinda took a step-by-step approach over 10 years Gates and Hewlett foundations pooled to cover all 10 provinces. The expansion resources and were willing to invest a of Lesson Study closely followed and substantial amount of long-term funding 11. Long-term aligned with the government’s policies for for something they believed would have continuing professional development for a significant impact on learning. Pratham teachers. The gradual scaling expansion responded with a proposal that outlined education financing: allowed the program to continuously refine the first generation of Read India. It was Stable and predictable support is essential. its implementation strategies. Similarly, by appropriate for a number of reasons, gradually expanding Lesson Study across including its ambition to go “for all of the country, the Zambian government could India.”158 This support lasted more than observe its impact, such as improved math eight years. and science pass rates, before incorporating In addition to flexible support, scaling at times support scaling, they often the method within its national development A long-term commitment is behind Room requires stable, predictable financing, lack a systematic scaling focus and plans.157 JICA has remained a constant to Read’s requirement that it must first 151 152 often for a decade or more. This approach. They also systematically partner throughout these 10 years. This long- raise funding for a minimum of three often does not align with politicians’ or underestimate how much trial and error term commitment and phased approach by years before starting operations in a new 153 funders’ timelines. A new administration will be involved in the scaling process. all partners has been an important aspect country with full program implementation. or government official may discontinue The German development agency to Lesson Study’s success in Zambia. The principle is that an effective program funding for a predecessor’s program Deutsche Gesellschaft für Internationale and commitments to communities and or policy. Meanwhile, a key conclusion Zusammenarbeit (GIZ) candidly reflected Aflatoun International supports partners countries requires building networks in the literature is that while donors that the timelines often envisioned by in civil society organizations to start social and establishing roots without fear of

108 109 discontinuing the program mid-stride. This generally delivered between one to innovations end up.163 Experience and Fundación Escuela Nueva managed to finding is reinforced by research in other three months that lead to measurable literature reviewed found that new ideas cross this valley of death, expanding an sectors, such as health, which found that improvements in young children’s reading and pilots generally had little difficulty initial pilot of 150 rural primary schools in more predictable funding at adequate and mathematics. The Sesame Street finding funding—“donors love to fund Colombia in the mid-1970s to eventually levels enabled the system required for approach focuses on more immediate promising innovative new ideas; ministers reaching more than 20,000 public scaling health innovations to work.159 learning improvements, in addition to its will push their pet project; and venture schools by 1988. No single factor was long-term, positive impacts on children.161 capitalists provide startup capital.”164 responsible for this journey, but rather a While ensuring that deep and lasting For example, in Indonesia, researchers At the other end of the spectrum, large, number of mutually reinforcing drivers, impact can take time, the cases found that children who watched “Jalan established programs sustain their including founder Colbert’s appointment also underscored the importance of Sesama” (the Indonesian co-production) funding from national budgets. What is as vice minister of education in 1982, demonstrating intermediate results. It regularly showed greater improvements often missing in terms of development evidence of the model’s impact, is unrealistic to expect public or private in tests of early cognitive skills, letter assistance, domestic budgets, and capital community mobilization and support, and funds to support an initiative for multiple recognition, number recognition, counting, markets is support for the middle stage of long-term financing from international years without seeing any results. Specific health and safety knowledge, social scaling.165 partners, such as USAID, the Inter- guideposts are needed to signal progress development, environmental awareness, American Development Bank, the United toward the achievement of desired final and cultural awareness than those who Bridging this middle phase requires Nations Children’s Emergency Fund outcomes. Larry Cooley, founder of MSI had no exposure to the show.162 greater clarity and transparency (UNICEF), and the World Bank. Escuela Worldwide, and Linn found that securing regarding who funds which stage of the Nueva also demonstrates that even after

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions and maintaining the commitment and scaling cycle. Domestic funding is clearly crossing the middle phase, sustainability resources needed to scale over this needed to scale and sustain any efforts is not guaranteed. While the model period requires “tangible milestones, While ensuring that deep and to improve learning at a national scale. reached a scale of 24,000 schools and strategic communications, and an explicit lasting impact can take time, Earlier on, external support in the form of became a national policy, the program strategy for maintaining momentum.”160 financial, in-kind, or technical assistance is has been shown to be susceptible to Pratham’s Read India program’s “learning the cases also underscored the often needed during experimentation and political and administrative changes. camps” focuses on low-cost, intensive importance of demonstrating adaptation of a model or approach. Over Over the years, as Colombia has bursts of teaching and learning activity, intermediate results. the past few decades, there has been a become decentralized, Escuela Nueva’s movement by the global community for implementation and sustainability have improving donor effectiveness through been highly dependent on the political greater harmonization or coordination of will and the technical and financial assistance.166 However, to date, frameworks capacity of the local and regional and policies calling for a clearer division educational authorities. In turn, the level of labor among donors have focused on of implementation of the program across financing across countries and sectors, the country is not consistent. 12. “Middle phase” financing: and not on the phases or processes being financed. A greater focus on scaling could Silicon Valley could be a useful model to Financing is required to bridge the critical stage between provide an opportunity for more rigorous consider in terms of how it embraces and pilot and broad uptake. research and data on funding along allocates resources according to risk. various stages of scaling. New ideas receive minimal funding, and

Crossing the “valley of death” requires greater coordination Without flexible and long-term funding, implementation at scale, is particularly many interventions die in the “middle phase” important but often neglected. Chandy and collaboration among government agencies and donors, of scaling. This phase, which occurs after and colleagues describe this “valley of including greater segmentation and transparency around prototype or proof of concept but before death” as where most creative ideas and who funds at which stage of the scaling cycle.

110 111 as they prove themselves, they are able Omidyar Network, Sida, and Australia’s to attract more financial support. The Department of Foreign Affairs and Trade. more proved and less risky the initiative, The GIF funding model utilizes venture the greater the sums available.167 Some capital approaches and experiences development agencies—such as USAID’s for development innovation through the Development Innovations Venture and the use of “a tiered financing model [that] Innovation Against Poverty by the Swedish offers three stages of funding.” Each International Development Cooperation stage is characterized by the level of (Sida)—have been experimenting with advancement of the innovation and “by this approach. Both of these initiatives the level of evidence that supports its have fed into the more recently launched potential for success.”168 Global Innovations Fund (GIF), which aims to connect “innovators with scaling Crossing the “valley of death” requires partners” to successfully scale social greater coordination and collaboration innovations and reach millions in the among government agencies and developing world. GIF partners have donors, including greater segmentation committed more than $200 million over and transparency around who funds at five years, and include DfID, USAID, which stage of the scaling cycle. Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions

112 113 Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions

ENABLING VIII ENVIRONMENT

114 115 13. Supportive policy Introduction environment: Government policy must safeguard every child’s right to a quality education while remaining open to a diversity of ideas and actors to contribute to this common aim.

Evidence from cases and literature shows reviewed, it was not that the government ENABLING Supportive policy that government policy space that was stepped aside to allow a free-for-all, but ENVIRONMENT environment open to innovation was key for a number of that it continued to play a prominent role, good ideas to flourish and scale. In many of particularly in regulating and monitoring the more successful case studies reviewed, any new initiatives. Driven by a commitment

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions 13 14 the government provided a platform for a to ensure that all children receive a quality multiplicity of actors to be involved in the education, these governments reached out provision or financing of quality learning to a range of partners and considered new A culture opportunities, or both. In the cases ways of improving learning. of R&D Evidence from cases and literature shows that government policy space that was open to innovation was key for a When it comes to scaling, resources about creating an enabling environment number of good ideas to flourish and scale. In many of the matter. Evidence matters. But scaling does for innovation to flourish than specific more successful case studies reviewed, the government not occur in a vacuum. The environment action required for an individual program provided a platform for a multiplicity of actors to be involved in which a program or policy operates to grow. We focus below on the role of in the provision or financing of quality learning opportunities. can be just as important in contributing policy, while acknowledging that the to or impeding the process of going to enabling environment includes human scale. Political, institutional, economic, capital, culture, and other critical aspects cultural, and other factors play a critical that affect scaling prospects. role in “success” or “failures” in going to scale. Defending the mission of quality

Ultimately, quality learning is the product Successful scaling efforts are education, not a model or institution of a complex, adaptive system, not a more often about creating single program or policy. As the Annie an enabling environment for As economist Lant Pritchett writes in The innovation to take root and spread did E. Casey Foundation’s CEO, Patrick Rebirth of Education, “the price of better not adhere rigidly to a single educational McCarthy, notes, “A bad system will innovation to flourish than education is allowing freedom, giving model but were open to a diversity of trump a good program every time.”169 specific action required for an choices and hence ceding power.”170 possibilities. As educator and author Successful scaling efforts are more often individual program to grow. Governments that provided space for Frederick Hess describes, these are

116 117 governments that defend the mission had a chance to complete a quality high Governments that provided space for innovation to take of quality education, as opposed to the school education—something that was root and spread did not adhere rigidly to a single educational institution.171 There are examples of this from unthinkable only a few years prior. all around the world. model but were open to a diversity of possibilities. Governments that provide a supportive Amazonas State Government’s Media policy environment for scaling quality adjustment that allowed for non-education nonformal schools, of which Bridge is one, Center is an example of a program learning know that they do not have to professionals to fast-track into open to freeze expansion until new regulations that benefited from a supportive assume all roles of education delivery and positions. These deficits were prevalent were released.174 policy environment, where the national financing. In the case of SAT in Honduras, in the highest-need, most remote areas government set clear, quality targets, while the government recognized that it needed of the country—precisely the communities While the ministry was working to create leaving states and schools free to choose to expand the provision of secondary school EP sought to serve. These changes have a legal environment that would incentivize how best to achieve them. As discussed to the rural population while faced with allowed EP to work with the public school registration under the soon-to-be-released above, SEDUC seized this opportunity and limited resources and capacity. Therefore, system to scale and serve disadvantaged regulations and ensure that schools would creatively designed a distance-learning as described above, it made the bold children in Peru.173 be measured by examination performance, model based on the realities of the region. decision to allow a local NGO, Asociación families were concerned that if their With the space to experiment and iterate, Bayán, to recruit, train, and manage a cadre Bridge International Academies is an children were enrolled in alternative or along with clear national standards and of teachers (or “tutors,” as they are referred example where the low-cost private school complementary schools they would not be regulations, Amazonas was able to ensure to in the program), while the government chain was able to initially grow rapidly able to sit for national exams and transition to

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions that hundreds of thousands of teenagers paid for their salaries as contract teachers. in Kenya—opening a new school every secondary school. The Ministry of Education 2.5 days—in the midst of new regulations worked closely with Bridge and other being drafted for non-state schools. This alternative schools to find a positive solution. process of drafting regulations took place The Cabinet secretary demonstrated his Regulating education quality and over seven years, creating significant leadership and commitment to the needs of ambiguity over government intent for the children by ensuring that while waiting for the standards with a multiplicity of actors sector. In the lead-up to new regulations new regulations to come out, children who for the “alternative” or “complementary” had been attending these schools could sit As governments move to recognize multiple government initially did not allow fellows education sector being released, the for the exam at public schools, if necessary. education models, regulating education with Enseña Perú (EP), Teach For All’s ministry issued new guidelines revoking As result of the government’s leadership, quality and standards becomes more partner in the country, to be employed as full- such schools’ registration as testing Bridge’s first class of 2,900 graduates complex. Schools may vary significantly in time public school teachers on the Ministry centers, in a move to ensure that only was able to sit for the national exams. The terms of student demographics, teaching of Education payroll, which significantly ministry-registered schools could enroll government released the new regulations techniques, use of technology, and other limited EP’s ability to place fellows in the children to sit for the mandatory national in January 2016, allowing Bridge to begin important factors. While the diversity may highest-needs schools and communities. primary exit examinations. In 2015, the the process of registering its 405 schools in lead to more rapid innovation and discovery This started to change two years ago with Ministry of Education also asked that Kenya as “alternative” schools. of improved methods, policymakers may the arrival of a new minister of education, struggle to ensure that their regulatory who prioritized leveraging human capital tools—such as standardized tests or in the education sector. In addition, EP’s curriculums—evolve as rapidly as the sector. alumni working across regions in the public Role of civil society in monitoring This gap has impeded progress in a number sector built the right national, regional, and of sectors worldwide.172 local relationships and helped construct a educational developments more complete understanding of the needs This tension for governments between and opportunities within the education However, monitoring quality and standards advocacy groups play a pivotal role in providing space for innovation while community. Furthermore, a teacher deficit is hardly the responsibility of policymakers validating the quality of education, identifying maintaining quality control played out in resulting from the increase of teaching alone. In many countries, civil society promising practices, and influencing policy many of the cases reviewed. In Peru, the hours in the public school day led to a policy watchdogs, social intermediaries, and reforms. Their participation can lead to a

118 119 richer public dialogue on education and Ultimately, it is about the accelerate improvements. government putting in place Box 17. Government action to create space for scaling enough controls to protect its One such example is the Annual Status quality learning of Education Report (ASER) survey citizens, particularly those most conducted in India each year by the disadvantaged, without stifling Research by Dembélé, Samoff, and Sebatane points to specific roles that national ASER Centre, an autonomous unit of innovation or growth. governments can play in order to create space for scaling of learning. These are the Pratham Network, which is now summarized here and illustrated through the Media Center case study:182 replicated in Pakistan, Kenya, Tanzania, 175 Uganda, Mali, Senegal, and Mexico. as advisors in assessment activities, such • SET APPROPRIATE TARGETS: Establish broad objectives and national ASER conducts an annual, nationwide as in the survey design process, helps to standards that provide clear targets for the country. In the case of Media Center, 179 household survey of children’s ability to increase institutional buy-in. Results from the federal government of Brazil established a coherent vision along with clear read simple text and do basic arithmetic by Uwezo, which has administered large-scale, national goals for learning for all children, while leaving it up to states and deploying about 30,000 volunteer data citizen-led household-based assessments municipalities to determine the specifics of what students should know and do collectors from partner organizations, in Kenya, Tanzania, and Uganda, have at each grade level. including colleges, universities, NGOs, been cited in some government reports and youth groups.176 The release of the as contributing to government’s renewed • PROVIDE A COHERENT SYSTEM OF ACCOUNTABILITY: This includes 180 results is timed with discussions about focus on learning outcomes. According tracking how children are faring in regard to meeting national targets and

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions the national budget and shared through to an official with the Kenya Ministry of ensuring transparency and availability of results. In the case of Media Center, various media outlets with simple, easy-to- Education, developing a strong partnership Brazil established a national assessment system, the Brazilian Education Quality understand charts and comparisons.177 between the agency and other civil society Index (IDEB), using an internationally benchmarked index to measure progress actors such as Uwezo and the Kenyan Head of all schools against a baseline. A multi-country study by Results for Teachers Association, among others, has Development in 2015 found that citizen- been particularly important in addressing • ENSURE STABILITY AND SET CLEAR RULES OF THE GAME: Put into place led surveys, such as ASER, have been learning challenges in the country. The rich and provide complete information on clear regulations and policies that allow for extremely effective in making the low information on learning provided by Uwezo a stable environment for engagement in the education system. In Brazil, there is quality of education much more visible has been especially useful in helping the a clear division between federal, state, and municipal governments in regards to at the global and national level.178 They Kenyan government understand the scope the enrollment of children. The new state-level mandate for high school education also have successfully generated much and scale of the problem.181 encouraged the Amazonas state government to be creative and experiment with more dialogue among key stakeholders more efficient ways to reach a greater number of rural youth across the region. about the state of education and, in some Drawing from cases reviewed, it appears cases, have influenced policy or budget that innovations have the greatest likelihood • PROVIDE RESOURCES: These can include financial, human, and technical allocations. ASER has had an impact on of successful scaling where policymakers resources. In the case of Media Center, the federal government of Brazil has policy in some Indian states and, in general, provide space for experimentation and steadily increased its investment in education—from 4 percent of GDP in 2000 the surveys have greatly empowered local allow for alternative forms of delivery if to 6.3 percent in 2012183—while more equitably distributing funds to resource- civil society organizations to target and necessary, while providing the necessary poor states.184 At the same time, the Amazonas State government did not wait strengthen their advocacy efforts. safeguards and regulations to ensure for sufficient funding from the federal government and allocated state funds quality for all. Often these are governments for initiatives, such as Media Center. Rather than criticizing or ignoring the that value flexibility, focus on outcomes findings, government engagement has (rather than on process), and promote While the specifics of these roles vary, what is clear is that without participation been particularly important in contributing information sharing. Ultimately, it is about and leadership from the public sector, it is virtually impossible to ensure the to the effectiveness of citizen-led monitoring the government putting in place enough necessary enabling environment for interventions to scale systematically, efforts. The Results for Development controls to protect its citizens, particularly sustainably, and equitably. evaluation found that including key those most disadvantaged, without stifling government actors at the national level innovation or growth.

120 121 outcomes and ensure they continue that provide a picture of what students 14. A culture of R&D: to as they scale. Data on learning are know and can do.186 especially needed if countries and the Ensuring that more children learn requires a strong ethos of global education community wants to The cases reviewed provide good experimentation, collecting learning data, and using it for shift from scaling access to scaling access examples of how different types of actors plus learning. Actors across the ecosystem collected and used data. For example, continuous improvement. need to be able to review and learn from in Brazil, significant improvements what they and others are doing across a in the country’s overall learning number of dimensions and especially how outcomes have been attributed partly their actions affect children’s learning. to systemic monitoring of progress and greater accountability.187 The federal The cases exhibited an adept use studies, the key drivers behind creating Too few developing countries have the government’s IDEB, which sets targets for of evidence plus a strong culture of this culture were to collect and utilize robust and relevant learning assessment schools based on each school’s trajectory, experimentation, collecting learning data data on learning and other key outcomes systems needed to support evidence- evaluates a school’s performance and using it for continuous improvement, to ensure impact. They also were not based changes in policy and practice. As against its past performance and not while iterating without fear of failure. afraid to take risks and learn from failure discussed earlier, developing countries against arbitrary targets for all schools. Education lacks the kind of research and whenever possible. On top of that, many have a range of snapshot assessments, The public nature of the index provides development (R&D) culture that many of the cases used a strong foundation in but what is needed are robust and a real incentive for states to use other sectors employ, investing large sums research when developing and scaling

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions systematic methods to collect relevant effective strategies and improve student in developing new products or solutions, their approaches. These strategies should learning data that can be regularly achievement.188 In the case of Amazonas, plus the subsequent experimentation be employed across the entire ecosystem used by all actors across the education SEDUC had the mandate and freedom to that allows the most successful ideas to of actors to create a shift in education to ecosystem. These systems include develop initiatives, such as Media Center, adapt and go to scale. Across the case this kind of culture. classroom-level learning data that as students test score and pass rates teachers can use to help each student all were monitored and held accountable to the way up to national-level assessments national targets. “We still lack a lot of basic information in the education sector, which is particularly shocking when compared with the health sector. The explanation is straightforward, though: there was and there is an underinvestment in data in the education sector. There is also no doubt that the data gap has a negative impact on the level of Prioritizing education R&D international funding for education.”185 R&D has been identified in many research may require research centers— — Jean-Marc Bernard, Global Partnership for Education sectors as a key driver of innovation and not program evaluation—embedded in improvement. In comparing the scaling of countries with the farthest to go to meet education to health, Colette Chabbott, EFA [Education For All] goals and with professor at George Washington close ties to research centers of excellence University, argues that, “Early investments in other countries.”189 According to 2008 Strengthening data collection in science enable faster scaling up later. data from Organisation for Economic Co- The education sector has not made operation and Development countries, systems on learning necessary investments in the types of public R&D spending in the sectors of basic research that can improve the health and defense is 15 times the level Data, and particularly data on learning experimentation. This is especially conventional, slow-expanding model of spending on education.190 Hungary, for outcomes, are the foundation of an important for systems focused on scaling of primary school and/or provide the example, earmarks 73 times as much of education ecosystem that fosters a effective approaches, in order to identify scientific foundation for simple, game- its public research budget for health as it culture of evidence, research, and which interventions improve learning changing innovations. That sort of does for education.191

122 123 Some of the most impactful R&D processes occur at the and evaluation officers in every country that implementers can work immediately local level through micro-studies and systems that collect in which it implements its direct services. and directly with schools on program Initial investments in research contributed improvement. This approach also allows and feed data immediately back into the system. to the development of tools and training some information to feed up into the larger for school-level implementers to collect country-level monitoring system to explain strategic information about program trends and modify program content or Similarly, sustained investment in within each critical component of the implementation and outcomes in ways implementation strategies as necessary.196 agriculture R&D was found to be vital education system, among practitioners, to developing country’s agricultural policymakers, and funders. development. As the Millions Fed study identified, “The critical role of long- When considering R&D in education, term public investment in science and the key is that it be undertaken on a technology plays out across the entire systematic basis and used to devise new developing world, from Asia to Latin applications. Cases reviewed effectively America and Sub-Saharan Africa and translated research discoveries into a range of successes from major food improved practices. This has been seen crops such as rice, wheat…to livestock in the partnership over the past decade and fisheries.”192 between MIT’s Abdul Latif Jameel Poverty

Millions Learning: Scaling up quality education in developing countries in developing up quality education Scaling Learning: Millions Action Lab (J-PAL) and Pratham in India. One theory as to why sectors such As results are gathered by J-PAL, an as health receive more investment is organization devoted to rigorous impact that the outcomes are more easily evaluation, from experimenting with measured and the impacts are more different variations of the teaching at the apparent to decision-makers, unlike the right level model, these lessons have been results of poor education investment.193 incorporated into Pratham’s work and Educational researchers Nora Sabelli the menu of options available for state and Dede suggest a combination of governments as they consider adopting efforts to increase educational research the approach. Bridge International and make it more practical. They Academies invested large sums of capital emphasize that reform is an iterative into R&D before its first pupil was even process, requiring long-term investment admitted. The company found that it and also an “interplay between was “imperative to continue this rigorous theory and experiment”—particularly development process until the highest involving practitioners in the process.194 levels of academic performance and Therefore, a culture of R&D must be operational effectiveness are achieved present in research institutes, as well as so that the 1st, 100th and 1000th pupils receive the same level of education as the 100,000th and the 1,000,000th.”195

When considering R&D in Some of the most impactful R&D education, the key is that it be processes occur at the local level through undertaken on a systematic micro-studies and systems that collect and feed data immediately back into basis and used to devise new the system. Room to Read, for example, applications. has invested in research, monitoring,

124 125 126

Millions Learning: Scaling up quality education in developing countries IX in achangingworld adaptive education ecosystems The needfor inclusive and THE FUTURE: TOWARD

127 128

Millions Learning: Scaling up quality education in developing countries its descendants didfor muscle power.” environment—what the steam engine and brains andunderstand theshapeof our for mental power—the abilitytouseour and otherdigitaladvances are doing technologists from MIT, claim,“computers and Andrew McAfee, optimistic Second MachineAge future filledwithuncertainty. In Either way, theevidence points toa paradigm isathand. computing power asproof thatanew Others point toexponential increases in always faced highdegrees of uncertainty. since thebeginningof time, peoplehave that astechnologyhasbeenevolving systems for betterorworse. Someargue that have thepotential todisruptcurrent rapid globalization, andsocietalchanges of significant advances intechnology, We often hearthattheworld isonthecusp scaling from themargins tellusthen? the 14ingredients andthenotionof in educationonthehorizon? Whatcan toward thepotential for atransformation several decades.Whatifwe lookforward practices into educationover thepast and policiesthatbrought newideasand have looked retrospectively toprograms The case studies discussed in this report for enablingscalingof qualitylearning. combinations dependingonthecontext, have allbeeninstrumental, indifferent The 14core ingredients we have identified in achangingworld. The needfor inclusive andadaptive education ecosystems Toward thefuture: , Erik Brynjolfsson The The 197

in schoolingandthe promise of new The combinationof massive failures will looklike for future generations. Nor can we predict what education needs of thefuture. innovating themselves tobest meetthe to changeby adapting,learning,and ecosystems willneedtoberesponsive with key academicknowledge, education global citizens. To foster theseskillsalong information, nottomention begood and analyze increasing amounts of new knowledge andskills,filter lifelong learn learners who can adapt, like, we doknow thatthey willrequire and societiesof thefuture willlook While we cannotpredictwhatthejobs Levy, andMurnane have quantified. hollowing outof thejobmarket thatAutor, from taxidrivers todoctors, leadingtothe thought of asrequiring ahuman element, professions thathave traditionally been This isleadingtotheautomationof power.” descendants didfor muscle what thesteam engineandits shape of ourenvironment— brains andunderstand the power—the abilityto useour advances are doingfor mental “Computers andother digital 198

around. ratherto solvethan theotherway withit, then lookingfor aneducationalproblem trap of choosing atechnologyfirst and interventions have often fallen into the educational outcomes, this approach hasyieldedfew useful politically appealing.Unfortunately, joining the21 computers are a powerful symbol of policies. the priorityinmost developing countries’ placing computers inclassrooms was technologies ineducationshowed that Africa of information andcommunication A 2007 World Banksurvey across ecosystem inthefuture isstill unknown. The role of technologyinchangingthis created those goodschools.” instead of theecosystem conditionsthat to strive toimitatebest practiceschools response to low learning isto them. “The on the underlying factors that gave rise they focus onbest practicesrather than allow for disruptive innovation because describes how educationsystems donot break withcurrent practice. innovation,” orthose thatcompletely ChristensenClayton would call“disruptive have notbeen what innovation scholar To date, most innovations ineducation needed anymore. are wondering ifclassrooms are even need tobea“sage onthestage,” many real time, andwhenteachers nolonger level andassessthestudent’s progress in when programs canadapttoastudent’s potential tobeafew keystrokes away, to the world’s best professors has the know it is even necessary. Whenaccess people askingwhetherschoolingaswe different learningenvironment hasmany methods andtechnologiestocreate a 202 More thanradio ortelevision, st century andhencevery 199

203 for these 200 201

Pritchett experimentation tothrive, andthen that provide spacefor innovation and forcalls educationecosystems This 21 meet thedemandsof learning in the commit to thestep changerequired to we recommend that allstakeholders and adaptive education ecosystems, To move to thisnew normof inclusive reaching alltheworld’s children. to theneedsoffuture anuncertain and are inclusive andadaptive, responding norm ineducationwhere ecosystems from recommendations basedonthefindings or policies. What follows here are andmimicspecificprogramsto identify innovations toscalerather thantrying the digitalfuture thatallow for effective Bank callsasetof “analog foundations” for ecosystem conditions Pritchett callsfor, developing the report demonstrate isexactly what What thecasestudies explored inthis advances toclose thedigitaldivide. to bedoneinscalingandspreading these of theworld’s poorest,there ismuch work great potential for transforming thelives hold new technology and connectivity people donotown aphone. mobile phonepenetration, nearly2billion unaffordable.” Even with rapid increases in remains unavailable, inaccessible, and of theworld’s internet population,“the marginalized far behind. For 60 percent have leftthepoorest andmost that most technological advancements rather thandigitaldivides,yet found focused oncreating “digital dividends” Bank’s 2016 World Development Report failure of technologicalprogress. TheWorld Widening equitygaps represent another st -century. Millions Learning

, orwhattheWorld for creating anew 204 So while Sowhile 129 130

Millions Learning: Scaling up quality education in developing countries ensuring thatthemost marginalized all assetstheseactorsbringaswell as must beinclusive andadaptive, leveraging sector, toacademia.Theseecosystems civil societyorganizations, totheprivate from households,tocommunities, to effectively contribute their expertise— environment where allactorscan but they must also actively foster an every child’s right toaqualityeducation, is ittheirresponsibility todeliver on pivotal role inthisecosystem. Not only improve learning.Governments playa ideas orapproaches thatmost effectively actively helpfacilitate the spread of new and adaptive educationecosystems are: five neededtocreate actions inclusive healthy, safe, andproductive lives. The necessary skillsandcompetenciestolead that millionsof children are learningthe practices andapproaches toensure recommendations to scale effective We offer the following five adapt towhatever thefuture holds. education ecosystem ready toface and Moreover, it is key to developing a nimble of whathassuccessfully scaledtodate. way to move forward basedonevidence children are reached. Thisisthebest Millions Learning We needinclusive andadaptive education AND SCALING DATA WHAT WORKS LEARNING THROUGHBETTER MEASURE ANDLEARN THE MIDDLEPHASE FUND TALENT ANDEXPERTISE OUTSIDETHECLASSROOM ACTIVATE NEW IDEASTHROUGHANETWORK OFIDEAHUBS SHARE A CULTURE OFR&D INEDUCATION DEVELOP ecosystems that: FIGURE 9.

recommendations

131 132

Millions Learning: Scaling up quality education in developing countries • • • education problems. This includes: space, andsupportfornew trying ways to tackle persistent Leaders ingovernment should provide fullencouragement, of R&D ineducation: ethos across theeducationcommunity. To develop ashared culture must extend beyond individualinitiatives andbecomeashared inclusive andadaptive educationecosystem, thisculture of R&D the formal schoolsystem inBangladesh.But totrulydevelop an with anaccelerated learningprogram tobringmore children into through BRAC peerlearning.SodidtheNGO whenexperimenting to experiment withnewways of theirown supporting development culture of R&D whenitsLesson Studypracticeempowered teachers develop andscalesolutions.TheZambian government embraced a are flexible andcreative, adaptingtodifferent contexts asthey on changestopracticeandpolicy. Alltheactorsinecosystem evaluated, failures are honestly shared, anddatainform decisions learning outcomesfor marginalized children, are triedand approaches toseeminglyintransigent problems, such asimproving R&D isonewhere ideasare exchanged openly, neweducation of R&D within the education ecosystem is a key step. A culture of approaches tosolvingproblems atscale. Building a strong culture government tocivilsocietybusiness—must embrace new Leaders across of theeducationecosystem—from allparts A CULTURE OFR&D INEDUCATION DEVELOP digital booksthroughout thecountry’s publiclibrary network. working withWorldreader toincorporate e-readers loadedwith flourish. The Kenyan government’s National Library Service is Providing infrastructure support secondary schoolreaching young peoplelivinginrural areas. Aprendizaje Tutorial program who teach in the alternative government pays thesalariesof tutors in theSistema de Funding new approaches to distance learning,which gave riseto theMedia Center model. allowed thestate ministry inAmazonas totest anewapproach tried. Thenationaleducationministry inBrazil, for example, Adapting rules andpolicies

that holdpromise. TheHonduran

to allow anewapproach tobe thathelps newapproaches world andfacilitate toidentify adoptionof innovations. publicsector from innovation units that exist in some governments around the The formation of Idea Hubs around learningcoulddraw lessons approaches andconditionsthatfacilitate learningatlarge scale. Idea Hubs, contributing to a global knowledge bank of effective for experiences andlessonsshared amongregional andnational These hubs shouldbelinked through aglobalnetwork toallow that isincreasingly usedasatooltoaddress educationalproblems. to future changes,such asnear-constant evolution intechnology tohelpgovernmentespecially important decision-makers adapt funding, orcapacitybuildingtodoso.support, Idea Hubs willbe in more contexts orcommunities, butrequire additionaltechnical homegrown innovations thatholdpromise toimprove education involvement. effortlocal, toidentify Thisshouldincludeaconcerted that could usefully be scaled up by government or with government to review evidence around theapproaches, theones andtoidentify decision-makers tostay uptodatewithnewapproaches beingtried, the Idea Hubs istoprovide animble mechanismfor government educators, business,civilsociety—should bediscussed.Thegoalof and scalingthem.Approaches ledby allactors—government, adapting, andsharingeffective approaches toimproving learning Leaders ingovernments shouldestablish Idea Hubs for identifying, NEW IDEASTHROUGHANETWORK OFIDEAHUBS SHARE society leaders, relevant mentoring, training, andin-kindexpertise. leaders shouldcontribute fundingaswell as,togetherwithcivil in leadership development. Philanthropy, government, andbusiness off what exists in each country and drawing from global expertise an initiative would behighlycollaborative andmultisectoral, building educators, principals, program officers, andportfolio managers. Such from senior policymakers, leaders, NGO and donor heads to of acohort leaders thatcutsacross levels andsectors, for example actively pursued. Such initiatives could seek todevelop the capacity an initiative leaders tosupport totake onthisapproach shouldbe continuous learning,and adaptation.Within eachcountry context, to embrace thisapproach, such as risktaking,experimentation, leaders atalllevels shouldbeequippedwiththecompetencies To truly shift to a culture of R&D throughout the education ecosystem, business shouldactively cultivate acohort of Learning Leaders. Leaders across philanthropy, civilsociety, government, and

133 134

Millions Learning: Scaling up quality education in developing countries to find pragmatic and sustainable ways to deploy the power of the expertise of thetechnologycommunity of withthesupport donors and other education personnel inthe developing world. appropriate ways for technology to be used to support teachers Donors and the technology community should actively develop disseminating schooldatatoparents. administrators through such as monitoring schools and activities struggling students. Thiswouldfor alsoincludesupport education for lessonplanpreparation, toreducing timeneededfor supporting management and administrative burdens, to reducing time needed Unburdening teachers cantake many forms, from reducing classroom assist tasksthatteachers withparticular would normallyperform. students, for example, by strategically deploying otherpeopleto ways. They couldensure teachers have more timetointeract with assistance toteachers andothereducationpersonnel inseveral these initiatives could strategically activate expertise to provide off theeffective ways community isalready support being tapped, other education personnel intheirrespective countries. launch boldAll-In Community initiatives to supportteachers and Governments, civilsociety, andthebusiness community should talent andexpertise from communities outsidetheclassroom: programming tomany communities where noneexists. To active outreach workers tobringqualityearlychildhooddevelopment Street channels the expertise of artists, media specialists, and parent shoulders of already overloaded teachers. Around theworld, Sesame secondary schools,takingthework of preparing off anewsubject the of entrepreneurs andbusinessleaders toteachentrepreneurship in not available. In Jordan, for example, INJAZ leverages theexpertise elevate themintheirroles, andreach children wheneducators are types of expertise canbestrategically tappedtoassist educators, and young graduates tobusinessprofessionals andparents, different From sourcecan beoneimportant of thissupport. nonprofit workers education problems, expertise from diverse actorsoutsidetheschool overburdened andrequire tangibleassistance. For addressing tough Teachers andothereducationpersonnel whoare onthefront linesare to reach, acreative of andenergy injection support isneeded. To scale quality learning, especially in communities that are hardest TALENT ANDEXPERTISE OUTSIDETHECLASSROOM ACTIVATE

Activating Building role inscalingqualitylearning. provide resources where noneexisted before canplayanimportant tools that can unburden teachers and education administrators or toeducationactors.Low-cost,support easy-to-use technological technology is a key strategy for creative injecting energy and tangible include more grades. To catalyze middlephasefunding: able tocontinue toreach many more communities andexpand to provided sufficient fundstoensure thatthe MediaCenter pilot was greater scaleacross India. TheAmazonas state government inBrazil has helpedPratham’s evidenced-based approaches toreach a much the Indo-American community. Flexible financingfrom Pratham USA foundation set up to raise awareness and mobilize resources among Pratham benefits from long-term funding from a Pratham USA, after proof of conceptbutbefore large-scale nationaladoption. financing, includingmore attention and support for thismiddlephase also requires greater coordination andsegmentation of education scaling cycle—from prototypes orideastonationalprograms. It entities, investors, and donors fund at which stage of the innovation- Bridging thismiddlephaserequires greater clarityof whatgovernment new ideasorprototypes andimplementation atanationallevel. “valley of death” toscale. Too often, there isafundinggapbetween This is crucial to help effective education approaches cross the including for core costs. Donor agencies andfoundations shouldprovide flexible support, funding for implementing effective approaches atscale. build scalingcapacity, andwhoisproviding predictable, long-term (crucially) middlephasefundingtohelpsustain issupporting and funders should know innovative whois supporting pilots, who investment strategies tobe complementary. In any given context, investors, must actively work togethertoaligntheirrespective support scaling. develop amore organized ecosystem of education fundingto Governments, donor agencies, foundations, and investors should THE MIDDLEPHASE FUND

Education funders, outsideof particularly business One strategy important tohelpeffective 135 136

Millions Learning: Scaling up quality education in developing countries used by policymakers. are learning,andnationallyaggregated dataonlearning canbe immediately usedby teachers toseethelevels atwhichchildren useful. For example, classroom-level dataonlearningcanbe anddisseminated inaway thatismaximallyshould becollected learning dataasaglobalgood.As dataare meant tobeused,they in line with the Learning Metrics Task Force recommendation on that are systematic, transparent, and housed in-country. This is governments tobuildthecapacityof nationalassessment systems regional or international assessments. A better approach is for achievement dataorresults every few years from large-scale main source of dataonstudent performance are project-specific learning. In many countries inthedeveloping world today, the needed toinform toscaleeffective actions approaches toimproving student assessment systems. Government anddonoragencies shouldstrengthen national more locations.To anduselearningscalingdata: collect approaches are adapted,spread, andtaken upby more actorsin requires understanding theprocess by whicheffective ideasand understanding isneededonhow itisdoingsoatlarge scale. This whatiscontributingto identifying children’s learning,abetter foundationlearning isanimportant toinform In action. addition country totrack globalprogress, information onwhatstudents are put into parents’ hands,todatathatare aggregated from each ecosystem. From datathatare usedby teachers, todatathatare a culture of continuous improvement across theeducation and accessibledataonchildren’s learning facilitates spreading success of allpriorrecommendations. For example, havingregular Systematic data on children’s learningis a crucial input for the to scale. organizations, isrequired tobringeffective ideasandapproaches developing core operational capacity, especiallywithin civil society for thatare theactivities most scaling.Often neededtosupport approaches scaleistoprovide flexible fundingthatcanbeused AND SCALING DATA WHAT WORKS LEARNING THROUGHBETTER MEASURE ANDLEARN

Data onstudent learningare

a focus of innovation intheassessment space. of theskillsaswell asthecompetenciesof theirstudents shouldbe offunction assessment asatooltoinform teachers’ understanding to both measurement specialists and teachers in classrooms. The problem solving.Theassessment of theseskillsposes achallenge of such skillsascollaboration, criticalthinking,innovation, and need tobeexplored of how toguidechildren inthedevelopment teaching, learning,andassessment practicestrategies. Models domains. Educators are seekingways tointegrate theseinto their will needtomeasure theirsuccess across abroad range of learning For educationecosystems toadapttheneedsof thefuture, they breadth of learningopportunities—beyond literacy andnumeracy. explore ways of teaching andassessing 21 Education actors attheglobal,national, andlocal levels should development, andotherdisciplinesmore broadly. on existing scalingknowledge, resources, andtoolsineducation, approaches andtoexchange ideasandresources. It would draw common experiences intheirefforts to expand anddeepeneffective officials, social entrepreneurs, funders, and researchers toshare also provide spacefor peer-to-peer learningamonggovernment corrections andadjustments. AReal-time ScalingLab would self-reflection andproviding moretomake opportunities course setting would approach scalingasalearningprocess, encouraging on how toscalequalitylearninginterventions. Alaboratory-type to learningasthey unfold, contributing tobuildingabodyof evidence examine anddocument theprocess of scalingeffective approaches a Real-time Scaling Lab. The research community shouldimprove data onscaling through Such a forum would provide spaceto st -century skills—or a-century skills—or 137 138

Millions Learning: Scaling up quality education in developing countries Endnotes 21. 20. 19. 18. 17. 16. 15. 14. 13. 12. 11. 10. 9. 8. 7. 6. 5. 4. 3. 2. 1.

UNESCO, “Education Transforms Lives” (Paris: UNESCO, 2013). Return,” RolnickArt andRob “Early Grunewald, ChildhoodDevelopment: Economic Development withaHigh Public James J. Heckman, “Skills,Schools,Synapses,” Discussion Paper no. May 35152008). (Bonn:IZA, Institute, 2006). James J. Economics Heckman, “The of Investing Early in Children,” Policy Briefing no.Geary 1(Dublin: UCD educationcommission.org/. The Education Commission,“The Commission,”accessedMarch 4,2016, http:// for Universal Education atBrookings, July 2015). Liesbet SteerandKatie Smith,“Financing Education: Action” Opportunitiesfor Global (Washington: Center Targets by 2030,” Policy Paper 18(Paris: UNESCO, July 2015). Education For Monitoring Report,“Pricing AllGlobal theRight toEducation: TheCost of Reaching New Systems. Discussion Prepared Document,” for theLearning Metrics Task Force 2.0, December 1,2015. “Assessment for Learning (A4L):AnInternational Platform National toSupport Learning Assessment to learn,anddeveloped asmallsetof indicators tomeasure those skills. The taskforce worked tobuildaconsensus onwhatskillsandcompetencies are for important allchildren working groups of 186technicalexperts, andmore than1,700 consultationin118countries. participants Center for Universal Education atBrookings from 2012 to2016, consisted memberorganizations, of 30 The Learning Metrics Task Force (LMTF), co-convened by theUNESCOInstitute of Statistics andthe 2013). (Montreal andWashington: UNESCOInstitute of Statistics andCenter for Universal Education atBrookings, Learning Metrics Task Force, “Toward Universal Learning: Framework AGlobal for Measuring Learning” 2016). UNESCO, World Inequality Database onEducation; Indicator: Out-of-School Children (Paris: UNESCO, Ibid. UNICEF, “Early ChildhoodDevelopment: AStatistical Snapshot” (New York: UNICEF, 2014). (Washington: Brookings Institution, June 2015). Rebecca Winthrop andEileen McGivney, Wait “Why 100Years? Education” inGlobal BridgingtheGap (Washington: Brookings Institution Press, 2013). Lant Pritchett, Twaweza, 2013). Uwezo, “Are Our Children Learning? Literacy andNumeracy Across East Africa2013” (Nairobi: Hivos/ “Annual Statusof Education Report Data Query” (New Delhi: ASERCentre, 2014). Community of Practice (Research Triangle Park, NC:Research Triangle Institute, 2011). GoveandP.A. Cvelich, E UNESCO, 2014). Education For Monitoring Report,”Teaching AllGlobal andLearning: Achieving Quality for All”(Paris: (Washington: Brookings Institution, June 2015). Rebecca Winthrop andEileen McGivney, Wait “Why 100Years? Education” inGlobal BridgingtheGap Rebecca Winthrop, “Education inAfrica:TheStoryIsn’t Over,” (Washington: Brookings Institution, June 2015). Rebecca Winthrop andEileen McGivney, Wait “Why 100Years? Education” inGlobal BridgingtheGap Fedgazett The Rebirth Schooling of Education: Ain’t Learning e, March 2003. March e, arly Reading: Igniting Education for All Current History , revised ed.,Early GradeLearning , CenterDevelopment for Global 110, no. 736 (May 2011). 38. 37. 36. 35. 34. 33. 32. 31. 30. 29. 28. 27. 26. 25. 24. 23. 22.

to Policies andProgrammes Service Innovations: AFramework for in Action,” (Washington: Brookings Institution Press, 2013); Ruth Simmons andJeremy Shiffman,“Scaling Up Health People Laurence Chandyet al., eds., (Brookings Institution, Washington, 2013). Center for Universal onLearning Education, Compact “2013Donor Global Network Mapping Results” O. Pidufala, “Scaling Up andAidEffectiveness: AnnotatedBibliography,” 2008. Washington, How to BringDevelopment Solutions to Millionsof PoorPeople Laurence Chandy, AkioHosono, Homi Kharas, andJohannes Linneditedthe2013 book, Workingof Brookings Global Papers authored by Johannes LinnandLaurence Chandy. Most recently, abookonMexico’sand Linn(2008); Progresa-Oportunidades by Santiago Levy (2006); andaseries period andbeyond hasincludedastocktaking of theliterature andexperience withscalingby Hartmann Brookings, where scalingwas the mainthematicfocus of work from to2010. 2005 Work duringthis least towhenformer World Bank president James Wolfensohn directedtheCenter for Development at A focus on scaling successful development interventions has a long history atBrookings, going back at Education Innovation inAfrica,”EdQual Working Paper 28(EdQual, September2011). Martial Dembélé, Joel Samoff,E. Molapi Sebatane, and “‘GoingtoScale’: the Nurturing Local Roots of February 2012). Implications for Australia’s Aid Program” (Canberra: Australian Agency for International Development, Organization, 2006),Ian Anderson, 1-30; “Scaling Up Development Results: A Literature Review and Chandy etal.,eds., D. Fixsen et al., “Implementation Research: A Synthesis of the Literature” (Tampa: Laurence NIRN, 2005); Experiments,” McEwan, “Improving Learning inPrimary Schoolsof Developing Countries: AMeta-analysis of Randomized Meta-analysis of Rigorous ImpactEvaluations,” submitted for Ph.D., ColumbiaUniversity, 2014; Patrick J. achievement/; Katharine Conn, “Identifying Effective Education Interventions in sub-Saharan Africa: A accessed February 18,2016, http://visible-learning.org/hattie-ranking-influences-effect-sizes-learning- “Hattie Ranking: Influences and Effect Sizes RelatedtoStudent Achievement,” 2015). John Hattie, Works “What Best inEducation: ThePolitics of Collaborative Expertise” (London: Pearson, Ain’t Learning schools andmost improved schoolsystems” series,andLant Pritchett, Development Institute’s Development Progress initiative, McKinsey’s “How theworld’s best-performing Development (DfID)-supported Research onImproving Systems of Education (RISE),theOverseas World Bank’s Systems Approach for BetterEducation Results (SABER), UKDepartment for International There isarange of initiatives thatfocus onsystem-wide reforms. Someof themore includethe important Educational Development, 1973). P. Coombs withR.Prosser andM.Ahmed,“New Paths toLearning” (New York: International Councilfor Policy Research Working Paper (Washington: 5342 World 2010). Bank, Brian Blankespoor etal.,“Adaptation toClimate Extremes inDeveloping Countries: TheRole of Education,” Journal Sentinel Kharas, Homi andRebecca Winthrop. 2016. “Want tosave theplanet? Invest ingirls’ education.” Investment Rebecca Winthrop Sperling, andGene Countries between 1970 and2009: ASystematic Analysis,” Emmanuela etal.,“Increased Gakidou Educational Attainment andIts Effect onChild in Mortality 175 paper for theWorld Development Report (Washington: World June Bank, 10, 2013). Cristian Aedo etal.,“From Occupations toEmbedded Skills:ACross-Country Comparison,”background Empirical Exploration,” David H.Autor, Frank Levy, andRichard Murnane, SkillContent of “The Recent Technological Change:An of Growth Eric Hanushek andLudger Woessmann, (Washington: Brookings Institution Press, 2013): 33. (Cambridge: MITPress Books,2016). (Washington, DC:Brookings Institution Press, 2015). (Washington: Brookings Institution Press, 2013). Review of Educational Research . February 13, 2016. Getting to Scale: How to BringDevelopment Solutions to Millionsof PoorPeople Quarterly Journal of Economics , ed.Ruth Simmons, Peter Fajans, andLaura (Geneva: Ghiron World Health Getting to Scale: How to BringDevelopment Solutions to Millionsof Poor The Knowledge Capital of Nations:Education andtheEconomics What WorksWhat inGirls’ Evidence Education: for theWorld’s Best 85, no. 3(September 2015): 353-94. Scaling Up fromPilot Health Service Delivery Innovations 118,no. 4(2003). Lancet . 376 (2010): 959-74. The Rebirth of Schooling Education: Visible Learning Getting to Scale: Milwaukee blog,

139 140

Millions Learning: Scaling up quality education in developing countries 54. 53. 52. 51. 50. 49. 48. 47. 46. 45. 44. 43. 42. 41. 40. 39.

Everett M.Rogers, error. This iswhatJohannes Linnrefers 2”(wrong toasa“type 1” scaling)vs “type (noortoolittlescaling)scaling groups#Low_income. World “Country Bank, andLending Groups”, http://data.worldbank.org/about/country-and-lending- org/Library/Documents/isced97-en.pdf. UNESCO, International Standard Classification of Education (Paris: UNESCO, 1997), http://www.uis.unesco. at Brookings, September 2013). Task Force” (Montreal andWashington: UNESCOInstitute for Statistics andCenter for Universal Education Learning Metrics Task Force, “Toward Universal Learning: Recommendations from theLearning Metrics Washington: UNESCOInstitute for Statistics andBrookings Institution, Center for Universal Education, 2013). Learning Metrics Task Force, “Toward Universal Learning: WhatEvery ChildShouldLearn” (Montreal and Laura (Geneva: Ghiron World Health Organization, 2006), 1-30. fromPilotService Delivery Innovations to Policies andProgrammes Jeremy Shiffman,“Scaling Up Health Service Innovations: A Framework in for Action,” at the Grassroots: Paths to Scaling Up,” “Scaling Up: Thinkingthrough theIssues” (Providence: Brown University, 1994); P. Uvin, “Fighting Hunger Practices,” Brookings Institution, C.J. 2008); Dunst etal.,“ScalingUp Early ChildhoodIntervention Literacy Learning Up: AFramework andLessons for Development Effectiveness from Literature and Practice” (Washington: and Shiffman(2006); Arntraud and andLinn (2008). Hartmann and Hartmann Johannes F. Linn,“Scaling These four pathways are drawn from Uvin andMiller (1994); Uvin Dunst (1995); etal.(2006); Simmons Researcher Coburn,“RethinkingCynthia E. Scale:Moving Beyond Numbers toDeep andLasting Change,” (2013). (2007); ExpandNet Hartmann and Linn (2008); andWorld Health Organization (2011); and Chandy et al. and Miller (1994); Samoff et al. (2001); MSI, and Coburn (2003); ExpandNet, World Health Organization This definitionsdraws from some of themost influential papers onscale, including Meyers (1984); Uvin Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development Alice Gugelev andAndrewAlice Gugelev Stern,“What’s Your Endgame?” Developing a Scaling-Up Strategy” (Geneva: World Health Organization, 2010). Innovations” (London: Nesta, July 2014); World Health Organization “Nine and ExpandNet, Steps for and World Health Organization, 2007; Madeleine“Making Gabriel, It Big:Strategies for ScalingSocial Questions for Developing a ScalingUp CaseStudy,” Management Systems International, ExpandNet, and Lessons” (Washington: Results for Development Institute, September 2014); Peter Fajans etal.“20 March 2006); Larry Cooley andJohannes F. Linn,“Taking Innovations toScale:Methods, Applications Kohl, “ScalingUp—From Vision toLarge-Scale Change” (Washington: Management Systems International, Other Programmatic Research for Successful ScalingUp” (Geneva: WHO, 2011); Larry Cooley andRichard ExpandNet andWorld Health Organization, “BeginningwiththeEnd inMind: Planning Pilot Projects and Johannes Linn,“Internal ConceptNote for Project onScaling Up,” October21,2013. Working Paper 28(EdQual, September2011). Molapi Sebatane,E. “‘Going toScale’: Nurturing theLocal Roots of Education Innovation inAfrica,” EdQual d’Ivoire: Association for theDevelopment of Education inAfrica,2001); MartialDembélé, Joel Samoff, and to ExpandEducational Reform Association for theDevelopment of Education for Africa”(Abidjan,Côte Martial Dembélé, Joel Samoff,E. Molapi Sebatane, and “Scaling Up by Focusing Down: Creating Space (New York: Harper, 1999). Moore, Geoffrey Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development 32,no. 6(2003): 3. Center for Early Literacy Learning (CELL) Crossing the Chasm: Marketing andSellingCrossing High-Tech theChasm: Products to Mainstream Customers Diffusion of Innovations World Development , 5thed.(New York: Free Press, 2003). Papers 1,no. 1-10; 2,2006: D. Miller andP. Uvin, Stanford Social Innovation Review 23, no. 927-39; 6 (1995): Ruth Simmons and , ed.Ruth Simmons, Peter Fajans, and Scaling Up Health (Winter 2015). Educational 75. 74. 73. 72. 71. 70. 69. 68. 67. 66. 65. 64. 63. 62. 61. 60. 59. 58. 57. 56. 55.

Randomized Experiments,” Patrick J. McEwan, “Improving Learning inPrimary Schools of Developing Countries: AMeta-analysis of Education Innovation inAfrica,”EdQual Working Paper 28(EdQual, September2011). Martial Dembélé, Joel Samoff,E. Molapi Sebatane, and “‘GoingtoScale’: the Nurturing Local Roots of a program orpolicytothemost disadvantaged rather thanattemptingnationalcoverage. also increase toapoint thatscalinganintervention becomesunviable. At times,thismight callfor targeting supervision, andadministration.transport, Where expansion mayachieve economies of scale, costs can and more difficultto reachbecomesincreasingly expensive asthere are oftenhigherunit costs of training, the case. thisisoften In notthecasewhenitcomestoeducation.Expandingcoverage fact, tothose more It isoften assumed thattheprocess of goingtoscalewillbring economiesof scale, butthisis notnecessarily Take?” PurnimaMenon, Gillespie, andAndrewStuart Kennedy, “ScalingUp ImpactonNutrition: WhatWill It Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development Countries,” Patrick McEwan, “Cost-effectiveness Analysis of Education and Health Interventions in Developing Crown Business, 2014). Robert I.Sutton andHuggy Rao, https://kmonadollaraday.wordpress.com/2013/12/18/the-next-innovation-scaling-up/. Ian Next Thorpe, “The Innovation: Scaling Up,” kmonadollaraday.wordpress.com (blog),December 18,2013, the Brookings Institution, Washington, June 2014). Larry Cooley, “Millions Learning: Role of Partnerships inBringingEducation Initiatives toScale” (meetingat PESRP, “SchoolCouncils,” http://www.pesrp.edu.pk/pages/School-Councils. youtube.com/watch?v=gjATVO3l-s0. YouTube, “Punjab, Pakistan, Education Reform Roadmap,” published onJanuary 22,2014, https://www. Cory Heyman, emailcommunication withJenny Perlman Robinson, February 17, 2016. 2013). “World Development Report Making 2004: ServicesWork for Poor People” (Washington: World Bank, People Laurence Chandyetal.,eds., Ibid. Scott Macmillan, “Glorious Failure: TheJoy of Learning from Your Mistakes,” wright. blogs/education-plus-development/posts/2014/04/23-colombia-southeast-asia-learning-gustaffson- Learn from ColombiatoSoutheast Asia,” Brookings.edu (blog),April23, 2014, http://www.brookings.edu/ Emily Gustafsson-Wright andEileen McGivney, “Fundación Escuela Nueva: ChangingtheWay Children no. 3(September 2015). Developing Countries: AMeta-analysis of Randomized Experiments,” (Nashville: Vanderbilt University Press, 2012); Patrick J. McEwan, “Improving Learning inPrimary Schoolsof Erin Murphy-Graham, Market-based Poverty” SolutionstotheChallengesof Global (Cambridge: Monitor Group, March 2009). Ashish Karamchandani, Michael Kubzansky, and Paul Frandano, “Emerging Markets, Emerging Models: Social Innovation Review Jeffrey“Transformative BradachGrindle, andAbe Scale:The Future Growingof What Works,” -Driven Development Approaches,” OEDWorking Papers (Washington: World 2005). Bank, Development for anEvaluation of theEffectiveness of World Bank Support for Community-based and Barbara Pozzoni andNalini Kumar, “A Review of theLiterature onParticipatory Approaches toLocal (Washington: Brookings Institution Press, 2013). Advances inNutrition Journal of Development Effectiveness Opening Minds, Improving OpeningMinds, Lives: Education andWomen’s Empowerment In Honduras (February 19, 2014). Review of Educational Research 6, no. 4(2015): 440-51. Getting to Scale: How to BringDevelopment Solutions to Millionsof Poor Scaling Up Excellence: Getting to MorewithoutSettling for Less 4,no. 2(June 2012): 189-213. 85, no. 3(September 2015). Review of Educational Research The Guardian , March 30, 2015. (New York: Stanford 85, 85, 141 142

Millions Learning: Scaling up quality education in developing countries 98. 97. 96. 95. 94. 93. 92. 91. 90. 89. 88. 87. 86. 85. 84. 83. 82. 81. 80. 79. 78. 77. 76.

Barbara Bruns and Javier Luque, Philippines,” IFPRI Discussion Paper (Washington: International Food Policy Research Institute, 2011). and Yanyan Liu,“Girls Take Over: Long-term Impacts of an Early StageEducation Intervention inthe Measures of Teacher Quality,” no. 627-42; 5(2008): Hanushek Eric A. andSteven G. Rivkin, “Generalizations aboutUsing Value-Added Achievement: Evidence From Five Indian States,” Amita Chudgar andVyjayanthi Sankar, Relationship “The between Teacher andStudent Gender Madeleine “Making Gabriel, It Big:Strategies for ScalingSocialInnovations” (London: Nesta, July 2014). Anna Davies, “Spreading SocialInnovations: 2015. ACaseStudyReport,”TEPSIE, Management Systems International, March 2006). Larry Cooley andRichard Kohl, “ScalingUp—From Vision toLarge-Scale Change” (Washington: Madeleine “Making Gabriel, It Big:Strategies for ScalingSocialInnovations” (London: Nesta, July 2014). Robinson, February 17, 2015. http://www.roomtoread.org/AnnualReport/2014/; CoryHeyman, emailcommunication withJenny Perlman Room toRead, “SolvingthePuzzle: Children’s Literacy andGirls’ Education,” Annual Report,2014 ed., ofteacher appointed toguidetheactivities center andtoprovide toteachers. academicsupport is usedfor teachermeetings,training, andadministrative Thecoordinator work. isusually asenior A cluster isusually agroup of 10-15 schoolsinaconfinedgeographic area. The Cluster Resource Center Bangladesh: Academy for Educational Development, July 1992). Rahman Rahman Huq, “Cost Comparisonof BRAC NFPandGovernment Primary Schools” (Dhaka, and meals. per month are closer to$16-$20 whenfactoring infees for such thingsasexams, uniforms, schoolsupplies, The average cost of BridgeInternational Academies hasbeencontested withsomeclaimingaverage costs bridgeinternationalacademies.com/approach/reach/. Bridge International Academies, “Reach,” accessedAugust 24, 2015, http://www. Personal communication, Vicky andRebecca Colbert Winthrop, November 5, 2015, Doha, Qatar. People Laurency Chandyetal.,eds., Kevin Carey andAndrew P Kelly (Cambridge: Harvard Education Press, 2013). in Michael Staton,“Unbundling Higher Education: Taking theComponents of Apart theCollege Experience,” 22, 2015. Interview withteacheratBridgeInternational Nairobi, Academy Jenny inGicagi, Perlman Robinson, April Rukmini Banerjiinterview by Jenny Perlman Robinson July, andShushmita ChatterjiDutt, 18,2015. International Institute for Educational Planning, 2008). Mark Bray, “Double-shift Schooling:Design andOperation for Cost-effectiveness” (Paris: UNESCO 2030,” UISFact Sheetno. 27 (Montreal: UIS2013). UNESCO Institute of Statistics, “A Teacher for Every Teacher Child:Projecting Global Needs from 2015 to More Teachers,” UISFact SheetNo. 33(Montreal: UIS,2015). UNESCO Institute of Statistics, “Sustainable Development For Goal Education CannotAdvance without Policy Report (Washington: Center for Universal Education atBrookings; CfBT Education Trust, 2012). Janice Dolan etal.,“Building Effective Teacher Salary Systems In Fragile andConflict-Affected States,” 2010). Militancy inPakistan,” Working Paper no. 2(Washington: Center for Universal Education atBrookings, Rebecca Winthrop andCorrinneGraff, “Beyond Madrassas: Assessing theLinksbetween Education and UNESCO, 2014). Education For Monitoring Report,“Teaching AllGlobal andLearning: Achieving Quality for All”(Paris: Frances Vavrus, Caribbean Stretching the Higher Education Dollar: How Innovation Can Improve Access, Equity, and Affordability, (Washington: Brookings Institution Press, 2013). (Advance ed.)(Washington: World 2014). Bank, Desire andDecline: Schooling Amid Crisis in Tanzania American Economic Review Getting to Scale: How to BringDevelopment Solutions to Millionsof Poor Great Teachers: Great How to Raise Student Learning inLatin America andthe Journal of andInternational Comparative Education 100, No. 2(May 2010): 267-71; Futoshi Yamauchi (New York: Peter Lang, 2003). 38, 38, ed. 118. 117. 116. 115. 114. 113. 112. 111. 110. 109. 108. 107. 106. 105. 104. 103. 102. 101. 100. 99.

Robert Hill, “Teach First: Ten Years of Impact” (London: Teach First, 2012). web.pdf. Teach First, “2014 Annual Report”, 2014, https://www.teachfirst.org.uk/sites/default/files/ar/pdf/TF_AR_ Sarabeth Berman,e-mailmessagetoJenny Perlman Robinson andEileen McGivney, January 4, 2016. (Washington: Center Development, for November Global 30, 2004). Ruth Levine, What Works Working Group, Molly Kinder, Keep Better,” Getting McKinsey andCompany, November 2010. Michael Barber, Chinezi Chijioke, andMona Mourshed, “How theWorld’s Most Improved SchoolSystems Erik Ramirez-Ruiz indiscussionwithJenny Perlman Robinson andSamantha Spilka, July 24, 2015. Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud Hartmannand Johannes F. Linn,“ScalingUp: AFramework andLessons for Development Center Development, for November Global 2004). What Works Working Group, Molly Kinder, Institute, “Development Progress Series” (London: Overseas Development Institute, 2011-2012); Ruth Levine, Development Millions Fed Series” Leadership was identified asa key driver intheOverseas Development Institute’s Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development www.gainhealth.org/sites/www.gainhealth.org/files/NaandiLR.pdf. Businesses: Naandi Foundation’s SchoolFeeding Program,” Alliancefor Improved Global Nutrition, http:// InnovationsGlobal Fund.com, https://www.globalgiving.org/pfil/6853/projdoc.pdf; Naandi, “Lessons from children-cannot-learn/; Naandi, “Naandi Foundation’s Centralized Kitchens for Midday Meal Programme,” “Our Work,” David Risher indiscussionwithJenny Perlman Robinson, December 1,2014. Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development (Stockholm: Sida, 2001). OstromE. etal.,“Aid, Incentives, andSustainability: AnInstitutional Analysis of Development Cooperation” et al.(2001). Cooley(2008), andKohl (2006), Simmons andShiffman(2007), BradachGrindle(2014), andOstrom and These principlesare drawn from casesandliterature reviewed, includingresearch by HartmannandLinn People Laurency Chandy et al., eds., 2007). Matters, andHow It CanBeAccelerated,” Working Paper (Oxford: Skoll Centre for SocialEntrepreneurship, Geoff Mulgan with Simon Tucker, Rushanara Ali,andBenSanders, “Social Innovation: What It Is, Why It People Laurency Chandy et al., eds., February 9, 2016. More Learning Ahead),”internal document shared withJenny Perlman Robinson andEileen McGivney, Sharath Jeevan, Do We “What Know aboutTeacher Motivation? .Our Learning atSTIR SoFar (with Much DoQSS Working Paper 10-18 (London: Institute of Education, University of London, 2010). Caste, Religious, Distance orGender between Student andTeacher Matter? SomeEvidence from India,” Reforms 10(Washington: World 1998); Bank, ShenilaRawal Kingdon, “Akin andGeeta toMy Teacher: Does The Community Program Support inRural Balochistan,” Working Paper onImpactEvaluation of Education Jooseop Kim, Harold Alderman,andPeter Orazem, “Can Cultural Barriers BeOvercome inGirls’ Schooling? 22, 2016. with ChrisMurphy, Vice President, Marketing andCommunications atRocketship Education onFebruary Preston SmithindiscussionwithRebecca Winthrop andEileen McGivneyonDecember 2,2015. Confirmed , Center Development’s for Global (Washington: Brookings Institution Press, 2013). (Washington: Brookings Institution Press, 2013). . David J. SpielmanandRajul Pandya-Lorch, eds., Naandi (Washington: International Food Policy Research Institute, Overseas 2009); Development , accessedMarch 2,2016, http://www.naandi.org/programmes-phased-out/hungry- Getting to Scale: How to Bring Development Solutions to Millions of Poor Getting to Scale: How to Bring Development Solutions to Millions of Poor Millions Saved Millions Saved: Proven Successes inGlobal Health , andInternational Food Policy Research Institute’s Millions Saved: Proven Successes in Global Health Millions Fed: Proven Successes for Agricultural “Development Progress (Washington: 143 144

Millions Learning: Scaling up quality education in developing countries 140. 139. 138. 137. 136. 135. 134. 133. 132. 131. 130. 129. 128. 127. 126. 125. 124. 123. 122. 121. 120. 119.

based Educational Improvement Chris Dede, James P. Honan, Laurence C.Peters. etal., Preeminence of Corporate Culture,” Gerard J. Tellis, Jaideep C.Prabhu, andRajesh K.Chandy, “Radical Innovation across Nations: The the Brookings Institution, Washington, June 2014). Larry Cooley, “Millions Learning: Role of Partnerships inBringingEducation Initiatives toScale” (meetingat Validity inKenyan Education,” Working Paper 321 (Washington: CenterDevelopment, for March Global 2013). teacher performance. Tessa Boldetal.,“ScalingUp WhatWorkings: Experimental Evidence onExternal union, adversely affecting incentives for government-employed teachers to therebyexperiment, weakening 18,000 newcontract teachers into theworkforce spurred organized resistance from Kenya’s nationalteachers student test scores. In terms of thelatterissue,affecting the program’s potential tointroduceultimately delays, in the government treatment arm received fewer monitoring visits and teachers underwent longer salary teacher intervention underwent several monitoringandevaluation obstacles. For example, Kenyan schools are implementation restrictions andpoliticaleconomy limitations.In termsof theformer issue, thecontract According toTessa Boldandcolleagues,two possible explanations for theintervention’s scalingfailure Paper 10, andDevelopment CommissiononGrowth (Washington: World 2008). Bank, Daron Acemoglu andJames Robinson, Role “The of Institutions andDevelopment,” inGrowth Working Ibid. Sapiens3,Management,” no. 1(2010). R. Bille, “Action withoutChange? On theUse andUsefulness of Pilot Experiments inEnvironmental stireducation.org/. “STIR Education: Empowering Teachers toBeChangemakers,” and Japan International Cooperation Agency, 2015), 6. Through Lesson Study—Implementation 5thed.(Lusaka, Guidelines,” Zambia: Ministry ofEducation General Zambia Ministry Education, of General “School-BasedContinuing Professional Development (SBCPD) claudia-costin-perlman-robinson. 7, 2010, www.brookings.edu/blogs/education-plus-development/posts/2015/01/07-education-reforms- Claudia Costin, SeniorDirector for Education attheWorld Brookings Institution Bank,” (blog),January Jenny Perlman Robinson, “Ensuring Lasting Education Reforms by Delivering Results: ADiscussion with Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development Rukmini BanerjiindiscussionwithJenny Perlman Robinson July andShushmita ChatterjiDutt, 18,2015. S. Fisch andR.Truglio (Mahwah, NJ: Erlbaum, 2001). J. G. Cooney, Foreword in Education Innovation inAfrica,”EdQual Working Paper 28(EdQual, September2011). Martial Dembélé, Joel Samoff,E. Molapi Sebatane, and “‘GoingtoScale’: the Nurturing Local Roots of Program” (Canberra: Australian Agency for International Development, February 2012). Ian Anderson, “ScalingUp Development Results: ALiterature Review andImplications for Australia’s Aid (Washington: Brookings Institution Press, 2013), 138-64. Michael eds., Barber andSaadRizvi, Problems aroundtheWorld Adele Cassola and Jody Heymann, general/7553-education-to-be-revised. Lewis S.Teh, “Education toBeRevised,” unesco.org/Education/Documents/OOSCI%20Reports/brasil-oosci-summary-2012-en.pdf. UNESCO, “GlobalInitiative onOut-of-School Children: Brazil” (Paris: UNESCO, August 2012), http://www.uis. Laurency Chandy et al., eds., opinionator.blogs.nytimes.com/2011/09/29/making-change-happen-on-a-deadline/?_r=0. Tina Rosenberg, “Making Change Happen, on a Deadline” Everett M.Rogers, Diffusion of Innovations “G” Is for Growing: Thirty Years of Research onChildren and“Sesame Street,” (Oxford: Oxford University Press, 2012). Getting to Scale: How to BringPeople DevelopmentSolutionsofPoor toMillions (San Francisco: Jossey-Bass, 2005). Asking More:The to Path Efficacy Journal of Marketing Lessons in Educational Equality: Successful Approaches to Intractable The New Dawn , 5thed.(New York: Free Press, 2003). Scaling Up Success: Lessons Learned fromTechnology- , June 4,2015, http://www.thenewdawnliberia.com/ 73, 2009. New York Times (Pearson, 2013). STIR Education (blog), September 29, 2011, http:// , 2015, http://www. ed.

163. 162. 161. 160. 159. 158. 157. 156. 155. 154. 153. 152. 151. 150. 149. 148. 147. 146. 145. 144. 143. 142. 141. Development of Indonesian Preschool Children: An Experimental Study,” D. G. L. Borzekowskiand H.K.M.Henry, Impactof Jalan “The SesamaontheEducational andHealthy education/. Sesame Workshop, “Early Education,” http://www.sesameworkshop.org/what-we-do/our-initiatives/early- (Washington: Results for Development Institute, September2014). Larry Cooley andJohannes F. Linn,“Taking Innovations toScale:Methods, ApplicationsandLessons” (Washington: Center Development, for November Global 2004). Alice Gugelev and Andrew Stern, “What’s Your Endgame?” Biggest inIndia,” Gaps (McKinsey andCompany, October2013). Ramesh Mangaleswaran andRamya Venkataraman, “Designing Philanthropy tothe for Impact:Giving Institution, 2012), http://www.brookings.edu/research/reports/2013/02/brooking-blum-roundtable-2012. the Fight against Poverty,” reportfor Laurence Chandy, Kemal Dervis, andSteven Rocker, “Clicks into Bricks,Technology into Transformation, and Development Ruth Levine, What Works Working Group, Molly Kinder, Madhav Chavan indiscussionwithJenny Perlman Robinson July andShushmita ChatterjiDutt, 16, 2015. Factors.” B. Bandaetal.,“Lesson StudyPractice of ScienceTeachers inZambia: Its Effects, Enhancing and Hindering Muhannad al-Jarrah indiscussionwithMayyada AbuJaber onApril6, 2015. People Laurence Chandyetal.,eds., GIZ, “ScalingUp inDevelopment Cooperation: Practical Guidelines” (Eschborn, Germany: GIZ,2010). emailcommunication withJennyColette Chabbott, Perlman Robinson, February 3, 2016. Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development Experience,” Rural Development” (Washington: World 2003); S.AhmedandM.French, Bank, “ScalingUp: TheBRAC Practice Spector, eds., Brookings Institution, Hans 2008); P. Hans P. Binswanger-Mkhize, Jacomina P. deRegt, andStephen A Framework andLessons for Development Effectiveness from Literature and Practice” (Washington: and French Binswanger 2006; andAiyer 2003). Arntraud HartmannandJohannes F. Linn,“ScalingUp: longer. Otherseminalstudies found scaling typically extends over 10-15 years (Hancock 2003; Ahmed foundHartmann andLinn(2008) successfully scalingprograms tonationalscalecantake 5-10 years or org/blog/posts/friday-note-zero-year. Ruth Levine, “Friday Note: TheZero Year,” Hewlett Foundation (blog),July 10, 2015, http://www.hewlett. 11, 2015. Alex Daniels, “Ford ShiftsGrantmaking toFocus Entirely onInequality,” The Overhead Myth, “Moving toward anOverhead Solution,”http://overheadmyth.com/. Grantmakers for Effective Organizations, 2014).February GEO, Philanthropy “Smarter for Greater Impact: Rethinking How Grantmakers Scale” Support (Washington: Verne Harnish, Ibid. People Laurency Chandy et al., eds., claudia-costin-perlman-robinson. 7, 2010, www.brookings.edu/blogs/education-plus-development/posts/2015/01/07-education-reforms- Claudia Costin, SeniorDirector for Education attheWorld Brookings Institution Bank,” (blog),January Jenny Perlman Robinson, “Ensuring Lasting Education Reforms by Delivering Results: ADiscussion with (Washington: Brookings Institution Press, 2013). (Washington: Brookings Institution Press, 2013). (Washington: World J. 2009); Bank, “ScalingUp Hancock, Practices theImpactof in Good New BRAC University Journal 35(2011): 169-79.

Scaling Up Local Community DrivenDevelopment: AReal World Guideto Its Theoryand Perspectives inScience Education Scaling Up: How Few Companies Make It and Why the Rest Don’t Getting to Scale: How to Bring Development Solutions to Millions of Poor Getting to Scale: How to BringDevelopment Solutions to Millionsof Poor Brookings Blum Roundtable onGlobal Poverty , 2003. , , 3rd ed.(Florence, Italy: 2014). Millions Saved: Proven Successes in Global Health Stanford Social Innovation Review International Journal of Behavioral Chronicles of Philanthroph (Ashburn: Gazelles, 2014). (Washington: Brookings (Winter 2015). y, June

145 146

Millions Learning: Scaling up quality education in developing countries 187. 186. 185. 184. 183. 182. 181. 180. 179. 178. 177. 176. 175. 174. 173. 172. 171. 170. 169. 168. 167. 166. 165. 164.

Ibid. Shifting theEducation Agenda” (Washington: Results for Development Institute, June 2015). Molly Jamieson Eberhardt etal.,“BringingLearning toLight: TheRole of Citizen-Led Assessments in www.asercentre.org/p/76.html. ASER Centre, “Citizen-led, Household BasedAssessments around theWorld” (New Delhi: ACER Centre), html. “Overview,” ASERCentre, 2016, http://www.asercentre.org/Survey/Basic/Pack/Sampling/History/p/54. www.asercentre.org/p/76.html. ASER Centre, “Citizen-led, Household Based Assessments around the World” (New Delhi: ASER Centre), David Herbling, “Ministry Stops Expansionof Low-Cost Schools,” Laura Lewis, emailcommunication withJenny Perlman Robinson, March 18,2016. Report Jacques Pelkmans andAndrea Renda, “Does EURegulation Hinder orStimulate Innovation?” Frederick M.Hess, 2013). Lant Pritchett, Social Innovation Review Jeffrey“Transformative BradachGrindle, andAbe Scale:The Future Growingof What Works,” GIF. “Transforming Ideas into Impact.” Adam Davidson, “SavingtheWorld, Startup-style,” Effectiveness, OECD(2005). Agreement for Effective Development Co-operation, OECD(2011); and Paris Declaration onAid These international agreements includeAccra Agenda for Busan OECE(2008); Partnership Action, Ibid. Effectiveness from Literature and Practice” (Washington: Brookings Institution,2008). Arntraud HartmannandJohannes F. Linn,“ScalingUp: AFramework andLessons for Development in Brazil: TheNext Agenda” (Washington: World 2012). Bank, and Development, 2011); Barbara Bruns, David Evans, and Javier Luque, “Achieving World-Class Education Strong Performers and Successful Reformers in Education,” (Paris: Organisation for Economic Co-operation Organisation for Economic Co-operation andDevelopment, “Lessons from PISA for theUnited States: proposed Partnership by theGlobal for Education. spearheaded at the UNESCO Institute of Statistics, and the Assessment for Learning (A4L) platform academics. Two initiatives are being developed to address thisgap, Alliance for Learning, theGlobal approaches andfoster collaboration withinandacross countries amongteachers, policymakers, and Robust learningdataare apublicgoodthatcanbeusedtoresearch anddevelop effective education August 5, 2014, www.globalpartnership.org/blog/data-not-just-please-statisticians. Jean-Marc Bernard, “Data Are Not Just toPlease Partnership Statisticians,” Global for Education (blog), Strong Performers andSuccessful Reformers inEducation,” (Paris: OECD, 2011). Organisation for Economic Co-operation andDevelopment, “Lessons from PISAfor theUnited States: data-catalog/world-development-indicators. World “World Bank, Development Indicators: Brazil” (Washington: World Bank),http://data.worldbank.org/ Africa” (Abidjan,Côted’Ivoire: Association for theDevelopment of Education inAfrica,2001). Down: Creating SpacetoExpandEducational Reform Association for theDevelopment of Education for These roles are basedonMartialDembélé, Joel Samoff,E. Molapi Sebatane, and“Scaling Up by Focusing Zambia. Personal communication withDarius Mogaka Ogutu andRebecca Winthrop, February 5, 2015. Livingstone, Ibid. (Brussels:Centre for European Policy Studies,November 2014). The Rebirth of Schooling Education: Ain’t Learning Same ThingandOver Over (February 19, 2014). Global Innovation Fund (Cambridge: Harvard University Press, 2010). New York Times . http://www.globalinnovation.fund. Business Africa Daily (Washington: Brookings Institution Press, , November 17, 2015. , September24, 2015. CEPS Special Stanford 204. 203. 202. 201. 200. 199. 198. 197. 196. 195. 194. 193. 192. 191. 190. 189. 188.

World “World Bank, Development Report 2016: Digital Dividends” (Washington: World 2016). Bank, Classroom intheDeveloping World” (Washington: Brookings Institution, January 2012). Marshall S.SmithandRebecca Winthrop, “A New Face of Education: BringingTechnology into the 2007). FarrellGlen and Shafika Isaacs, “Survey of ICT and Education in Africa” (Washington: Infodev/World Bank, 2013). Lant Pritchett, Harvard Business SchoolPress, 1997). Christensen,Clayton Superpower, eds. Madhav Needs Chavan, Classrooms?” “Who in Empirical Exploration,” David H.Autor, Frank Levy, andRichard Murnane, SkillContent of “The Recent Technological Change:An of BrilliantTechnologies Erik Brynjolfsson andAndrew Mcafee, Cory Heyman, emailcommunication withJenny Perlman Robinson, February 17, 2016. bridgeinternationalacademies.com/approach/model/. Bridge International Academies, “Model,” accessedAugust 24, 2015, http://www. Interconnecting Scholarship andPractice,” H2OProject (Cambridge: Harvard University, 1999). Nora SabelliandChrisDede, “Reconceptualizing andProcess theGoals of Educational Research Funding: Countries CanLearn About Evidence fromEvidence. Caroline Fiennes andSarah Gupta. (Washington: International Food Policy Research Institute, 2009). David J. SpielmanandRajul Pandya-Lorch, eds., education-plus-development/posts/2015/08/18-cholera-education-bangladesh-chabbott. against Cholera inBangladesh,” Brookings.edu (blog),August 18,2015, http://www.brookings.edu/blogs/ “Blueprints, Bureaucrats,Collette Chabbott, andScalingUp: Lessons for Education from BRAC’s Fight Board, “Innovation Strategy for Education andTraining: Progress Report no. 5” (Paris: OECD, 2011). OECD Directorate for Education Centre for Educational Research andInnovation (CERI) Governing education-plus-development/posts/2015/08/18-cholera-education-bangladesh-chabbott. against Cholera inBangladesh,” Brookings.edu (blog),August 18,2015, http://www.brookings.edu/blogs/ “Blueprints, Bureaucrats,Collette Chabbott, andScalingUp: Lessons for Education from BRAC’s Fight Strong Performers andSuccessful Reformers inEducation,” (Paris: OECD, 2011). Organisation for Economic Co-operation andDevelopment, “Lessons from PISAfor theUnited States: The Rebirth of Schooling Education: Ain’t Learnin Clay ChandlerandAndadilZainulbhai (New York: Simon &Schuster, 2013). The Innovator’s When New Dilemma: Technologies Cause Firms Great to Fail Quarterly Journal of Economics (New York: W. W. Norton,2014), 7. Getting Better Education II:What inLess Economically Developed The Second MachineAge: Work, Progress, andProsperity inaTime Millions Fed: Proven Successes for Agricultural Development Reimagining Unlocking India: thePotential of Asia’s Next (London: Evidence, Giving forthcoming). 118,no. 4(2003). g (Washington: Brookings Institution Press, (Boston: 147 148

Millions Learning: Scaling up quality education in developing countries Bibliography 169-79. Indonesian Preschool Children: AnExperimental Study.” Borzekowski, D. G., L. andH.K.M.Henry. Impactof Jalan “The SesamaontheEducational andHealthy Development of Development, March 2013. Experimental Evidence onExternal Validity inKenyan Education.” Working Paper 321.Washington: Center for Global Bold, Tessa, Mwangi Mwabu, Kimenyi, Germano AliceNg’ang’a, andJustin Sandefur. “ScalingUp WhatWorkings: 2010. Developing Countries: TheRole of Education.” Policy Research Working Paper Washington: 5342. World Bank, Blankespoor, Brian Susmita Dasgupta, BenoitLaplante, andDavid Wheeler. “Adaptation toClimate Extremes in Development: AReal World Guideto Its TheoryandPractice Binswanger-Mkhize, Hans P., Jacomina P. deRegt, andStephenSpector, eds. 2003.Bank, Underpinnings andProgram Design Implications.” Policy Research Working Paper no. 3039. Washington: World Binswanger, Hans P., andSwaminathan Aiyar. “ScalingUp Community-Driven Development: Theoretical Sapiens Bille, R.“Action without Change? On theUse andUsefulness of Pilot Experiments inEnvironmental Management.” 5, 2014. Bernard, Jean-Marc. “Data Are Not Just toPleasePartnership Statisticians.” Global for Education (blog),August June 2012. World DominicBarton, etal.“The atWork: Jobs, Pay andSkillsfor 3.5BillionPeople.” McKinsey Institute, Global Barber, Michael, eds. andSaadRizvi, Enhancing and Hindering Factors.” Banda, B.,V. Tindi, Mudenda, E. andK.Nakai. “Lesson StudyPractice of ScienceTeachers inZambia: Its Effects, Exploration.” Autor, David H.,Frank Levy, andRichard Murnane. SkillContent of “The Recent Technological Change:AnEmpirical Discussion document prepared for theLearning Metrics Task Force 2.0. December 1,2015. “Assessment for Learning (A4L):AnInternational Platform National toSupport Learning Assessment Systems.” Sampling/History/p/54.html. ASER Centre. “Overview,” New Delhi: ASER Centre, 2016. http://www.asercentre.org/Survey/Basic/Pack/ asercentre.org/p/76.html. ASER Centre. “Citizen-led, Household-based Assessments around theWorld.” New Delhi: ASERCentre, www. ASER Centre. “Annual Statusof Education Report Data Query”. New Delhi: ASERCentre, 2014. Canberra: Australian Agency for International Development, February 2012. Anderson, Ian. “ScalingUp Development Results: ALiterature Review andImplications for Australia’s AidProgram.” 2013). American Educational Research Association. “Lead theChangeSeries:Q&AwithVicky no. (August Colbert,” 29 Ahmed, S.,andM.French. “ScalingUp: TheBRAC Experience.” the World Development Report.Washington: World June Bank, 10, 2013. Aedo, Cristian etal.“From Occupations toEmbedded Skills:ACross-Country Comparison.”Background paperfor 10, andDevelopment. CommissiononGrowth Washington: World 2008. Bank, Acemoglu, Daron, and James Robinson. Role “The of Institutions and Development.” in Growth Working Paper no. 3, no. 1(2010). Quarterly Journal of Economics New Perspectives inScience Education Asking More:The to Path Efficacy 118,no. 4(2003). International Journal ofDevelopment Behavioral . Washington: World 2009. Bank, BRAC University Journal . Pearson, 2013. . 3rd ed.Florence, Italy: 2014. Scaling Up Local Community Driven , 2003. , 35 (2011): 35 (2011): International, March 2006. Cooley, Larry, andRichard Kohl. “ScalingUp—From Vision toLarge-Scale Change.” Washington: Management Systems International, 2012. Cooley, Larry, and Rajani R.Ved. “ScalingUp–From Vision toLarge-Scale Change.” Washington: Management Systems Results for Development Institute, September2014. Cooley, Larry, and Johannes F. Linn.“Taking Innovations toScale:Methods, ApplicationsandLessons.” Washington: Impact Evaluations.” Submitted for Ph.D., ColumbiaUniversity, 2014. Conn, Katharine.“Identifying Effective Education Interventions in sub-Saharan Africa: A Meta-analysis of Rigorous Escuela Nueva Model.” Vicky.Colbert, “Improving Educational Quality andAccess inColombiathrough Innovation andParticipation: The 32, no. 6(2003). “RethinkingCoburn, Cynthia E.. Scale: Moving Beyond Numbers toDeep andLasting Change.” Journal Simon, and MatinaClark, Stevis. “Zuckerberg-backed Seeks to Shake Startup Up African Education.” from Five Indian States,” Chudgar, Amita,andVyjayanthi Sankar. Relationship “The between Teacher andStudent Achievement: Gender Evidence Business SchoolPress, 1997. Christensen, Clayton. edited by Clay ChandlerandAndadilZainulbhai. New York: Simon &Schuster, 2013. Chavan, Madhav. Needs Classrooms?” “Who In http://www.brookings.edu/research/reports/2013/02/brooking-blum-roundtable-2012. against Poverty.” Report for Chandy, Laurence, Kemal Dervis, andSteven Rocker. “Clicks into Bricks,Technology into Transformation, andtheFight Environments.” Working Global Papers no. 41.Washington: Brookings Institution, September2011. Chandy, Laurence, andJohannes F. Linn.“Taking Development toScaleinFragile Activities andLow Capacity Solutions to Millionsof PoorPeople. Chandy, Laurence, AkioHosono, Homi Kharas, andJohannes Linn,eds. posts/2015/08/18-cholera-education-bangladesh-chabbott. in Bangladesh.”Brookings.edu (blog),August 18,2015. http://www.brookings.edu/blogs/education-plus-development/ Collette. “Blueprints, Bureaucrats,Chabbott, andScalingUp: Lessons for Education from BRAC’s Fight against Cholera Brookings Institution, 2013. Center for Universal onLearning Education. Compact “2013 Donor Global Network Mapping Results.” Washington: around theWorld Cassola, Adele, andJody Heymann. Technologies. Brynjolfsson, andAndrew Erik, Mcafee. Advance ed.Washington: World 2014. Bank, Bruns, Barbara, and Javier Luque. approach/model/. Bridge International Academies. “Model.” Accessed August 24, 2015, www.bridgeinternationalacademies.com/ about/. Bridge International Academies. “About.” Accessed August 24, 2015. www.bridgeinternationalacademies.com/company/ for Educational Planning, 2008. Bray, “Double-shift Mark. Schooling:Design andOperation for Cost-effectiveness.” Paris: UNESCO International Institute Bradach, Jeffrey,“TransformativeGrindle, andAbe Scale:The Future Growingof What Works.” Bradach, Jeffrey L., “Scaling Impact.” Review (February 19, 2014). , March 13, 2015. New York: W. W. Norton,2014. . Oxford: Oxford University Press, 2012. The Innovator’s WhenNew Dilemma: Technologies Cause Firms Great to Fail. Journal of Education for International Development Journal of andInternational Comparative Education Brookings Blum Roundtable onGlobal Poverty. Great Teachers: Great How to Raise Student Learning in Latin America and the Caribbean. Washington: Brookings Institution Press, 2013. Stanford Social Innovation Review Lessons inEducational Equality: Successful Approaches to Intractable Problems The Second Machine Age: Work, Progress, andProsperity inaTime of Brilliant Reimagining Unlocking India: the Potential of Asia’s Next Superpower, (Summer 2010). Getting to Scale: How to BringDevelopment 3, no. 3(2009). Washington: Brookings Institution, 2012. 38, no. 38, 627-42. 5(2008): Stanford Social Innovation Educational Researcher Boston: Harvard Wall Street

149 150

Millions Learning: Scaling up quality education in developing countries Corporate Culture.” Gerard J. Tellis, Jaideep C.Prabhu, and Rajesh K.Chandy. “Radical Innovation across Nations: ThePreeminence of for Effective Organizations, 2014.February http://docs.geofunders.org/?filename=geo_2014_ssir_supplement_final.pdf. GEO. Philanthropy “Smarter for Rethinking GreaterImpact: How Grantmakers Scale.” Support Washington: Grantmakers and Its Effect onChildin Mortality 175 Countries between 19702009: and A Systematic Analysis.” Emmanuela,Gakidou, Krycia Cowling, Rafael Lozano, andChristopher J. Murray. L. “Increased Educational Attainment Madeleine.Gabriel, “Making It Big:Strategies for ScalingSocialInnovations.” London: Nesta, July 2014. International Labour Organization, 2013. Comparative Analysis Experiencesfrom Brazil, Chile, India, Mexico andSouthAfrica.”Working Paper no. 4.Geneva: Fultz, Elaine, andJohn Francis. “Cash Transfer Programmes, Poverty Reduction andEmpowerment of Women: A Educational Patents.” OECD Education Working Paper no. 84.Paris: OECDPublishing, 2012. Foray, Dominique, andJulio Raffo. “Business-driven Innovation: Is It Making a Difference in Education?of AnAnalysis Fixsen, D. etal.“Implementation Research: ASynthesis of theLiterature.” Tampa: NIRN,2005. Farrell, andShafika Glen, Isaacs. “Survey of ICT and Education inAfrica.” Washington: Infodev/World2007. Bank, Management Systems International, andWorld ExpandNet, Health Organization, 2007. Fajans, Peter, Laura Richard Ghiron, Kohl, andRuth Simmons.Questions “20 for Developing aScalingUp CaseStudy.” Programmatic Research for Successful ScalingUp.” Geneva:World Health Organization, 2001. ExpandNet and World Health Organization. “BeginningwiththeEnd in Mind: Planning Pilot Projects andOther Education For Monitoring Report.“Teaching AllGlobal andLearning: Achieving Quality for All.”Paris: UNESCO, 2014. Policy Paper 18.Paris: UNESCO, July 2015. Education For Monitoring Report.“Pricing AllGlobal theRight toEducation: TheCost of Reaching New Targets by 2030.” The Education Commission.“The Commission.”Accessed onMarch 4,2016. http://educationcommission.org/. European Centre for Development Policy Management andOverseas Development Institute, 2015. ECDPM andODI.“Case StudySummary: Material Wellbeing, Education andEmployment.” Maastricht andLondon: Led Assessments inShiftingtheEducation Agenda,” Washington: Results for Development Institute, June 2015. MollyEberhardt, Jamieson, Michelle Engmann, Tara Hill, andDaniel “BringingLearning Plaut, toLight: TheRole of Citizen- (CELL) Papers C.J.Dunst, etal.“ScalingUp Early ChildhoodIntervention Literacy Learning Practices.” Trust, 2012. and Conflict-affected States.” Policy Report. Washington:Center for Universal Educationat Brookings; CfBT Education Dolan, Janice, Susy Ndaruhutse, AprilGolden, andRebecca Winthrop. “Building Effective Teacher Salary Systems in Fragile Development of Education inAfrica, 2001). Educational Reform Association for theDevelopment of Education for Africa”(Abidjan,Côted’Ivoire: Association for the Dembélé, Martial,Joel Samoff,E. Molapi Sebatane. and“Scaling Up by Focusing Down: Creating SpacetoExpand In Africa.”EdQual Working Paper 28.EdQual: September2011. Dembélé, Martial, Joel Samoff,E. Molapi Sebatane. and “‘Goingto Scale’: the Nurturing Local Roots of Education Innovation Educational Improvement. Dede, Chris,James P. Honan, andLaurence C.Peters, etal. Davies, Anna.“Spreading SocialInnovations: 2014. ACaseStudyReport.”TEPSIE, Davidson, Adam. “SavingtheWorld, Startup-style.” Daniels, Alex. “Ford ShiftsGrantmaking toFocus Entirely onInequality.” C and R.Truglio. Mahwah, NJ: Erlbaum, 2001. Cooney, J. G. Foreword. In Development, 1973. Coombs, P., withR.Prosser andM.Ahmed.“New Paths toLearning.” New York: International Councilfor Educational Institution, Washington, June 2014. Cooley, Larry. “Millions Learning: Role of Partnerships inBringingEducation Initiatives toScale.” Meeting attheBrookings 1,no. 2(2006): 1-10. Journal of Marketing SanFrancisco: Jossey-Bass, 2005. “G” Is for Growing: Thirty Years of Research onChildrenand“Sesame Street,” 73 (2009). New York Times, Scaling Up Success: Lessons Learned fromTechnology-Based November 17, 2015. hronicle of Philanthrophy, Center for Early Literacy Learning June 11,2015. Lancet editedby S.Fisch 376 (2010). February 13, 2016. Kharas, Homi, andRebecca Winthrop. “Want toSave thePlanet? Invest inGirls’ Education.” Social Impact.” Khan, Zia, Isabel Guerrero, Harish Hande, andJeffreyL. Bradach. “Creating a Funding Environment for Scaling Up Poverty.”Solutions totheChallengesof Global Cambridge:Monitor Group, 2009. Karamchandani, Ashish, Michael Kubzansky, andPaul Frandano. “Emerging Markets, Emerging Models: Market-based 2015. A Cluster Randomized Controlled Trial of SocialandFinancial Education inRwanda.” Budapest: Aflatoun International, Kaneza, Yves, Paul Moclair, andDaniel Shephard. “Financial Education for Children: WhatCurriculum? WhichMethods? Learning Ahead).”Internal document shared withJenny Perlman Robinson andEileen McGivney, February 9, 2016. Jeevan, Do Sharath. We “What Know aboutTeacher Motivation? .Our Learning at STIR SoFar (with Much More brac.net/images/what%20does%20it%20take%20to%20scale%20social%20impact.pdf. Ishtiaque Does Hussain It etal.“What Take toHave SocialImpactatScale?” BRAC, October 2014. http://innovation. isced97-en.pdf. International Standard Classification of Education. Paris: UNESCO, 1997. http://www.uis.unesco.org/Library/Documents/ Academy for Educational Development, July 1992. Huq, Rahman Rahman. “Cost Comparisonof BRAC NFPandGovernment Primary Schools.”Dhaka, Bangladesh: Hill, Robert. “Teach First: Ten Years London: of Impact.” Teach First, 2012. Hess, Frederick M.S Herbling, David. “Ministry Stops Expansionof Low-Cost Schools.” 2006. Heckman, James J. Economics “The of Investing Early in Children.” Policy Briefing no.Geary Institute, 1. Dublin: UCD Heckman, James J. “Skills,Schools,Synapses.” Discussion Paper no. 3515. May 2008. Bonn:IZA, Hattie, John. Works “What Best inEducation: ThePolitics of Collaborative Expertise.” London: Pearson, June 2015. Literature andPractice.” Washington: Brookings Institution, 2008. Hartmann, Arntraud, andJohannes F. Linn.“ScalingUp: AFramework andLessons for Development Effectiveness from Harnish, Verne. Cambridge: MITPress, 2016. Hanushek, Eric,andLudger Woessmann. American Economic Review andStevenHanushek, Eric A., G. Rivkin. “Generalizations aboutUsing Value-Added Measures of Teacher Quality.” J.Hancock, “ScalingUp Practices theImpact of inRural Good Development.” Washington: World 2003. Bank, development/posts/2014/04/23-colombia-southeast-asia-learning-gustaffson-wright. Colombia toSoutheast Asia.” Brookings.edu (blog),April23, 2014. http://www.brookings.edu/blogs/education-plus- Gustafsson-Wright, Emily, andEileen McGivney. “Fundación Escuela Nueva: ChangingtheWay Children Learn from Gugelev, Alice, andAndrew Stern.“What’s Your Endgame?” Practice. Research Triangle Park, NC:Research Triangle Institute, 2011. andP.Gove, A., Cvelich. Business School,July 2008. PaulGompers, etal.“Performance A. Persistence inEntrepreneurship.” Working Paper 09-028. Cambridge:Harvard www.gainhealth.org/sites/www.gainhealth.org/files/NaandiLR.pdf. Alliancefor ImprovedGlobal Nutrition. “Lessons from Businesses: Naandi Foundation’s SchoolFeeding Program,” http:// GIZ. “ScalingUp inDevelopment Cooperation: Practical Eschborn, Guidelines.” Germany: GIZ,2010. Nutrition PurnimaMenon, Andrew Kennedy. Stuart, Gillespie, “ScalingUp ImpactonNutrition: WhatWill It Take?” GIF. “Transforming Innovation Global Ideas Fund. into Impact.” http://www.globalinnovation.fund. 6, no. 4(2015): 440-51. Stanford Social Innovation Review Scaling Up: How Few Companies Make It andWhy theRest Don’t. ame ThingandOver Over Early Reading: Igniting Education for All. 100, no. 2(May 2010): 267-71. The Knowledge Capital of Nations:Education andtheEconomics of Growth . Cambridge:Harvard University Press, 2010. (Spring 2016). Stanford Social Innovation Review Business Africa Daily Revised ed.Early GradeLearning Community of Ashburn: Gazelles, 2014. , September24, 2015. Milwaukee Journal Sentinel (Winter 2015). Advances in , . 151 152

Millions Learning: Scaling up quality education in developing countries “Innovation Strategy for Education andTraining: Progress Report no. 5.” Paris: OECD, 2011. OECD Directorate for Education Centre for Educational Research andInnovation (CERI) GoverningBoard. ODI. “Development Progress Series.”London: Overseas Development Institute, 2011-2012. int.co.uk/wp-content/uploads/2015/10/Insights-for-taking-RBF-to-scale_Policy-Brief_NU-Health-2015.pdf. NU Health. “Insights for Taking Results BasedFinancing toScale.” UKAID, August 2015. http://resources.healthpartners- London: Overseas Development Institute, September2015. Nicolai, Susan, ChrisHoy, Tom Berliner, andThomasAedy. “Projecting Progress: Reaching by theSDGs 2030.” Accessed January 15, 2016. http://www.nsf.gov/statistics/randdef/fedgov.cfm. National ScienceFoundation. “Definitions of Research and Development: AnAnnotated Compilation of OfficialSources.” learn/. Naandi. “Our Work.” Accessed March 2,2016. http://www.naandi.org/programmes-phased-out/hungry-children-cannot- www.globalgiving.org/pfil/6853/projdoc.pdf. Naandi. “Naandi Foundation’s Centralized Kitchens for Midday Meal Programme.” Innovations Global Fund.com. https:// Vanderbilt University Press, 2012. Murphy-Graham, Erin. How It CanBeAccelerated.” Working Paper. Oxford: Skoll Centre for SocialEntrepreneurship, 2007. Mulgan, Geoff, with Simon Tucker, Rushanara Ali,andBenSanders. “Social Innovation: What It Is, Why It Matters, and Harper Publishing, 1999. Geoffrey. Moore, Miller, D., andP. Uvin. “ScalingUp: Thinkingthrough theIssues.” Providence: Brown University, 1994. Education_Intro_ Standalone_Nov%2026.pdf. Better,” McKinsey andCompany, November 2010, http://mckinseyonsociety.com/downloads/reports/Education/ Michael Barber, Chinezi Chijioke, andMona Mourshed, “How theWorld’s Most Improved SchoolSystems Keep Getting Meyer, Robert G. “GoingtoScale.” New York: UNICEF, 1984. Brookings Institution, 2015. McGivney, Eileen, and Rebecca Winthrop. Wait “Why 100 Years? Education.” in Global Washington: Bridging the Gap Experiments,” McEwan, Patrick J. “Improving Learning inPrimary Schoolsof Developing Countries: AMeta-analysis of Randomized of Development Effectiveness McEwan, Patrick J. “Cost-effectiveness Analysis of Education and Health Interventions in Developing Countries.” Effective Organizations,2014. http://www.Geofunders.Org/Storage/Docs/Geo_2014_Field_Study.pdf. McCray, J. “IsSmarter? ANational Grantmaking Getting Studyof Philanthropic Practice 2014.” Grantmakerse for in India.” McKinsey andCompany, October 2013. Mangaleswaran, Ramesh, andRamya Venkataraman. “Designing Philanthropy totheBiggest for Gaps Impact: Giving “GloriousFailure:Macmillan, Scott. TheJoy of Learning from Your Mistakes.” Linn, Johannes. “Internal ConceptNote for Project onScalingUp.” October21,2013. friday-note-zero-year. Levine, Ruth. “Friday Note: TheZero Year.” Hewlett Foundation (blog),July 10, 2015. http://www.hewlett.org/blog/posts/ Washington: Center Development, for 2004. Global Levine, Ruth, andtheWhatWorks Working GroupwithMolly Kinder. UNESCO Institute for Statistics andCenter for Universal Education atBrookings, 2013. Learning Metrics Task Force. “Toward Universal Learning: WhatEvery ChildShouldLearn.” Montreal andWashington: 2013. Montreal andWashington: UNESCOInstitute for Statistics andCenter for Universal Education atBrookings, September Learning Metrics Task Force. “Toward Universal Learning: Recommendations from theLearning Metrics Task Force.” Washington: World 1998. Bank, Community Program Support in Rural Balochistan.” Working Paper on Impact Evaluation of Education Reforms no. 10. Kim, Jooseop, Harold Alderman,andPeter Orazem. “Can Cultural Barriers BeOvercome inGirls’ Schooling? The Review of Educational Research Crossing the Chasm: Marketing andSellingCrossing High-Tech theChasm: Products to Mainstream Customers. Opening Minds, ImprovingOpening Minds, Lives: Education andWomen’s Empowerment inHonduras. 4,no. 2(June 2012): 189-213. 85, no. 3(September 2015). Millions Saved: Proven Successes inGlobal Health. The Guardia n, March 30, 2015. Nashville: New York: Journal Journal STIR Education. “STIR Education: Empowering Teachers toBeChangemakers.” 2015. http://www.stireducation.org/. Universal Education atBrookings, July 2015. Steer, andKatie Liesbet, Smith.“Financing Education: Washington: Action.” Opportunities for Global Center for Carey andAndrew P Kelly. Cambridge:Harvard Education Press, 2013. Stretching theHigher Education Dollar: How Innovation CanImprove Access, Equity, andAffordabilit Staton, Michael. “Unbundling Higher Education: Taking theComponents of Apart theCollegeExperience.” In International Food Policy Research Institute, 2009. Spielman, David J., andRajul Pandya-Lorch, eds. Developing World.” Washington: Brookings Institution, January 2012. Smith, Marshall S.,andRebecca Winthrop. “A New Face of Education: BringingTechnology into the Classroom inthe Laura Geneva:World Ghiron. Health Organization, 2006. fromPilotHealth Service Delivery Innovations to Policies andProgrammes Simmons, Ruth, andJeremy Shiffman.“Scaling Up Health Service Innovations: A Framework for Action.” Sesame Workshop. “Early Education.” http://www.sesameworkshop.org/what-we-do/our-initiatives/early-education/. Stanford Social Innovation Review Sahni, Nikhil R., Maxwell Wessel, M.Christensen. andClayton “Unleashing Breakthrough Innovation inGovernment.” Interconnecting Scholarship andPractice.” Cambridge:Harvard H2OProject. University, 1999. Sabelli, Nora, andChrisDede. “Reconceptualizing andProcess theGoals of Educational Research Funding: opinionator.blogs.nytimes.com/2011/09/29/making-change-happen-on-a-deadline/?_r=0. Rosenberg, Tina. “Making ChangeHappen, onaDeadline.” roomtoread.org/AnnualReport/2014/http://www.roomtoread.org/AnnualReport/2014/. Room toRead. “SolvingthePuzzle: Children’s Literacy andGirls’ Education.” Annual Report,2014 ed.http://www. Fedgazette andRobRolnick, “Early Art, Grunewald. ChildhoodDevelopment: Economic Development withaHigh PublicReturn.” Rogers, Everett M. edu/blogs/education-plus-development/posts/2015/01/07-education-reforms-claudia-costin-perlman-robinson. Costin, SeniorDirector for Education attheWorld Brookings Institution Bank.” (blog),January 7, 2010. www.brookings. Robinson, Jenny Perlman. “Ensuring Lasting Education Reforms by Delivering Results: ADiscussion withClaudia JuneInstitution, 2011. Robinson, Jenny Perlman. “A onLearning.” Compact Global University of London, 2010. and Teacher Matter? SomeEvidence from India.” DoQSS Working Paper 10-18. London: Institute of Education, Rawal, Kingdon. “Akin Shenila,andGeeta toMy Teacher: Does Caste, Religious, Distance orGender between Student Lant.Pritchett, Approaches.” OEDWorking Papers no. 52233. Washington: World 2005. Bank, for anEvaluation of theEffectiveness of World Bank Support for Community-Based and-Driven Development Pozzoni, Barbara, andNalini Kumar. “A Review of theLiterature onParticipatory Approaches toLocal Development scaling-up-aid-linn/10_scaling_up_aid_linn_bibliography.pdf. Development atBrookings, August http://www.brookings.edu/~/media/Research/Files/Papers/2008/10/ 2008. Pidufala, Oksana. “Scaling Up andAidEffectiveness: AnnotatedBibliography.” Washington: Wolfensohn Center for PESRP. “SchoolCouncils.”http://www.pesrp.edu.pk/pages/School-Councils. Brussels: Centre for European Policy Studies,November 2014. Pelkmans, Jacques, and Andrea Renda. “Does EU Regulation Hinder or Stimulate Innovation?” The Overhead Myth. “Moving toward anOverhead Solution.”http://overheadmyth.com/. of Development Cooperation.” Stockholm:Sida, 2001. S.Shivakumar,Ostrom, C.Gibson, E., andK.Andersson. “Aid, Incentives, andSustainability: AnInstitutional Analysis Organisation for Economic Co-operation andDevelopment, 2011. OECD. “Lessons from PISAfor theUnited States:Strong Performers andSuccessful Reformers inEducation.” Paris: , March 2003. The Rebirth Schooling of Education: Ain’t Learning. Diffusion of Innovations (Summer 2013). . 5thed.New York: Free Press, 2003. Millions Fed: Proven Successes for Agricultural Development. Washington: Center for Universal Education Brookings at Washington: Brookings Institution Press, 2013. New York Times , editedby Ruth Simmons, Peter Fajans, and (blog),September29, 2011. http:// CEPS Special Report. y, editedby Kevin Washington: In Scaling Up 153 154

Millions Learning: Scaling up quality education in developing countries International Cooperation Agency, 2015. Lesson Study—Implementation 5thed.Lusaka, Guidelines,” Zambia: Ministry ofEducation General andJapan Zambia Ministry Education. of General “School-BasedContinuing Professional Development (SBCPD) Through com/watch?v=gjATVO3l-s0. YouTube. “Punjab, Pakistan, Education Reform Roadmap.” Published on January 22, 2014. https://www.youtube. in thePhilippines.” IFPRIDiscussion Paper. Washington: International Food Policy Research Institute, 2011. Yamauchi, Futoshi, andYanyan Liu. “GirlsTake Over: Long-term Impactsof anEarly StageEducation Intervention World Health Organization “Nine andExpandNet. Steps for Developing aScaling-Up Strategy.” Geneva:WHO, 2010. World “World Bank. Development Report 2016: Digital Dividends.” Washington: World 2016. Bank, World “World Bank, Development Report Making 2004: ServicesWork for Poor People.” Washington: World 2013. Bank, catalog/world-development-indicators. World “World Bank. Development Indicators: Brazil.” Washington: World http://data.worldbank.org/data- Bank. Washington: World 2016. Bank, World “Leveraging Bank. Urbanization inSouthAsia: Managing SpatialTransformation for Prosperity andLivability.” World August Bank, 2015. World “EducationPractice Bank. Snapshot: Global TheRise of Results-based Financing inEducation.” Washington: income. World “Country Bank. andLending Groups.” http://data.worldbank.org/about/country-and-lending-groups#Low_ Washington: Brookings Institution Press, 2015. Winthrop, Rebecca, Sperling. and Gene Pakistan.” Working Paper no. 2.Washington: Center for Universal Education atBrookings, 2010. Winthrop, Rebecca, andCorrinneGraff. “Beyond Madrassas: Assessing theLinksbetween Education and Militancy in Wasserman, Founder’s Noam. “The Dilemma.” February 18,2016. http://visible-learning.org/hattie-ranking-influences-effect-sizes-learning-achievement/. Visible Learning blog.“Hattie Ranking: Influences and Effect Sizes Relatedto Student Achievement.” Accessed Vavrus, Frances. Uwezo. “Are Our Children Learning? Literacy andNumeracy Across East Africa2013.” Nairobi: Hivos/Twaweza, 2013. Uvin, P. “Fighting Hunger attheGrassroots: Paths toScaling Up.” UNICEF. “Early ChildhoodDevelopment: New AStatistical York: Snapshot.” UNICEF, 2014. www.education-inequalities.org/. UNESCO. World Inequality Database onEducation. Indicator: Out-of-School Children. Paris: UNESCO, 2016. http:// Teachers.” UISFact SheetNo. 33. Montreal: UIS,2015. UNESCO Institute of Statistics. “Sustainable Development for Education Goal CannotAdvance withoutMore Fact Sheetno. 27. Montreal: UIS2013. UNESCO Institute of Statistics. “A Teacher for Every Teacher Child: Projecting Global Needs from 2015 to 2030.” UIS org/Education/Documents/OOSCI%20Reports/brasil-oosci-summary-2012-en.pdf.UNESCO. UNESCO. “Global Initiative onOut-of-School Children: Brazil.” Paris: UNESCO, August 2012. http://www.uis.unesco. UNESCO. “Education Transforms Lives.” Paris: UNESCO, 2013. kmonadollaraday.wordpress.com/2013/12/18/the-next-innovation-scaling-up/. Thorpe, Next Ian. “The Innovation: ScalingUp.” kmonadollaraday.wordpress.com (blog),December 18,2013. https:// Teach First. “All OfUs: Teach First Annual Review.” London: Teach First, 2014. Business, 2014. Sutton, Robert I.,andHuggy Rao. Desire andDecline: Schooling Amid Crisis inTanzania Scaling Up Excellence: Getting to MorewithoutSettling for Less. What WorksWhat in Girls’ Evidence Education: for the World’s Best Investment. Harvard Business Review World Development . New York: Peter Lang, 2003. (February 2008). 23, no. 927-39. 6(1995): New York: Crown process Millions Learning Annex 1. and broader development efforts. health, agriculture, social innovation, education andothersectors, including a wide-ranging literature review in and analysis. We alsoconducted for investigationthen selected further countries andsubnational regions were where progress hadbeenmade. Specific at theirlearningoutcomesandidentify assessment datatolookacross countries of international andregional learning the study. Thisincludedananalysis uncover interventions toconsiderin backgroundconducted research to studies, theMillions Learning team In additiontoissuing thecallfor case improved. known examples of where learninghad cases butalsofrom local,lesswell- draw evidence notonlyfrom well-known call for casestudies, we endeavored to learning takes place. In issuing aglobal nonformal settings where intentional or policiesthattake placeinformal or interventions, whichwe definedprograms learning. We focused oneducation up effective approaches toimproving secondary educationthathave scaled early childhood development through examples of programs andpoliciesfrom case studies inJune 2014 toidentify study, CUEissued aglobalcallfor To contribute tothe Millions Learning case study selection in alow- or middle-incomecountry. learning atscale(to varying degrees) of demonstrating an improvement in foremost becausethey metourcriteria These caseswere first selected and review. 14casesforwe selected amore in-depth for the approximately 12 cases to review in-depth criteria with the aim of selecting andevaluativebased onthequalifying evaluating theshort-listed casestudies Advisory Panel thenassisted usin Our Millions Learning International needed. address gaps orlingeringquestions as we requested additionalinformation to inclusion inthefinalstudy. In doingso, below) andwere strong candidatesfor (see “Case Criteria” StudySelection thatmetourcasestudy criteria30 studies, we identified approximately 100 casestudies intotal.Ofthesecase research, we considered more than submissions, andwithourindependent case studies, we received 75 casestudy Ultimately, asaresult of the global call for that warranted additionalstudy. informants cases alsohelpedtoidentify meetings, andinterviews withkey Recommendations from consultations, Millions Learning study. Ultimately, 155 156

Millions Learning: Scaling up quality education in developing countries well-being, andsciencetechnology. and mathematics,social andemotional literacy andcommunications, numeracy Force, including physical well-being, as defined by the Learning Metrics Task progress across multiple competencies, broadly definedandcouldinclude rates. Animprovement inlearningwas rates andincreases inprogression indicators such as decreases in dropout indication of learning,along withproxy learning, or in a few instances, a strong clear andmeasurable improvement in There shouldhave beenevidence of a Impact: for in-depthreview:studies case following the14 criteriatohelp identify The Millions Learning team developed the Case study selection criteria cases was toshedsomelight onthe scaling. Rather, thepurpose of the to provide any recipe orblueprint to approaches. They alsowere notselected ongoing debatesregarding someof the to improving there are learning;infact, represent thebest modelsorapproaches These casestudies were notmeant to in geography andpopulations reached. secondary programs, aswell asdiversity of interventions from early childhood to a range of scalingpathways andtypes balanced against the desire to identify that hadyet tobeexplored. This was all of expanding, anewangleoraspect course correctionmadeintheprocess contentious issue addressed, somebold was an interesting story to tell—some Beyond we believed that, thatthere future efforts around scalingeffective And how couldallof theselessonsinform more children learningvaluable skills? events, orprocesses ledtomillionsof order toexpand? Whatkey decisions, confronted thathadtobeaddressed in not inanother? Whatobstacles were made themeffective in onelocationbut implementation. For example, what importantly—the stories behindtheir but additionally—and perhaps more that demonstrated learninggains the technicaldesignof theinterventions They were meant touncover notonly most marginalized anddifficulttoreach. amongthechildrenparticularly whoare success inimproving learningoutcomes, nuances anddrivers behindlarge-scale subnational-level commitment. included ifthey represented anational-or population orgeographic area were reached. Interventions targeting aspecific total target populationmust have been achieving scale, asignificant share of the to multiple country contexts. To as qualify national, orsubnational scale, oradapted of beingimplemented) onaregional, implemented (or beenin the process The policy or program should have been Scale: studies withstronger evidence. evaluations. Preference was given tocase assessments, household surveys, or as international, regional, ornational the assessment methodused,such Nominations shouldhave clearlydefined of achieving theSDGs? learning approaches, especiallyinlight implemented. was current when the intervention was or anequivalent classification that classifies economiesbaseduponGNI, as defined by the which World Bank, middle-, orhigh-middle-incomecountry operational inatleast onelow-, low- The intervention must have been Low- ormiddle-income country: cost-efficiency. partnerships, data,andtechnology; and thematic interest inthestudy—financing, time; focus onone of four areas of communities; progress sustained over afocus onmarginalized included: process income, othercriteriaintheselection scale, andcountryIn additiontoimpact, Additional factors for consideration: 157 158

Millions Learning: Scaling up quality education in developing countries 14 Annex 2. Bridge Bridge Education Primary Non-Formal BRAC, Media Center Government’s State Amazonas International Aflatoun NAME Zambia Lesson Study, Jordan I Nueva Escuela Fundación Educate! Academies International NJAZ, Millions Learning provinces) Zambia (all 10 governorates) Jordan (all 12 Colombia starting in 16 countries, Rwanda, Uganda Uganda Kenya, Nigeria, South Sudan Philippines, Pakistan, Bangladesh, Afghanistan, Brazil Amazonas State, starting inIndia 116 countries, LOCATION

Secondary Secondary Primary, Secondary secondary Primary, Lower Secondary Primary Pre-Primary, Primary Secondary Secondary Primary, Early childhood, LEVEL EDUCATION case studies Teacher training development Entrepreneurial Work readiness, Financial literacy, program secondary school primary andlower Alternative Teacher training development, Entrepreneurial education, Secondary private school) Primary (low-cost development, Early childhood school program Alternative primary learning program school distance Formal secondary financial education Social and development, Entrepreneurial FOCUS AREA 2,200 tutoring tutoring 2,200 300,000 students, each year and peereducators teachers, facilitators, students &34,000 Almost 4million NUMBERS REACHED date 46,000 teachers to 1.8 millionstudents & volunteers todate & over 23,000 millionstudents 1.2 Over 5millionstudents secondary schools or 12% of Ugandan schools, partner 350 120,000 students in teachers todate students &8,000 Over 100,000 countries eachyear in otherfour NFPE 900,000 students Bangladesh &over 670,000 students in teachers todate teachers, &60lecturing Worldreader Teach For All System (SAT) Learning Tutorial Tutorial / Aprendizaje Sistema de Street Sesame Workshop, Sesame Program Literacy to Read, Room Read India Pratham, NAME starting inGhana 69 countries, United Kingdom United Statesand starting inthe 39 countries, Nicaragua Honduras, Ecuador, Brazil, Colombia, States in theUnited countries, starting More than150 Zambia Tanzania, Vietnam, Africa, SriLanka, Laos, Nepal, South Cambodia, India, Bangladesh, states) India (across 23 LOCATION Secondary Primary, Secondary Primary, Early childhood, Secondary Early childhood (Grades 1-2) Primary (Grades 3-5) Primary LEVEL EDUCATION

Literacy teaching pathways to Alternative school program secondary Alternative mathematics reading & Early grade development, Cognitive development, Early childhood Teacher training reading, Early grade education Remedial FOCUS AREA month reading digitalbooksper date &1.1 millionpeople Over 5.6 millionpeopleto alumni todate teachers52,323 and 1.1 millionstudents & students todate Estimated 300,000 children Approximately 156million to Read overall) todate 10 millionstudents (Room Program specifically) and 2,000 teachers (Literacy 110,000 students & government partnerships via state ordistrict million students indirectly from 2014-2015 &over 6 directly inrural areas Over 350,000 students NUMBERS REACHED 159 160

Millions Learning: Scaling up quality education in developing countries Gardiner, Shurahbeel Laura Gharaibeh, Michael Shivani Gibbons, Ghiron, Ghosh, French, Nora Fyles, Orlando Gabarrete, Jenifer Gamble,AmandaGardiner, Sophie Heriberto Flores, John Floretta, Marcelo Campbell Fonseca, Nadine Freeman, Anna Filipe from Tres Unidos community andclassmatesinAmazonas, ArielFiszbein, Eversmann, ChineloEzeobi, Mutaz Faza, Mary Fearnley-Sander, Caroline Fiennes, mentors, Nada Elattar, Pius Elumeze, Eliza Erikson, KhaledEstitieh, ArielEvans, Eric Durham, RalfGeorge Dürrwang, Molly Jamieson John Eberhardt, Edeh, the Educate! Balmurugan Devraj, Ngoc Dinh, Carolina Dominguez, SoheilDooki, Meghan Duffy, Malek Abu Deeb, Michael Deich, Natasha de Marcken, Joseph DeStefano, Pingla Devi, Loren Crary, Luis Crouch, Raoul Davion, Dawodu, Susan Davis, Chris Dede, Gbolahan Me Chung, Elia delCid, Vicky CharlotteCole, Peter Colbert, Colenso, BitaCorrea, and theteachers attheNgungu Primary SchoolinZambia, PunamChuhan-Pole, An- Musounda Chileshe, CharlesAnthony Chisanga,Esvah Chizambe, Malonga Christine SharanyaColette Chabbott, Chandrana, Laurence Chandy, Maria Cherono, Fridah Byrd, Ron Nick Canning, Esther Cammaert, Care, Tim Carlberg, Andrew Carruthers, Brown, Barbara Bruns,Ellen Buchwalter, Trang Bui, BorisBulayev, Canh Buon, Andrew Maria Brindlmayer, teachers attheBroadway SecondarySchoolinZambia, Katherine Unmesh Brahme, theteachers attheBridgeInternational Nairobi, Academy inGicagi, Bosch, Jo Bourne, Brackin, Geordie Jeffrey Bradach, Lucy Bradlow, AmandaBraga, Elizabeth Binder, SoledadBoas, Funke Amy Black, Bolaji,Andrew Bollington,Talya Joyce Patricia Bheeka, Manisha Bhinge, Deema Bibi, Nguyet Bich, Jeroo Billimoria, Berman, Desmond Bermingham,James Bernard, Jean-Marc Bernard, Averil Besier, Middle Schoolof Patna, Reyes Regis Benedict, Benedith,Teodora Berkova, Sarabeth Banu, Omar AliBarishi,Jorge Baxter, Christine Beggs, parents of BegumpurGirls’ Balarabe, BensonBanda,Sashwati Banerjee, Sourav Banerjee, Rukmini Banerji,Nikhat Tamar Manuelyan Atinc,Mario Ayala, AbeerAzem, Simon Bailey, Reena Bajaj,Mallam Appleton, Clarita Arboleda,James BeatriceAsamoah, Arinaiwte, Hassan Bani‘Ata, Amorim,Kateof educationGedeão Anderson, Manos Antoninis, Maggie Shehan team atAmazonas StateSecretariat of Education, Nezar Ammari,former secretary Aleng, Marcelo Peréz Alfaro, KhaledAlmasri,Omar al-Momani, Omar al-Nimri, the Amrita Ahuja, parents of Akhetwara School inBihar, Rasha Al-Akkad, BlessedGrace Philip Adigun, Sharad Aggarwal, Omar Agha, Mustapha B. Ahmad, Maimuna Ahmed, Ibrahim Jnyas Abubakar, Karabi Acharya, Samuel Yalew Adela, Adeyemi Adeyemo, its various stages, including: their knowledge andfeedback ontheindividualcase studies andreport throughout We are extremely grateful to awide-range of colleagues whogenerously shared Acknowledgements Annex 3. Sani Abdullahi,Tara Abrahams, AliyuAbubakar, Alhaji Sharma, Daniel Shephard, Kazuro Shibuya, BassemShlul,Jack Shonkoff, Adam Short, Ambassador Molapi Heidi Sebatane, Schubert, Babita Shankar,Schmidt, Shailendra Alodia Santos, Taigo Sasaki,Ruhollah Sayyah, Patricia Philipp Scheid,Dana Schmidt, Sabani, Urvashi Sahni,Olalekan Saidi,AnaSalinas, Justin Sandefur, Kayode Sanni, former minister of educationinNigeria Ruqqayat Rufai, Julia Ruiz, Sara Ruto, Nora Marcelino Moreno Rivera, AnaLuis Correia Rodriquez, Mark Roland, ShariRosenfeld, Nancy Refki, Fernando Reimers, Michael Rettig, Michael Richards, David Risher, Quijada, Le ThiQuy, Moad Rabbabah, Hamza Raed, Erik Ramirez-Ruiz, Renan Rapalo, Resource Centre Coordinators inBihar, Xavier Prats-Monne, Lant Gerardo Pritchett, Penélope Pérez, Devyani Pershad, AlinePognonec, Elana Pollak, Pratham Cluster Ramnaresh Patel, Harry Patrinos, Evgenia Peeva, Sara Peracca, Juan Miguel Pérez, Daniel Ortega,Akira Oshima,Juan Pablo Ospino, Giraldo Adam Parker, ShamaParvee, James Omondi, Katherine Onorato, Neil O’Reilly, Katherine Onorato, Doyin Orogun, DanielObeidat, Obus, Tayo Odekunle, ChrisOkoro, Oluwafunmi Olalusi, Martine Hisanao Noda, Assem Nofal, Hugo Ñopo, Milena Novy-Marx, Ndidi Nwuneli, Huda Neto, Nguyen ThiThanhNga, Elizabeth Mwila Ngonga, Susan Nicolai, Sylvia Njoroge, Kazuyoshi Nakai, Dorothy Namubiru, Ola Natour, AhmedNayef, José Augusto deMelo Mercy Mwiya, Hawah Nabbuye, Zulfa Nafees, Rhoda Nanre Nafziger, Jordan Naidoo, Muskin, Lazarous Mutale, Sujatha Muthayya, Joan Mwachi, Paxina Mwenga Mweba, Maria Eugenia Munaretto, Lynn Murphy, Erin Murphy-Graham, AnnaC.Murru, Joshua John Mugo, teachers atMukonchi SecondarySchoolinZambia, Anthony Mullen, Kristen Molyneux, Roeland Monasch, Ives Morales, Trang Nguyen More, Caitlin Moss, McQueston, Media Center teachers, Edward Mfumblwa, Shuaibu Dabo Mohammed, McAlpin, John McArthur, Phil McComish, Patrick McEwan, Eugenia Kate McGill, Matachi, AmbassadorSteve Matenje, Mineko Matsumoto, Nadim Matta, Renee Kevin Marinacci, Alejandro Martinez, Yesenia Martinez, Marwan Mashriqy, Atsushi Venkatesh Malur, Adnan Mansour, Wendy Mantamba, Jonas Marcello, Erum Mariam, Madrine, Bronwen Magrath, Luckson Malambo, Nuzhat Joyce Malik, Malombe, Luwo, Viston Machiko, ScottMacMillan, Mallam AhmadAbdullahiMadana, Musonda Manh Loi, Joan Lombardi, Zev Lowe, teachers atLusaka SecondarySchool,Godfrey teachers and students atLight High inKampala, Johannes Linn,Cynthia Lloyd, Vu Filip Lenaerts, Laura Lewis, Emily Leys, Marie Leznicki, Helen Licata,Sandra Licon, Lake, Jeff Lansdale, Phong Le, Charles Leadbeater, June Lee, Suezan Lee, Sun Lei, Kohl, Jonas Koll, Wendy Kopp, Nasko Koprivshtitsa, Jarika Kumari, Sunita Kumari, Lucy Khadra, Deepali Khanna,Iyke Kimeminhia, Mwangi Kimenyi, Jay Kimmelman, Richard Caroline Kayoro, Esther Kazeze, LizKellison, Darrin Kerr, Kimberley Kerr, Bayan Abu Ihano Mulako Kapanda, Vishnu Karki, AliaKawasmi, SolomonKayiwa, Sarah Kayondo, Sharath Jeevan, Mahachi John, Tressa Johnson, Kasongo Joseph, Matthew Jukes, Islam, Eiko Izawa, Sarah Jaffe, Cornelia Janke, Muhannad Jarrah, Shuba Jayaram, Thu Huong, Huong Huyen, Clare Ignatowski, Miki Inaoka, Danilo Iraheta, SheikTanjeb Secondary School in Zambia, Razan Hindawi, Rachel Hinton, Ngoc Hoang, Nguyen Ross Hall, Yoshie Hama, Per Havgaard, CoryHeyman, teachers attheHilltop Wright, Tran Ha, Oliver Haas, Lauren Hadi, Melen Hagos, Nguyen ThiHong Hahn, LaurenGordon, Greubel,BethanGrillo, Randa GrobZakhary, Emily Gustafsson- Maria GonzalezdeAsis, Gisore,AmandaGlassman, Holly Gibson Gillespie, Gannon 161 162

Millions Learning: Scaling up quality education in developing countries our copy-editor, Marcia Kramer; designer, blossoming.it; andprinter, TheYGS Group. Lastly, we would like to extend aspecialthankyou to thefollowing: and AlixZwane. Ximenes, Yasmina Zaidman, Noor AbuZant, Teodora (Teddy) Zareva, Tassew Zewdie, and theteachers inKamulanga SecondarySchoolinZambia, Bruce Wolpe, Augusta Andrew White, Laura White, Brett Wigdortz, Matthew Williams, Patricia Wmudenda Resource Center Vietnam, Haruna Wakili, Joe Wales, Ian Walker, LillianWamuyu, Vázquez, Emiliana Vegas, BalaVenkatachalam, Kim Wright Violich, VUFO-NGO Leeuwen, Rebeca Vargas, Ingrid Vasquez, Secretary of Education for PueblaPatricia Rabecca M.Twelasi, Paul Ugwu, Mustapha Salisu Umar, Justin van Fred Fleet, Van Thanh, Edward Tindi, ChrisTinning, Joice Toyota, Sarah Troup, Rosemarie Truglio, Le Tan, Shinichiro Tanaka, Bessie Tembo, Miriam Temin, Pham Vu Thang, Tran van Hideharu Tachibana, Tina Tam and Worldreader and Uganda, colleagues in Ghana MorganSolanki, Michael Strecker, Staton,AnitaStewart, Hajiya Jummai Suleiman, Katie Smith,Secretary of Education Rossieli Soares daSilva, StephenSobhani,Ajit Ruth Simmons, Heather Simpson, Vikram Singh, Richard Singoyi, LisaSlifer-Mbacke,

1775 Massachusetts Ave., NW Washington, D.C. 20036 202 797 6000 www.brookings.edu/universal-education