Strategies to Scale Up Social Programs : Pathways , Partnerships and Fidelity

Strategies to Scale Up Social Programs: Pathways, Partnerships and Fidelity (Larson, Dearing, & Backer, 2017) summarizes a comparative case study of 45 social programs that scaled up. Three pathways, various partnerships, and fidelity considerations are discussed. This resource, commissioned by The Wallace Foundation, will be helpful to youth development organization leaders desiring to scale up. Youth development practitioners will benefit from the discussion on fidelity considerations in implementation. As the field of youth development continues to grow and programs are refined, leaders at youth development organizations may feel a responsibility to expand interventions to more individuals in need of services. For example, YMCA of the USA heeded the call to address the achievement gap (http://www.ymca.net/achievement-gap) back in 2011, with seven sites in an after-school program, and most recently in more than 260 sites across the nation. Expanding services must be planned methodically, with structures and resources in place to successfully reach more youth. Youth development organization leaders considering or planning to expand should read a report published in September 2017 titled Strategies to Scale Up Social Programs: Pathways, Partnerships and Fidelity (Larson, Dearing, & Backer, 2017). The report contains results of a comparative case study of 45 social programs that were all brought to scale. To identify programs to include in the study, evaluators searched for and reviewed documents from organizations and obtained recommendations from expert informants. Programs were selected if they focused on education, youth development or health and if they effectively scaled up; documents were reviewed for evidence of program scaling, Journal of Youth Development | http://jyd.pitt.edu/ | Vol. 12 Issue 4 DOI 10.5195/jyd.2017.579 Strategies to Scale Up Social Programs 108 such as number of sites and/or beneficiaries served and indication of at least one partner. Once the 45 programs were selected, 100 leaders at 44 of the social programs were interviewed to gather supplementary information to what was included in the documents. Finally, four programs were selected for additional in-person and on-site interviews so that the evaluators could gain a deeper understanding of their scaling efforts and to serve as illustrations for the pathways. The report identifies 3 components in scaling social programs: Pathways Although there are different paths to scaling, the report highlights three that were common among the 45 social programs:  Branching: A lead organization offers a program to new sites or to new target groups (Mulgan, Ali, Halkett & Sanders, 2007; Dees, Anderson & Wei-Skillern, 2004; Kalafatas, 2006)  Affiliate: An implementing organization buys the rights, often through a contract, from a lead organization to offer the social program (Gabriel, 2014; Mulgan, Ali, Halkett & Sanders, 2007)  Distribution network: A lead organization “piggybacks” on another organization’s network and structures (Gabriel, 2014) Partnerships Social programs in the study did not scale up without assistance; the following were the most common types of partners:  Lead partner: leads scaling efforts  Distribution partner: provides connection to local implementers  Implementing partner: offers direct services  Supporting partner: gives funding and/or expertise Fidelity Scaling up might require making one or two of these types of changes to the program intervention such as:  Reinvention: Changes made to the program prior to scale-up by the lead partner. The report identifies three types of partner involvement in reinvention: o Integrative reinvention: The lead partner shares control in reinvention decisions with a partner organization which leads to changes in structure or practices of the partner organization; roles may blur. Journal of Youth Development | http://jyd.pitt.edu/ | Vol. 12 Issue 4 DOI 10.5195/jyd.2017.579 Strategies to Scale Up Social Programs 109 o Interactive reinvention: The lead partner seeks program modification information and is the sole decision maker, or a partner may also have a say; organization identities remain separate. o Contractual reinvention: The lead partner manages reinvention through a contract with a partner; organization relationships are low in interdependence.  Adaptation: Changes made to the program for implementation purposes by implementers. The resource classifies four types of adaptation decisions: o Prescriptive fidelity: Adaptation is discouraged, and implementation is monitored. o Expectant fidelity: Adaptation is discouraged, and implementation is not monitored. o Responsive adaptation: Adaptation is encouraged as the lead partner responds to feedback, and implementation is monitored. o Independent adaptation: Adaptation is encouraged, and implementation is not monitored. The sections on fidelity and implementation define types of program adaptations and how they fit into the larger picture of scaling up, which may be useful to practitioners. However, given that practitioners are not the primary audience, there is an absence of in-depth information on processes implemented by practitioners at the 45 social programs to make successful adaptation. Moreover, it would have been helpful to learn how practitioners’ common concerns about scaling and implementation were addressed, how shifts in program and evaluation were successfully managed on the front lines, and what resources implementers found helpful. This type of information would also benefit leaders in organizations by providing insight on the type of support, training, and resources implementers found most effective. Since practitioners have the most contact with program participants and communities served, it would have been useful to understand what kinds of feedback strategies they used when sharing implementation and fidelity observations with leadership in order to provide context to adaptation monitoring. Given the magnitude of the study, it is understandable that topics delving into practitioner support were mainly out of scope for the project. This 90-page free report, commissioned by The Wallace Foundation, is an excellent resource for individuals in leadership roles who make decisions about scaling and forming partnerships. There is an executive summary to provide a general overview for individuals who want to quickly understand the study and have the option to dive deeper into subsequent chapters. It fully describes three types of pathways in separate chapters while tying in partnership roles and fidelity considerations. Conducting a comparative case study of 45 programs is no small feat, Journal of Youth Development | http://jyd.pitt.edu/ | Vol. 12 Issue 4 DOI 10.5195/jyd.2017.579 Strategies to Scale Up Social Programs 110 and the authors carefully integrate information from the programs to concretize how scaling dynamics look in practice. The authors include scaling research findings as well as a list of funders who support organizations’ programs and their scale-up efforts. This type of resource is needed as youth development organizations yield to the responsibility of reaching individuals in need of social services. The report is available at: http://www.wallacefoundation.org/knowledgecenter/Documents/Strategies-to-Scale-Up-Social-Programs.pdf

Scaling what works is a crucial component of systems change. The 45 education, youth development, and health programs in this study have scaled to varying degrees using different pathways, but all involve some type of collaboration or partnership and all must address issues of program fidelity. Though our study is exploratory and not representative of all social program scale up activity, we hope that the present report can help move the field of scale up studies and practice forward.
iii       To help answer this question, we collected 497 documents about 45 social programs to systematically code for strategic decisions made by lead organizations-those initiating a social program or bringing it to scale. We selected the 45 programs based on the advice of expert informants, database searching and web searches, and inclusion criteria such as program effectiveness, evidence of scale up through one of three pathways, and topical focus. We focused on the pathways they use for scaling, how they chose and then work with partner organizations, and how the program was reinvented and adapted-or not-as issues of fidelity were addressed. Because some information in which we were interested was not well represented in the available written documents, we interviewed 100 leaders and implementers associated with the 45 programs through use of a semi-structured interview protocol. Finally, we selected four programs for in-depth study through more interviews and site visits, which enriched our understanding of the work of these leadership teams and the challenges they have overcome. This purposive sampling procedure does not ensure a valid representation for extrapolating our results to all scale up activity in the given domains; rather, it is suggestive of questions and topics for further work. Ours is an exploratory study, as is much case study research.

Figures and Tables
» Implementation guidance is available. Each lead partner in this study provided guidance to implementers about how to best deliver their program. Programs had implementation guides, tutorials, and trainers available. Online resources are plentiful and include training videos and compelling first-person stories from implementers and beneficiaries.
» Adaptation is often expected and sometimes encouraged. The leaders we interviewed had personal experience with field conditions; they believed that even the best program could fail if put into place in an ill-suited environment. A number of our cases specified certain program components as prescribed and others as flexible as a way to allow for or encourage responsible adaptation.
» Monitoring for control and learning varies. Across the branching and affiliate programs are many examples of contracts, memorandums of understanding, formal agreements, data reporting requirements, dashboards, ongoing evaluations, and required training to monitor and control the implementation process. This shifts in a distribution network pathway. These programs tend to have greater collaboration in design and reinvention, so perhaps monitoring is less necessary. But this may also suggest a trade-off for where the developer may need to allow others to address issues of implementation fidelity.
The 45 cases we studied suggest that the strategic decisions about pathway, partners and fidelity are interrelated. A choice of one of three pathways-the steady establishment of branch offices in new locations, the selection and acculturation of affiliate partners in new locations, or collaboration with an established network of individuals or organizations to distribute a program more rapidly-may affect fidelity and partner choice. We also found cases where a choice of partner early on is made because of that partner organization's strength or established relationships for working through a particular pathway. And the desire or requirement for fidelity may lead to a decision about which partners make the most sense and which pathway will best facilitate greater impact.
"Partnering with universities and students is the sweet spot," said Laura Toscano, Director of The Campus Kitchens Project. "Instead of more branch offices or trying to provide all the services in-house, we find universities and students to partner with who already have the necessary resources on campus and want to do this kind of community impact work. Turning our nation's institutions of higher education into hubs for replication and ongoing innovation could be the superhighway to scaling up for any nonprofit." To become an affiliate of CKP, a student leadership team on a prospective campus is coached over a period of months. The cost associated with coaching students is lessened by using an online planning tool. Each student-led group must establish a community partnership with local organizations that provide services to food insecure residents, such as senior housing facilities, youth outreach groups, and churches. The campus-based student team also must establish and develop partnerships on campus with dining services contractors that operate large food service kitchens and dining halls, and with student groups and faculty advisors. This process can take months. When approved, they become a licensed Campus Kitchen and often receive a startup grant. So affiliation as a Campus Kitchen has to be earned.
The Campus Kitchens Project model has components of (1) food recovery, (2) meal preparation, (3) meal delivery, and (4) empowerment and education. Students apply organizational and recruitment skills across the four components. The model plays out differently in each community according to the opportunities that student leaders themselves identify and pursue in their community in response to a needs analysis. Some teams focus on augmenting the existing preparation and delivery of meals by homeless shelters, other student teams develop community gardens, while student teams at other campuses may host communal dinners or partner with senior centers or improve transportation so that access to food improves. For example, students at Auburn University raised money to purchase a cooler that would expand a local food bank's capacity to store food and thus serve additional clients. This degree of customization means that a CKP affiliate may emphasize one or more of the four model components in a service activity, and implement that component in a unique way. Still, taken together each Campus Kitchen affiliate addresses all four model components.
At the national level, The Campus Kitchens Project staff is just 12 people. The vast majority of the personnel that make this model work are volunteers based in each community. CKP has never been subject to a controlled research study, but data-collection and reporting is a required ongoing activity at each campus. "We stay on top of evaluation by requiring students to submit a monthly report through an online database system that is tailored for each campus," said Toscano. "We use this to write basic outcome reports. Twice a year we create an infographic of all of the results and send it out to all Campus Kitchen affiliates. Once a year we work with each Campus Kitchen to do a client food insecurity survey. Is the service effectively increasing the food security of those we serve? Are clients eating balanced meals? And once a year we conduct a student leadership development survey where we ask volunteers about their outcomes from participation." Monitoring is an ongoing activity for CKP and includes collecting stories and experiences from all Campus Kitchen affiliates that are then shared through an online tool and through meetings, newsletters and conferences.
Just as each college team develops local partnerships, the national office has developed partnerships with numerous funders including Sodexo Stop Hunger Foundation and AARP Foundation. The Sodexo Stop Hunger Foundation, as a primary and founding supporter, is a natural and complementary partner since the international Sodexo Corporation contracts for food service operations on many college campuses; thus, it serves not just as a source of financial support but brings the corporation's expertise and contacts with campus food service operations and sources of extra, unwanted, or unused food, as well as professional kitchen operations and advisors that can be available to student teams during off-hours. Many Sodexo employees on campuses satisfy an ongoing corporate volunteerism requirement by helping out in a Campus Kitchen.

THE IMPORTANCE OF SCALE UP
More than a half-century ago, the Ford Foundation and the Rockefeller Foundation began funding scale up projects both in the U.S. and especially internationally (Dowie, 2001;Rogers, 1994). By scaling up we mean a process for significantly increasing the number of sustained implementations of a successful program, thereby serving more people with comparable benefits. In what many observers consider to be a largely one-way "pipeline" trajectory of research-to-practice translation, scale up begins late in translation, after researchers develop a new program, and the program then debuts as a pilot, is tested to determine if it functions as intended, and then if it does, is revised and expanded and subjected to larger scale testing under varied community conditions (Sampson, et al., 2016;Lenfant, 2003). In this translational model, scale up remains successful as long as the program continues to expand and provide the intended benefits to participants.
This same standard applies to social programs that do not begin as research projects: They must continue to provide intended benefits to participants. Practice-based social programs like The Campus Kitchens Project that are created and grow in various community-based settings via affiliate organizations can mature in their degree of effectiveness through the gradual processes of making improvements as staff implement the program. These practice-based efforts also typically gather evidence of effectiveness, both to persuade funders or implementers (and partners in the scale up process) to invest in the program, and to enable its improvement over time.
For both research-based and practice-based social programs, two conditions must persist in order to conclude that scale up was successful: Increases in quantity, as in the number of sites or the number of beneficiaries served, and the maintenance of quality as in later iterations of the program continuing to provide benefit to participants as did earlier iterations of the program (Sabelli & Harris, 2015;Dede, Honan & Peters, 2005).
Over the years, public and private funders have joined in and supported the work of organizations that identify and nurture social programs and their scale up. Examples of these organizations include:  Significant gaps between the identification of effective programs and their broad adoption as standard practice have persisted in fields like healthcare, K-12 education, public health, and higher education (Green, Ottoson, Garcia & Hiatt, 2009). While many foundations, nonprofits and government agencies are involved in the scale up of social programs and consider it to be one of their most critical activities, their leaders are uncertain about which approaches to scale offer the greatest return on investment (Massarsky & Gillespie, 2013). In practice, organizations promoting scale up are likely to take advantage of opportunities as they arise, with growth a response to opportunity rather than the result of strategic decisions (Bridgespan, 2005).

KEY STRATEGIC DECISIONS FOR SCALING UP SOCIAL PROGRAMS
As an area of study, scale up is at an exciting pre-paradigmatic stage. Its core concepts, definitions, and strategies show some agreement among scholars but much is unresolved (Milat, King, Bauman & Redman, 2012). While there are many published studies and conceptual approaches in a variety of fields that might contribute to coalescence, few empirical studies have been conducted that investigate the scale up of social programs. In the meantime, policy makers faced with real-world problems are forging ahead and providing funding as well as guidance for communities and organizations to scale up effective solutions to social problems.
In this report we set out to explore and expand on questions and interests raised in recent work about alternative methods for organizing scale up (see Management Systems International, 2012; Hussein & Kerrissey, 2013;Clark, Massarsky, Raben & Worsham, 2012;Beckmann & Zeyen, 2014). We are specifically interested in elements that appear fundamental to the strategy used in scaling up social programs. Strategy refers to decisions made to achieve the objectives set out by an organization. While strategy has numerous dimensions, our focus is on three components that are interrelated and whose analysis can help us understand social program scale up: » Pathways. The organized activities by which a social program reaches more people we call a pathway. In studies of the diffusion of production technologies, the initial path selected can greatly determine the eventual scale up outcome (Greve & Seidel, 2015). Does "path dependence" characterize the spread of social programs, too? The organization, or a set of organizations, that controls the program makes decisions about which pathway to use to achieve scale (Dees, Anderson & Wei-Skillern, 2004). Pathways vary on a number of factors, and most especially in the degree of central control that a lead organizational partner has over the scale up process (Gabriel, 2014;Management Systems International, 2012;Sezgi & Mair, 2010). In this study we focus on three pathway types: (1) Branching pathways, (2) affiliate pathways and (3) distribution network pathways. We look at partnerships and program fidelity as they relate to pathways, and what differential impact pathways have on scale up.
» Partnerships. Which organizations to involve as partners and the roles that partners take in scaling up a program is a major strategic consideration. Within an overall partnership constellation, defined as all organizations working together to scale up a social program, may be several types of national or regional organizational partners: (1) Lead partners that direct scaling efforts, (2) distribution partners that provide connection to local implementers, (3) supporting partners that provide expertise and/or funding, and (4) implementing partners that provide direct services to intended beneficiaries. All these types of partners play roles in scaling up social programs, though not all have to be present for scale up to occur. In this study, we identify core scale up partners within the larger partnership constellation for each social program, identify the role they play, and recount how partners found each other since the initial social connection is a crucial step for many social entrepreneurs.
» Fidelity. Program fidelity is an important and ongoing challenge for scaling up social programs. Fidelity is the extent to which a program is implemented in the way intended by its developers, particularly with respect to critical elements whose benefits are supported by evidence (Mowbray, Holter, Teague & Bybee, 2003). In this study, we identified two sources of influence on program fidelity. The first is the extent to which developers of social programs deliberately reinvent their program with core partners to increase its chances of achieving scale. The second is program adaptation where developers have less control since adaptations to the program are made by implementers in far-flung communities. And while some program developers create feedback systems to monitor for program changes by implementing staff, a fidelity documentation review suggests that most often there is little to no feedback that is valid and reliable (Slaughter, Hill & Snelgrove-Clarke, 2015).
Dynamic change is a reality for successful social program scale up. The programs themselves evolve as leadership responds to opportunities to serve more communities and individuals. The organizations in the scaling partnership change by losing and gaining staff and capabilities, by expanding to offer new programs, and sometimes by spinning off a separate entity to focus exclusively on the scale up challenge while the parent organization goes about its core mission. And there is ever-present change in the sociopolitical environment as social problems rise and fall on media and public and policy agendas (Dearing & Rogers, 1996) and as funding priorities ebb and flow. In this research, while we did trace the history of the social programs studied, we made the decision to collect most of our data about a particular scale up phase in each program's history. And although we did not set out to measure the complexity of each program, social programs do vary in complexity and this reality must figure into how scale up partnerships strategize about pathways, partnerships, and fidelity.

RESEARCH QUESTIONS
The primary questions that guide our exploratory study across 45 social programs that have each achieved some degree of success in scale up are: 1. How

3.
To what extent do social program leaders work with partners to reinvent programs prior to attempting scale up? How do scale up partners manage program adaptation by implementers?

Strategic Choices: Pathways
Researchers, consultants and practitioners have produced models and guidelines about scale up. These models are often multifaceted and focus on several different variables. Typically, they assume that scale up occurs more or less in stages, with feedback loops tying the stages together. Embedded in these models are descriptions of pathways to scale up. These pathways and the language used to describe them typically come from business and management literature. The pathways to scale up presented vary on dimensions such as the degree to which a lead organization continues to control the delivery of products and services as the enterprise grows.
In applying ideas about pathways to the social sector, Simmons and Shiffman (2007) refer to a scale up strategy by which social innovations are communicated, transferred or otherwise promoted.
Clark, Massarsky, Raben and Worsham (2012) describe a business model where decisions are made to engage in geographic replication and non-replication options, such as affiliating with new partners or working to change policy environments. Management Systems International (2012) describes expansion, replication and collaboration as distinctive approaches for achieving scale. Gabriel (2014) writes about general scaling routes such as building a delivery network or forming strategic partnerships which are then further refined into pathways such as franchising and strategic alliances, respectively. These descriptions of pathways to scale can be viewed as theories of scale or a reasoned articulation of how to achieve scale up (Patrizi, Stephens, & McMullan, 2014).

BRANCHING PATHWAYS
Branching occurs when a lead organization increases its own capacity to offer the program at multiple sites in new locations or to new target groups (Mulgan, Ali, Halkett & Sanders, 2007;Dees, Anderson & Wei-Skillern, 2004;Kalafatas, 2006). In this type of pathway, the lead organization develops the program, distributes and implements it. Branching allows for considerable control over implementation because implementers are typically employees of the lead organization and go through the same extensive training with the same managerial and technical support that characterized earlier implementations of the program.
This may seem the most straightforward approach to scale and it is often the first pathway considered by social entrepreneurs (Gabriel, 2014). For complex social programs that depend on high implementation fidelity or for organizations that want to minimize variation, this model may be well-suited as this pathway allows the greatest level of control over how a program spreads. Scale up via this approach is usually deliberate and incremental, with multi-site additions often dependent on infusion of external investments, such as foundation grants or fee-for-service funding from participants or agencies in the new localities.

AFFILIATE PATHWAYS
Corporations have long offered franchises to investors as a means of scaling up a business. This occurs in the social sector through affiliates, which are not always formal franchises although arrangements do usually involve a legal relationship (Oster, 1995). Affiliation occurs when implementing organizations in the field buy or license the rights from a lead organization to offer the social program and the infrastructure that goes with it. This often includes a contractual agreement to follow specific procedures and processes (Gabriel, 2014;Mulgan, Ali, Halkett & Sanders, 2007). Dees, Anderson and Wei-Skillern (2004) define affiliation agreements as being prescriptive on many dimensions including the use of a common brand name, program content, funding responsibilities, and quality control. Affiliates reduce financial risk to the developer and lead organization, helping them bridge the gulf between a small-scale operation and a large-scale objective, partly by accessing local and regional resources to which developers otherwise would not have access (Beckmann & Zeyen, 2014).
Compared to branching, an affiliate pathway makes it possible for a lead organization to more quickly scale up a social program, as they are not solely responsible for providing the capital for new sites, hiring more staff, creating new relationships with communities, and so on. These responsibilities belong to the affiliate or licensee. A major trade-off is that the lead organization gives up some control-implementing partners are not their staff, lead partner personnel may never visit the location, and an affiliate may not only modify but simply end implementation for various reasons.
A distribution network pathway involves a lead organization working with a distribution organization to tap into the latter's existing networks of implementing organizations. Often the distribution partner is a national organization with many local member agencies, such as the YMCA or Boys & Girls Clubs of America. This strategy is variously described as a distribution network (Hussein & Kerrisey, 2013) a strategic alliance (Gabriel, 2014), "piggybacking" on another organization's infrastructure (Gabriel, 2014), or a networked approach (Wei-Skillern, 2014). For clarity we use only the term "distribution network" in this study. A distribution network allows partner organizations to focus on their shared mission (such as reducing poverty) rather than their own growth (Wei-Skillern, Silver & Heitz, 2014;Waitzer & Paul, 2011). This pathway can resemble a supply chain: One organization supplies and supports the social program, while a distribution partner delivers it and other partners support the program's effective use in organizations that adopt and implement the program.
In a distribution network, the lead organization may need to modify the social program, sometimes with the aid of the distribution partner, to increase the likelihood of successful local implementation (Scearce, 2011). Local implementing partners may or may not report back about how they change the social program to fit their local conditions. Lead and distribution partners may want to plan for that reality by creating methods for encouraging adaptations that customize to a specific implementing partner's capacities and to local clientele so as to maintain program fidelity.
There are other pathways to scale beyond the three types reviewed here, and social program leaders may scale using more than one pathway or transition from one path to another. In this exploratory study, however, for analytic purposes we locate each of the 45 social programs in just one of these pathways at a given point in time.

Strategic Choices: Partnerships
In the past, social entrepreneurs were portrayed as "solo operators, pursuing their agenda quixotically with little support from others. Today, this assessment has changed and it is generally accepted that successful social entrepreneurs are masters at mobilizing alliances of groups and individuals to all work together for a cause" (Bloom & Chatterji, 2009, p. 119). Many studies have noted that scale up often relies on inter-organizational partnerships in which partners complement one another's strengths, share a common mission, and extend reach into priority populations, (e.g., Bloom & Chatterji, 2009;Patrizi, Stephens & McMullan, 2014;Coburn, Catterson, Higgs, Mertz & More, 2013;Ackermann, 2013;Bradach & Grindle, 2014;Dearing, Maibach & Buller, 2006;Hussein & Kerrissey, 2013). The benefits of partnerships for scaling up include greater reach to appropriate local implementers, access to greater resources for scale up, access to skilled personnel with logistical expertise, shared risk and enhanced credibility.
Partnerships bring together partner organizations at local, state and/or national levels to achieve some common purpose. The relationship can be temporary or permanent, informal or structured through contracts or other legal agreements, and can be very limited or quite broad in scope (Backer, 2000). Partnerships can be defined based on the extent to which they are mutually dependent such as sharing goals and objectives, and maintain their organizational identity via their commitment to mission, core values and constituencies (Brickerhoff, 2002). In this study, we narrow the range of organizations on which we focus to core partners that comprise the leadership of a scale up initiative.
We explore the types of core partners found in selected scale up efforts and how they came to this partnership. Of particular interest is how core partners do, and don't, work together to modify the social program for scale up and control for its adaptation during and after implementation. Core partners are identified by leaders of the scale up initiatives.

TYPES OF PARTNERS
Organizations may perform one or more functions in the scaling process (Patrizi, Stephens & McMullan, 2014). Almost always, one organization plays a leadership role, linking the partner organizations, and steering the work of the partnership as strategic questions about scale up are addressed (Frost & Reich 2008;Trent & Chavis 2009). When scaling a social program, this leadership role is often attributed to the organization responsible for pushing to scale a social program. We call this type of organization the lead partner. The lead partner may have been the organization that created a program, led its initial implementation and tested it. In some cases, such responsibilities may have belonged to an earlier group and then the lead partner came along to push the scaling effort. Distribution partners work with lead partners to offer a program to members of its network of individuals or organizations. Local organizations are the implementing partners that directly deliver the program to beneficiaries. Other organizations are frequently involved in critical roles for financing, as a source of volunteers, as local community-level collaborators for reaching high-risk disadvantaged populations, and nationally for management, research, evaluation, communication and human resource expertise. We refer to all these additional organizations collectively as supporting partners.
Strategic scale up decisions are usually the domain of lead, distribution and/or supporting partners that gave rise to a program, provide major funding, partner for wide distribution or provide technical assistance. Yet social programs also may rely heavily on implementing partners for input about strategy concerning national or regional roll-out, partner selection, and pathways.

HOW KEY PARTNERS MEET
Studies about partnerships between organizations suggest that partners that have compatible goals or a shared purpose can reach consensus and retain commitment to a scale up objective (Tsasis, 2009). But finding a compatible partner poses challenges since many nonprofit organizations lack common, widely used communication channels and do not always have easy ways to learn about prospective partners (Austin, 2000). How do partners find each other? We identified five scenarios that describe how partners involved in scaling up an innovation may find each other: Happenstance, having a previous relationship, matchmaking by a third party, a search by the lead partner, or a search by a core partner other than the lead.

Strategic Choices: Fidelity
As defined earlier, fidelity is the extent to which a program is implemented in the way intended by its developers. Fidelity is often measured as the correspondence between how a program is delivered in tests prior to scale up compared with how the program is later offered by implementing partners in the field (Dusenbury, Brannigan, Falco & Hansen, 2003;Perez, Van der Stuyft, Zabala, Castro & Lefevre, 2016). Lead partners differ in the degree to which they reinvent programs in preparation for scale up, and how much they seek to maintain control over potential adaptations by field-based implementers. Some scale up researchers believe that a transference or shift in who feels ownership of a program or practice can by definition imply considerable adaptation, and is a necessary indicator of scale (Coburn, 2003;Dede, Honan & Peters, 2005).
We find reason to expect that the core partners in scale up partnerships can, by working together, affect both (1) the proactive reinvention of effective programs prior to scale up, and (2) shape the extent and nature of reactive adaptations made to programs so that they are better suited to implementing organizations and their clientele (Dearing, 2014;Leonard-Barton, 1988;Perez, Van der Stuyft, Zabala, Castro & Lefevre, 2016) as depicted in Table 1.1.

PROGRAM REINVENTION
Reinvention is a strategy where a developer or lead partner changes a program to heighten its chances of being considered, tried and effectively implemented. Not all social programs are reinvented. When reinvention does occur, it may be a decision by a lead partner alone or acting with other partners. Reinvention may reduce program effectiveness while increasing other characteristics (such as affordability). Done well, reinvention retains an acceptable degree of program effectiveness.
Why would lead partners reinvent their own program? A partnership may have developed a program for third and fourth graders but then realize that need is greater among fifth and sixth graders. Or a team may design materials to appeal to African American youth but then realize that need exists among Native American youth, too. Sometimes a funder learns of a successful program that is benefitting one population segment and wants to fund its conversion and delivery for a different segment of the population. Then there are considerations for how a program may be targeted to the right population, but needs to be less expensive, more culturally compatible, or made simpler to implement so that teachers or volunteers or nurses in low resource settings can do a good job in providing it to children or teenagers. Changes like these may reduce the effectiveness of a program but still be made because the reinvented program will benefit more people at still acceptable levels of benefit. An interest of ours is how, or if, core partners work together to reinvent programs. While continuums, categories and forms of inter-organizational partnerships have been posited and tested in the nonprofit sector (e.g., Austin, 2000;Austin & Seitanidi, 2012;Brickerhoff, 2002;Proulx, Hager & Klein, 2014;Kohm, La Piana & Gowdy, 2000), very little is known about if and how partners work to successfully reinvent programs for scale up. Drawing on the work of inter-organizational partnerships, and based on observations from this study, we identify three categories to describe the role of partners in reinventing social programs prior to scale up (Table 1.1).
Integrative reinvention occurs when the lead partner shares control in reinvention decisions which also require modifications to at least one of the core partner organizations. Integrative reinvention may consist of a distribution partner working closely with a lead partner to jointly redesign a social program, which, in turn, leads to transforming the structure and practices of the partner organizations. The partnership may be integral to the strategic success of each organization. Roles blur, projects are jointly developed and shared strategies are in place.
Lead and other core partners may engage in interactive reinvention. The lead partner actively seeks information from core partners for the purpose of program modifications and modification decisions are made by the lead partner or in conjunction with the core partners. For example, a lead partner seeks information about program implementation from distribution and early implementing partners that then informs modifications in the social program so that it better aligns with the perceived needs and interests of the distribution and implementation partners. The lead partner could also work closely with a foundation, a supporting partner, and draw on their internal expertise to modify the program prior to scale up, while still maintaining separate organizational identities.
Contractual reinvention consists of a lead partner directing or managing reinvention by way of a contract or other type of formal agreement with distribution, supporting or implementing partners. A lead partner may have a contracting relationship with a university or consulting group to provide an evaluation of the social program or with a publishing company to produce training materials. In some instances, an implementing partner may contract with a lead partner to deliver the social program but have little autonomy in determining goals or methods for scaling up the program. These relationships are low in mutual interdependence.
In some cases, partners working together in scale up initiatives may change their organizations in wholly unanticipated ways at least partly as a result of their interactions and shared experiences. So what may begin as a relatively modest program reinvention may escalate into organizational level change.
Little is known about if and how partners work to successfully reinvent programs for scale up.

PROGRAM ADAPTATION
The designers and proponents of some social programs demand a high degree of fidelity as a condition of scale up, site by site, in the belief that a given program can only be assured of being effective if it relies on the same principles, with the same manifest components, and is delivered just as it was originally delivered, when it was earlier tested and found to work. Extensive research (Backer, 2000) shows, however, that adaptation happens even when those demands are in place, and this reality must be taken into account by lead partners. Allowing for or encouraging customization or creativity when implementing an innovation is positively related to the likelihood of program adoption (Rogers, 2003). It may lead to greater commitment by implementers to a program and be positively related to its sustained use. In community settings, especially, an insistence on program fidelity is often unrealistic (Blakely, et al., 1987). Often the key is to find an appropriate, healthy balance between program fidelity and adaptation-and that balance may shift over time (Backer, 2000).
Adaptation is not necessarily detrimental to program effectiveness. This is so because those staff who implement and deliver programs to students, patients, clients, residents and other community members have direct contact with those people and also understand the capacities of their own implementing organizations-the schools, clinics, community centers and other types of local organizations. When frontline staff are trained and/or receive feedback about their performance, they can learn to make good decisions (Berta, et al., 2015) about adaptation of a program as well as to their implementing organization (Leonard- Barton, 1988).
Whereas our focus on reinvention activity is in the relations among core partners in making revisions prior to scaling a program, adaptation activity concerns the relations among lead partners and their implementing partners. How much input do implementing partners have in the evolution of social programs? Are systems created to capture their insights on a regular basis? Accordingly, we focus on four lead partner orientations to adaptation: Prescriptive fidelity, expectant fidelity, responsive adaptation and independent adaptation (Table 1.1).
Prescriptive fidelity occurs when implementers are discouraged from making adaptations and the lead or other partners are monitoring or controlling implementation. For example, a social program is to follow a specific order and scripting and those delivering the program are required to be certified in the program and/or they provide evaluation feedback to the lead organization who can then monitor outcomes.
Expectant fidelity occurs when implementers are discouraged from making adaptations but no one is monitoring to ensure high fidelity; that is, fidelity is expected but not controlled. We might Adaptation activity concerns the relations among lead partners and their implementing partners.
observe this when a well-tested social program comes with optional training and guidelines which can be reasonably followed without ongoing monitoring. It is also possible that there is little training and no monitoring and the lead partner simply expects fidelity.
When adaptation is encouraged or expected, and it is monitored or controlled, we term this responsive adaptation because the lead partner responds to feedback from implementing partners. In such a situation, a lead partner sees the importance of achieving compatibility between a program and each local implementing organization, but monitors adaptation through evaluation feedback or by requiring specialized training for implementation. The lead partner is thus ensuring responsiveness to the audience as well as the program's model.
Independent adaptation occurs when the lead partner anticipates or encourages some modification, and may provide training, but does not monitor implementation and adopters can act independently. We might observe this when a program consists of general principles that can be mixed and matched and benefit a population with little probability of experiencing a negative result.

Study Design and Methods
This study is a comparative case study of social program scale up. As is typical of case study research, we collected information about each case through multiple methods (Brewer & Hunter, 1989), chiefly by reviewing publically available documents and by interviewing program leaders. The primary unit of analysis in this kind of study is the case (Ragin & Becker, 1992), which also may be written about or referred to as an intervention or, for present purposes, a social program. Thus the primary unit analyzed is not a document or a smaller unit of text that describes an aspect of a social program. Instead, multiple documents and interviews are analyzed, each of which describes the same social program, with the objective of combining multiple sources and inputs to create one holistic portrayal of the scale up effort concerning that case.
As the number of cases increases in a comparative case study design, pattern-matching across cases takes on more importance in order to learn lessons across the cases, while qualitative detail about any one case recedes. We did this by following a protocol known as the case survey method, which decomposes cases into common variables, essentially treating qualitative data categorically, somewhat quantitatively (Yin & Heald, 1975;Yin, 2014). This is a way to compile qualitative information about a program's topical domain and objective, achievements, scope of operations, intended beneficiaries, pathway to scale being pursued, the organizations in the partnership, program characteristics and orientation to fidelity, etc., that is manageable for drawing conclusions across a large number of cases.

The Scale Up Challenge for Social Programs
This study is a comparative case study of social program scale up.
We chose the case survey method because, while essentially qualitative, it combines some of the objectivity of quantitative studies with the qualitative richness of case studies (Larsson, 1993;Yin & Heald, 1975) and can involve greater consistency in the collection and analysis of case data than narrative reviews of one or a few cases (Littell, Corcoran & Pillai, 2008;Wood, 2000). The method suited our objective since its strength is the ability to identify patterns across cases and enable pattern-matching during interpretation. For each case, as sources of information are combined and a common rubric or coding sheet is applied to them, the information from them constitutes a built case (Newig & Fritsch, 2009;Bullock & Tubbs, 1987). The set of built cases serves as the data-base from which information can be extracted and then analyzed to answer a study's research questions. This is the study design that we followed. Our design allows us to draw exploratory comparisons and associations that help with developing hypotheses for testing in other settings.

SEARCH PROCEDURE
Our search for effective social programs that had scaled up was based on recommendations from expert informants, database searching and web searches. Our research team, along with staff of The Wallace Foundation, identified colleagues knowledgeable about social programs in general, as well as specific partnerships for scale up, in particular. Expert informants are listed in Appendix A at the end of this report.
We searched multiple databases (e.g., Catalogue of Nonprofit Literature, Issue Lab, Education Resources Information Center, ProQuest Social Sciences) and the web using keywords, alone and in combination, including "social/innovation," "intervention", "social program", "nonprofit", "scale up", "collaboration," and "partnership." Databases included those focused on peer reviewed studies in academic journals, sources in the gray literature, and unpublished papers, including dissertations and working papers. We pursued a wide variety of sources to reduce bias related to prominence of cases, published versus unpublished status, and other characteristics (Larsson, 1993).
Expert identification coupled with systematic database and web searches generated an initial set of 105 cases that, on text review, we reduced to 79 based on whether each was a social program and was likely to provide relevant information about scale up.
Our design allows us to draw exploratory comparisons and associations that help with developing hypotheses for testing in other settings.

INCLUSION CRITERIA
Case by case, we reviewed the set of available documents to determine whether or not each social program met the following inclusion criteria: » Evidence that scale up occurred using a branching, affiliate, or distribution network pathway » Data gathered through research or program evaluations suggesting effectiveness of the social program » Evidence of at least one partner that appears to have facilitated or influenced the scaling up of the social program » Written records in English in the form of published studies, evaluations, review articles, web-based information, presentation slides, proposals and reports sufficient in amount and quality to provide data relevant to the study interests » Topical focus on education, youth development or health

THE SOCIAL PROGRAMS OF STUDY
Application of these inclusion criteria resulted in a purposive sample of 45 social programs in health, education, or youth development (listed in Table 1.2 and described in Appendix C), with representation of each scale up structure: Branching = 12, affiliate = 20, and distribution network = 13. Variety among these programs was intentional. We wanted to learn what was common among them as well as the degree to which branching, affiliate and distribution network pathways may be effective for different types of social programs. Programs using affiliate pathways were more prevalent in our initial set of 79 social programs so we included more of them in our purposive sample. Brief descriptions of these programs, including links to their websites, are in Appendix C. In chapters 2, 3, and 4 of this report we group the programs by pathway. To address questions that the written records did not contain answers to as well as to update the information gleaned from written documents, personal interviews were conducted with 100 leaders and staff associated with 44 of the social programs (one program leader did not participate).
Interviewees are listed in Appendix B at the end of this report. We began with the developer or leader associated with the social program and sometimes added one or more additional respondents if their perspective would add considerably to our understanding of the scale up experience. The focus of the semi-structured interviews was on partners considered core to the scale up efforts. We coded several aspects of the partnership such as the type of partner, their role in scale up and how they collaborated and identified themes across interviews. Interviewees were sent a draft of the results chapters for the present report with a request to review for accuracy. As with the document coding protocol, the interview protocol was drafted and redrafted several times, pretested, and finalized.
For four of the 45 programs we sought out additional interviewees, some of them interviewed in-person and on-site. We did this because we wanted the ability to write about some programs in narrative detail that would involve the perspective of implementing partners, not just national leadership. Even though this study is primarily about organizations that form the core of regional and national scale up efforts, the majority of the people who work on behalf of these programs work locally in their communities. These four cases were selected on the basis of them having strong education components, being willing to spend time with us to help us understand their program, and taking affiliate or distribution network pathways to scale up since these approaches involve more extensive partnerships than branching pathway programs.
All interview protocols were approved for use by the Western Institutional Review Board.

LIMITATIONS AND REPRESENTATIVENESS
Our purposive sampling procedure cannot ensure a valid representation for extrapolating to all scale up activity in the given domains to a larger population of initiatives. Ours is an exploratory study, as is much case study research. Our purposive sample netted fascinating but unrepresentative findings. It is possible that not all social programs require partnerships to scale up, other pathways to scaling can be used, etc. Strategic decision making can extend beyond a consideration of pathways, partners and fidelity.
In this study we selected only social programs that have had successful scale up activity. To varying degrees, proponents of these programs have bases for claiming program effectiveness, they have gone on to reach more places and more people, and they have bases for claiming that they are providing continued positive outcomes to beneficiaries. While it would be informative to compare effective with ineffective cases of scale up in order to draw distinctions between them, there are challenges with deriving such a population of scale up cases. First, there is no ready source for deriving a population of all scale up initiatives even within a single domain like youth development. Second, ineffective efforts at scale tend not to generate a record that can be accessed for analysis. Programs that fail tend not to publicize that fact, nor are their former proponents necessarily eager to talk about those experiences for the record. Even if written and submitted for publication, manuscripts about failed scale up efforts face long odds for acceptance.
Successes, on the other hand, may generate a lengthy paper trail that can be accessed and analyzed. As we have suggested, a purposive sample of successful efforts can enable the drawing of comparisons and associations, especially if the number of cases in the sample is large enough for the data and interpretation to gain credibility, such as Yin, Heald and Vogel's (1977) study of 140 innovations in state and local government. Our sample of 45 programs enables us to analyze the resulting data-set by qualitative pattern-matching, which can then serve the exploratory research purpose of generating hypotheses for future inquiry. We note that few studies of social program scale up have been conducted, and none with a number of cases this large, so there was an opportunity to contribute to an evolving field in a new and useful way.
Finally, another limitation of this study is that we did not set out to assess the effectiveness of each program. Others had done that. We did look for evaluations and studies that indicated some level of effectiveness. Many of the social programs in this study were studied using experimental methods. Others had well thought through evaluations. All had change models. But we cannot confirm the effectiveness of these social programs, especially as they have scaled up. Readers who wish to learn more about the effectiveness of these programs are encouraged to visit the websites listed in Appendix C.

Organization of this Report
The next three chapters present our results by pathway. We organize the chapters around pathways since pathway decisions are, in many cases, determinants of other strategic considerations in scale up.
Chapter Two focuses on social programs that have scaled using a branching pathway. Chapter Three looks at those programs that have taken an affiliate pathway to scale. Chapter Four does the same for social programs that take a distribution network pathway to scale up. We begin each of these chapters by illustrating one social program that has scaled using the featured pathway, then describe program characteristics such as evidence of scale up and time to scale up. Within each of the pathway findings chapters, we then focus on our remaining two major strategic considerations of which partners can best help with a scale up effort, and the fidelity concerns of reinvention and adaptation of the programs themselves. Throughout, we make liberal use of examples from our 45 programs.
We conclude with a chapter in which we discuss main results, implications and next steps for studies about scale up and include appendices about our expert informants, interviewees, and a description of the 45 programs of study.

Branching Pathways
Branching typically consists of a lead partner organization developing and expanding its own capacity to offer a social program at multiple sites in new locations or to new target groups. The lead partners pursuing branching pathways that we studied are nonprofit organizations, but not colleges or universities. Many, though not all, of these lead partners were founded to host the social program we studied. The social programs were typically offered in one or just a few locations for about 10 years before they started to scale. Implementing and supporting partners played limited, though important, roles, including involvement in decisions about reinvention and adaptation. Monitoring for fidelity is extensive for these programs. Interviews with leaders suggest that some branching pathway programs are in the process of changing to other pathways, partly at the urging of funders and partly based on learning from others who are scaling in different ways. Branching may remain the pathway of choice, however, for lead partner organizations that require or desire maximum control over social program spread and fidelity.

SCALING BY BRANCHING: PLAYWORKS
"Our whole model is based on solving a principal's problem, removing the chaos and injuries from recess," said Jill Vialet, Founder and Chief Executive Officer of Playworks, a national nonprofit based in Oakland, California, that supports learning and physical health by providing safe and inclusive play to low-income students in urban schools. In 1996, on a visit to a local Oakland school, Vialet saw how normal, unorganized recess activity left some kids out socially and physically, and how others routinely got into trouble during recess and caused problems for teachers. She began Sports4Kids to make recess safe and fun for all kids.
Sports4Kids came to the attention of the Robert Wood Johnson Foundation (RWJF) through a recommendation from Bill Drayton, the founder of Ashoka, a global organization that identifies, supports and invests in leading social entrepreneurs. Nancy Barrand, a program officer at RWJF, began a series of conversations with Vialet. The scale up of Sports4Kids eventually involved four activities, each of which also happens to be a division of the RJWF organization: Program development, evaluation, communication and finance. RWJF provided an initial round of funding for program validation and scale up. The program was re-branded Playworks, as part of a more specifically-targeted communications approach recommended by the foundation to better spread the word about the initiative. RWJF was a centrally important partner for Playworks. "I turned to them just like I did to my Board," said Vialet. "We talked about challenges. They helped me to find consultants to talk with about issues of scale up. They even helped me to interview our CFO candidate." Vialet's team began to scale up the program by expanding their organizational capacity. At each participating school, Playworks supported a full-time, year-round program with a "coach" employed by Playworks, with schools paying an annual fee of about $30,000. The coach delivered activities for recess, recruited students to co-lead games and activities, and coordinated play activities with classroom lessons. This personnel-heavy approach to the program meant that Playworks needed regional offices-essentially replicating a central administrative office in several field settings.
As the number of schools implementing Playworks grew, Vialet saw that not every school needed or could afford a full-time coach. This observation led to two lower-cost reinventions of the program: One in which the coach is shared by four schools, and one in which school staff are trained to do their own coaching. These leaner staffing approaches provided additional, complementary paths to scale up that could allow for more rapid growth than the regional office-supported branching approach with full-time Playworks coaches in each school.
Increasingly, local implementers also are adapting the Playworks program to their own circumstances. This is particularly true for those that only receive training and do not have a Playworks coach on site. Vialet said that learning to allow and support variance came through experience: "We had to learn how not to impose our definition of the value of play." Yet designing variants of a social program and encouraging adaptation at the school-level comes with risks. When is a variant no longer "evidence-based"? By providing guidance and regularly collecting evaluation data, Playworks staff are able to assess the effectiveness of the different ways that Playworks is being deployed.
The Playworks infrastructure has grown to include a national headquarters in Oakland and 23 regional offices overseeing implementation in 1,300 schools for 700,000 children, with plans for 7,000 schools by 2020, with the growth coming largely through the less personnel-intensive versions of the program and partnerships.

Program Characteristics
In the present study, 12 of the 45 social programs we assessed scaled using a branching pathway (Table  2.1.). Of these 12, several used other pathways, too; nevertheless, based on documents we reviewed and on conversations with program informants, our assessment is that branching was the primary pathway for these 12. Below, we review evidence of effectiveness and spread of the programs. We also briefly discuss the focus, time-to-scale and funding of these programs. We conclude this section with examples of how these programs started. We then discuss partnership and fidelity decisions.  All of these programs had an explicit theory of change that connects inputs to processes to outcomes. The change models are available on their web sites, or can be found in annual reports or in program assessments. Most of the programs had been evaluated quantitatively using randomized controlled trials. Others relied on evaluations that measured process and outcome variables quantitatively.
The Center for Employment Opportunities, which addresses employment barriers facing the formerly incarcerated, assessed its effectiveness through an independent 3-year random assignment research study conducted by MDRC, a research organization, and sponsored by the U.S. Department of Health and Human Services. This study found that Center for Employment Opportunities participants were less likely than individuals not in the program to have their parole revoked, be convicted of a felony, or to be re-incarcerated. In addition, the program was proven to provide taxpayer savings in the form of reduced criminal justice expenditures, up to $3.30 for every dollar spent on the program.
As another example, Public/Private Ventures conducted a four-year randomized trial of Higher Achievement. The study found mixed results. Findings led to changes in the program and additional evaluation. In 2014, the U.S. Department of Education named Higher Achievement a "highest rated applicant" for an Investing in Innovation (i3) Validation grant.
Sometimes, effectiveness is determined by process and/or outcome evaluations and not by randomized controlled trials. For example, College Summit relied on ongoing evaluation to assess effectiveness. They developed an extensive data gathering system with support from Deloitte, a key corporate sponsor. Weekly reports at the school and district level tracked how many ninth graders had developed a postsecondary plan, how many seniors had written an admission essay, where they applied, and so on. These data were used as evidence that the program was meeting its objectives to get high school students from low-income communities to and through college. These data were also used to modify the program and to make changes in its implementation.

EVIDENCE OF SCALE UP
All 12 branching pathway social programs have scaled up to varying degrees. For example, WINGS, an afterschool education program that teaches children how to behave, make good decisions and build healthy relationships, began in 2007-08 in one school. They have added schools one at a time and are now in 11 schools. WINGS has expanded out-of-state and expects more rapid growth moving forward. A deliberate pace for growth has also been established by Posse Scholars. Deborah Bial, President and Founder, described the first ten years of operation (1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999) as making sure "we knew what the program looked like. Then the second phase began in 1999 with a grant from FIPSE [the U.S. Department of Education's Fund for the Improvement of Post-Secondary Education] to pilot test a replication site in Boston. Now we have opened a new site every other year." In 10 regional sites they have more than 50 academic institutional partnerships. Posse Scholars plans to grow to 100 partnerships with colleges and universities by 2020 and recruit 1,000 students a year. Reading Partners, a children's literacy program, has grown from six schools in Silicon Valley in 2006 to more than 60 school districts nationwide, including some in New York City, Los Angeles and Washington, D.C.
In the 2014-15 school year, Reading Partners served over 8,500 students in 160 elementary schools.

PROGRAM FOCUS
Education was the focus of 10 of the 12 social programs using a branching pathway. Programs focus on elementary, middle and high school students, and some follow students across multiple years. The National Writing Project benefits schoolchildren but focuses on k-16 teachers. The Center for Employment Opportunities is unique in its focus on employment for formerly incarcerated persons, as is AARP's Experience Corps focus on adults over 50 and schoolchildren.

TIME TO SCALE UP
Programs were well established prior to scaling up. On average, scale up occurred after a program had been in operation for 10-12 years (see Figure 2.1.). Recall that Posse Scholars spent 10 years developing and codifying their program and then slowly began to scale. Earth Force, which started in 1993, began scaling in three years as it skipped the development phase by purchasing the intellectual property of a program from another organization. Scale up occurred in two phases for the National Writing Project. It began with creating a federation of local sites from 1974 -1990. During that period the National Writing Project created over 100 writing project sites with some states building out state networks. In 1991, with directed investment from Congress, they were able to expand further with attention to underserved areas across states, eventually reaching 205 sites.

SOURCES OF FINANCIAL SUPPORT
In the early years of developing programs, lead organizations relied primarily on foundations and individual contributions for financial support. By the beginning of scale up, the funding picture becomes more complex and most programs were relying on multiple funding sources, with foundations still playing a lead role, but federal funding and corporate funding becoming more apparent. College Summit illustrates the kinds of changes in funding that many of these social programs experienced. It started off as a grassroots effort and for its first several years operated on less than $250,000 per year. A grant in 1998 of $30,000 from the Knight Foundation evolved into a challenge grant of $480,000 from Knight for organizational development and expansion of regional services. By 2002, College Summit had operating revenues of $3.5 million. In 2004, foundations provided 23 percent of the total funds raised by College Summit. Another 46 percent came from federal government sources like the U.S. Department of Education, AmeriCorps, as well as from contracts with school districts (Bridgespan, 2004). In fiscal year 2015, College Summit had revenues exceeding $19 million.

PROGRAM ORIGINS
These programs largely start from the personal interests of one or a small group of leaders who then go on to create and then expand a community-based organization as a way to host the program. Many of these nonprofits were established expressly for the purpose of developing and implementing the social program of study. Reading Partners had a founding story that is similar to other organizations in this study. In 1999, three community leaders, Mary Wright Shaw, Molly McCrory, and Jean Bacigalupi launched a one-on-one tutoring program at Belle Haven Community School in Menlo Park, California to help children who had fallen behind in reading. In 2001, the three founders recruited a dedicated group of volunteers to tutor students, and in 2001, YES Reading was incorporated as a 501(c)(3). In 2002, they expanded to a second school. In 2004, they hired a full-time executive director. The program continued to expand and was renamed Reading Partners in 2008. By 2014, Reading Partners was serving 8,900 students in 169 partner schools and had more than 12,000 volunteers and an annual income of $25 million, most of it from foundation grants. Higher Achievement, Playworks, Posse Scholars, Summer Search, WINGS, and College Possible have similar stories of highly engaged individuals who started small and gradually grew to larger organizations with more staff and more complex budgets serving hundreds, if not thousands, of students.
It is common that these nonprofit organizations and the programs they developed share the same name. For example, Summer Search is a leadership development program designed to help Many of these nonprofits were established expressly for the purpose of developing and implementing the social program of study.
students from low-income backgrounds transform what they believe is possible for themselves and develop the skills they need to become college-educated leaders who give back to their families and communities. Summer Search is also the name of the host organization. Giving the program and the host organization the same name may help initially with brand identity, but it may lead to confusion and a need later on to distinguish the two if the host organization decides to change or add more programs to its work portfolio.

Strategic Choices: Partnerships
Lead partner organizations in a branching pathway control decisions about scale up. During phone interviews, we asked leaders of lead organizations to identify their core scale up partners. The core partners mentioned most frequently were implementing partners, followed by supporting partners. They do not identify distribution partners. Lead partners find core partners by actively looking for them.

LEAD PARTNERS
All 12 lead partners are nonprofit organizations, and none are universities or colleges. Most of these lead partners have annual revenues in 2014 or 2015 fiscal years in excess of $15 million. Fewer than a quarter had revenues of less than $10 million a year. Some lead partners are quite large, such as the AARP Foundation with revenues of more than $146 million in 2015. Experience Corps is just one of many programs managed by the foundation. The AARP Foundation Experience Corps staff hosts Experience Corps programs in four cities. Seventeen additional "affiliates" are hosted by other non-profit or public institutions and look for financial commitments from the community to sustain their work.
Lead partner organizations typically have regional offices that manage programs in their areas. Higher Achievement has Achievement Centers in Baltimore, Washington, DC, Pittsburgh and Richmond, VA. Playworks has 23 office locations other than its headquarters. The National Writing Projects operates sites at nearly 200 universities.
Many of these nonprofit organizations were initiated to support the program studied. For example, College Possible and College Summit programs are housed in organizations of the same name that started because of the programs. Universities played instrumental roles in the development of two programs.

IMPLEMENTING PARTNERS
Implementing partners are core to branching pathway efforts. Elyse Eidman-Aadahl, Executive Director of the National Writing Project, said "We believed from the beginning that we couldn't achieve the outcome we wanted without partners. We can only succeed collectively." The most common types of implementing partners were schools, summer program partners and colleges and universities. Specific schools, programs and colleges were not identified. For example, Bridget Laird, CEO with WINGS, said, "Our biggest partners are schools where we operate and local partnerships with organizations that have afterschool activities like Junior Achievement, Girl Scouts, local chess, and gymnastics groups that offer these activities." Several leaders mentioned AmeriCorps members as essential for implementation. Jim McCorkell, with College Possible, described AmeriCorps members as "key to the work that we do in high schools." AmeriCorps members help students in preparing for college admission tests and help to put together their applications, find financial aid, and work with the students to develop leadership and other skills. Not all of the leaders included in our study whose programs depend on AmeriCorps members identify this federal program as a core partner, though community service organizations clearly play critical roles in staffing the scale up of these programs.
Sam Schaffer, the Executive Director and Chief Executive Officer with the Center for Employment Opportunities, identified a state agency as a core implementing partner. Schaffer said, "For us to do our work and concentrate on serving high risk men coming from prison we need probation and parole buy-in. The state department of corrections needs to be aware of and support what we're doing with the people they're helping to re-enter the community. When we're looking to expand into a new state, our first meetings are with the state department of corrections." "We believed from the beginning that we couldn't achieve the outcome we wanted without partners. We can only succeed collectively."

SUPPORTING PARTNERS
Foundations were named as core partners by about half of the leaders interviewed who were using branching pathways. Jill Vialet said that RWJF was a centrally important partner for Playworks and described them as playing more than a funder role -they were consultants and thought partners. Lynsey Wood Jeffries with Higher Achievement said, "The Wallace Foundation invested in a formative assessment of Higher Achievement to help us understand whether we're having school-wide impact. They also put together a professional learning community that has been really effective. We've learned so much from other Wallace grantees." The Edna McConnell Clark Foundation (EMCF) was mentioned by several respondents because of the foundation's roles as an intermediary for Social Innovation Funds. For example, Michael Lombardo, former Chief Executive Officer with Reading Partners, said that EMCF helped with their scaling and sustainability strategy citing their funding as making it possible for Reading Partners to "go public." Program leaders did not identify consultants as core scale up partners although they referenced them during interviews and in documents. For example, Deloitte became a corporate supporter and worked with College Summit to help improve its methodology in measuring and increasing college enrollment rates across schools and school districts. Deloitte ultimately developed the organization's data gathering system, assisting College Summit in being able to scale to more schools. College Possible worked with McKinsey & Company to develop a strategic planning process that resulted in the organization's plan for further national expansion as a nonprofit. Bridgespan was referenced several times because of their role in developing business and scale up plans.
While essential to growth, providing financial resources or consulting expertise may be insufficient to be considered a core partner. This point was made by Deborah Bial, President and Founder of Posse Scholars, who said, "Funders like FIPSE, public school districts, and corporations that provide jobs are not partners. They are allies. They support the program but the complex partnership relationships that make Posse Scholars run are with the group of higher education institutions."

DISTRIBUTION PARTNERS
Leaders of programs that scale via branching do not mention distribution partners as core partners. This was expected, since branching pathways do not involve distribution partners. But changes are underway in several of these programs and distribution partners could be on the horizon. During phone interviews, we heard about very recent or planned scale up activities that may alter or expand the choice of pathways in the future. One example came from Experience Corps. During an interview with Barb Quaintance with the AARP Foundation, home to the Experience Corps program, she talked about shifting models for scale. She described the current in-school model that includes branches wholly owned by the AARP Foundation that run centrally, as well as affiliate sites hosted by other non-profit or public agencies. She also mentioned piloting projects with afterschool programs as a new distribution channel.

HOW KEY PARTNERS MET
The lead partner organizations using branching pathways to scale met their core partners primarily by searching for them. Several leaders indicated that they choose their expansion sites and implementing partners and did so carefully. Posse Scholars has "a very slow process of developing partnering relationships with colleges and universities because there is such complexity to the relationships and it requires a lot of resources. These relationships are developed very carefully and are quite complex and custom-tailored," said President and Founder Deborah Bial. The president of the university or college must be fully committed to the program, she said, especially since they are making a significant financial contribution that includes operating funds for the Posse Foundation and 10 full-tuition scholarships a year.
Bridget Laird, with WINGS for Kids, also described a careful process of identifying their expansion sites. She described partnerships that were unique as "each partner focuses on a distinct set of outcomes." WINGS uses milestones to determine if a region is ready for expansion which include a commitment of funds from schools, local foundations, corporations, and/or individuals, as well as support from each school, the creation of a regional advisory board, and student recruitment plans. Their close attention to scale up process was partly a response to an earlier attempt to scale through an affiliate pathway that compromised the program.
In several cases, supporting partners approached lead partners. For example, Bloomberg Philanthropies asked College Possible to be part of a coalition of organizations with the aim of helping high achieving low-income and moderate-income students earn admission and enroll at selective colleges and universities.
Far less common were third party connections or a reliance on previous relationships. No one identified meeting a partner through happenstance or serendipity.

Strategic Choices: Fidelity
Minor modifications by lead partner organizations were common across these programs and although core partners informed reinvention decisions, these were not necessarily shared decisions. Programs are nearly evenly split in the extent to which adaptation is encouraged, but all of these 12 programs are closely monitored for fidelity of implementation and outcomes. Implementation fidelity is of primary importance for lead partners.

REINVENTION
Nearly all of the lead partner organizations in this pathway have reinvented their social programs prior to scaling them up. Some reinvention involved making modifications to a program but keeping the basic program model intact. For example, Summer Search piloted a modification in the delivery of the program from 1:1 to group-based mentoring. AARP Experience Corps changed the focus, delivery and cost of its program. The focus changed from being about the health of older adults to a balance between students, schools and older adults. Leadership also reduced the time commitment required of volunteers and placed restrictions on stipend availability to reduce program cost.
Reinvention was sometimes more radical, though, such as adding complementary programs and adding or changing scaling pathways. As the number of schools implementing Playworks grew, the founder, Jill Vialet, observed a great deal of variation. For instance, a Teach for America instructor used the Playworks Playbook (a publication which summarizes the program and how to implement it) to create a Playworks-type program that was implemented entirely without assistance from the Playworks organization. The impact of play was also being defined in a number of different ways: In California it was social-emotional learning, in Colorado it was physical activity, in Philadelphia it was violence prevention. Vialet saw a need to create variants of the original Playworks model. The original model called for full-time, year-round program coaches who are employees of Playworks in low-income urban schools. A revised model was developed to train school personnel and youth development agency staff, but Playworks is not involved in the implementation or sustainability of the program. A third model was added where Playworks employs a site coordinator who works with four schools at a time with participating schools providing a "recess team" and a coach. In this model, most of the responsibility for the success and sustainability of the program rests with the school. The leaner staffing approaches provided additional, complementary paths to scale up that could allow for more rapid growth than the branching path with full-time Playworks coaches in each school.
Reinvention can occur at multiple points in time. Earth Force was started in 1993 with a grant from the Pew Charitable Trusts with the goal of engaging young people to improve their communities through practical activities. The initial interventions were communication campaigns run from a central office to raise awareness about environmental issues. Effects from these campaigns were small. Earth Force changed the intervention to be a curriculum-based program framed by a six-step process (that is still the core of their programs) and they opened community-based offices. This was their initial point of scale up. Reinvention did not stop there. The economic downturn of the early 2000s led to cutbacks by funders and staff layoffs. Leadership had to reevaluate how they were going to survive and grow their impact. A strategy of partnering with organizations that had similar objectives and values (civic engagement, environmental education and action, community organizing, youth mobilization), and brought resources that Earth Force needed, offered a solution. This strategy included working with corporations to have their employees volunteer. "We realized that we could really scale up once we got into discussions with corporations," said Chief Executive Officer and President Vince Meldrum. "Having trained professionals work as volunteers meant that they could take on much of the support burden with local educators that had been the responsibility of paid Earth Force staff." Higher Achievement provides another example of repeated reinvention. Higher Achievement grew its infrastructure beginning in 2006 to support scale with a high degree of fidelity, an effort undertaken after receiving an investment from Atlantic Philanthropies. They centralized their existing functions, such as program implementation monitoring, data evaluation and financial management, and they standardized their operating procedures for every element of the program model. Higher Achievement practiced oversight of each sites' quality with regular assessments, role-specific observation tools, weekly staff coaching and more. But after several years, Higher Achievement leaders realized that their model alone was not sustainable. So they launched a new approach to scale up that expanded their reach to middle school youth. They continued to replicate their traditional program, but added partnering with other youth-serving organizations to provide academic experiences while also lowering costs and avoiding duplication of services.
Higher Achievement writes in its business plan, "[We are] a learning organization and dedicated time is allocated throughout the year for debriefing and reflection. While the goal of serving more youth does not change, annual strategies reflect information gleaned from review of the changing landscape (education, economic, funding, etc.)." This comment could be said about many, and perhaps all, of the organizations operating in branching pathways in this study.
We examined how partner organizations work together to reinvent social programs. Not surprisingly, in a branching approach to scale up, partners do not figure prominently in reinvention decisions. We did not find examples of integrative reinvention where partners jointly decide to adapt programs knowing that these changes will necessitate organizational changes as well. Nor did we find examples of these organizations engaging in contractual reinvention. But we did find examples of core partners engaged in interactive reinvention where the lead partner sought information from other core partners that resulted in changes to the social program.
Reinvention can occur at multiple points in time.

INTERACTIVE REINVENTION
Although final decisions about programmatic or scale up changes typically reside with a lead partner in branching pathways, decisions about modifications to promote scale up were influenced by supporting partners, especially funders and consultants. Funders influenced scale up decisions when they provided fiscal resources, but more so when they brought expertise. For Playworks, RWJF brought more than funds; they brought expertise in communications. "Having the Robert Wood Johnson Foundation as a key partner has shown us how important communication is to scale up. Communications will now be the centerpiece of our new strategic initiatives going forward," said President and COO Elizabeth Cushing.
Several of the programs in this study received Social Innovation Funds via the Edna McConnell Clark Foundation and these funds, and the foundation, have influenced decisions about scale. In a blog posted by Michael Lombardo, former Chief Executive Officer with Reading Partners, he wrote that the investment by the Social Innovation Fund enabled Reading Partners to go from a promising local program to a nationally replicated model. According to Lombardo, "The prominence of the SIF opportunity was alluring. In some ways, winning SIF funding from EMCF was our 'IPO,' the thing that allowed us to take our intervention public" (Kersey, 2014). Higher Achievement's director, Lynsey Wood Jeffries, describes a professional learning community convened several years ago by The Wallace Foundation where groups like hers "facing the same basic challenges could dive into policy issues and discuss lessons learned from programs like ours. We have learned so much from BELL and Citizen Schools as well as from other Wallace grantees." Federal funding from the U.S. Department of Education was instrumental to scaling up the National Writing Project. But Elyse Eidman Aadahl, the Executive Director, said that these funds also came with constraints since the Department of Education did not encourage the seeking of other funding.

ADAPTATION
The social programs using branching pathways are committed to monitoring the implementation of their program. Amy Saxton, Chief Executive Officer with Summer Search, said, "There is no point to scale up if there are no outcomes." This was a common view expressed in interviews and in documents. Across the programs are many examples of the use of contracts, memorandums of understanding, formal agreements, data reporting requirements, dashboards, ongoing evaluations and reviews, required trainings and meetings to monitor and control the implementation process. They were somewhat evenly split in terms of whether or not the social program could or should be adapted to local circumstances.
"There is no point to scale up if there are no outcomes."

RESPONSIVE ADAPTATION
Vince Meldrum of Earth Force, said, "We encouraged adaptation at the local site level because whether we encourage it or not, it happens." Still, the lead partners that encouraged or anticipated adaptations actively monitored for changes. The National Writing Project (NWP) provides a good example of balancing adaptation with implementation fidelity. NWP is a nationwide professional development network that serves teachers across disciplines and at all levels. Its mission is to improve student achievement by improving the teaching of writing in the nation's schools. NWP works to meet this mission by using a specific model that is customized for local needs. NWP has a national program model, adheres to shared principles and practices, and offers programs that are common across the network. Built into NWP is the opportunity for local programs to identify their needs and interests and to build strategies to address them, and modify strategies as needed to achieve success. NWP prescribes elements of each site's design, such as a requirement to provide continuing education and research opportunities for teachers, even as it leaves other elements to local adaptation, such as conducting programs for youth, parents and/or community members.
The NWP has several components in place to track progress. They have a rigorous annual review process where a group of 50 writing project directors and lead teacher-consultants read and review proposals submitted for renewed funding. Reviewers attend to the basic model of the summer invitational institute, school year in-service programs, continuity programs for teachers in the local service area, and development of teacher leadership at the site. All of these components are deemed essential to ensuring the health of a writing project site. Site leaders facing challenges in implementing the model are supported through a program of technical assistance coordinated by the NWP national office and planned jointly by national and local leaders. If a site cannot make the necessary changes after significant technical assistance has been provided, it loses funding. NWP closes three sites a year on average. The training approaches used by Playworks draw on the program's basic concepts, but there is room for adaptation. Process and outcome evaluations from the many implementing sites inform checklists and free guides that help implementers adapt the program to their schools. By gathering and sharing stories about the successes and challenges from across the country, Playworks can both learn about implementation but can also direct or steer implementation. Playworks also conducts an annual survey of school staff to learn more about program impact on students and school climate.

PRESCRIPTIVE FIDELITY
About half of the programs seek to limit adaptations and monitor implementation to ensure fidelity to the program. Posse Scholars illustrates the type of monitored fidelity that several programs use. Posse Scholars is a four-year college scholarship and peer support program for public high school students with academic and leadership potential, and is negotiated carefully with university presidents at potential scale up sites. Deborah Bial said that negotiations include an agreement that the program is "implemented quite precisely in each scale up location. This is a condition of establishing the partnership." Bial described partnering as a "very slow process of developing a relationship because there is such complexity to the relationship and it requires a lot of resources. Each college or university requires a specialized, difficult-to-negotiate way of work and this is instrumental to success." Each university has a unique contract. Carefully negotiated agreements set out the conditions for each partnership and Posse requires working with the president to finalize any deals. Bial described Posse Scholars as "a unique model that is not easy to generalize. We are tightly controlled as an organization and our program is narrowly defined." Posse Scholars, then, represents a form of prescription where each contract is nuanced, responding to the local culture and conditions, but the program components are specified and the contract must be followed.
Posse Scholars, like several other programs, increases the likelihood of meeting their goals by conducting feasibility studies to support the decision to expand. They also centralized training, finance, evaluation and communications. Regional or local sites have directors who oversee local training, fundraising and staffing. A core program curriculum is used across all the branches to ensure consistency across sites. Annual retreats, site visits, regular meetings, required reporting and ongoing evaluation ensure quality and compliance.
The program is "implemented quite precisely in each scale up location. This is a condition of establishing the partnership."

Affiliate Pathways
Most of the lead partners we studied taking an affiliate pathway for scale up were nonprofits. A few of them were colleges or universities, or private organizations. Their core partners were support organizations, such as foundations and consulting groups, and implementing organizations. How partners met is more varied than what we found among branching partners, though partners in affiliate paths still primarily meet through the initiative of lead partners.
Most affiliate pathway partnerships had modest national offices with few staff though a few of these partnerships had large numbers of national and regional staff. Overall, however, the lead partner organizations appear to be leaner in an affiliate versus a branching pathway, probably because the responsibility for service delivery shifts from a national to a local level. Affiliates or implementing partners typically have to prove their worth to national or regional leadership and contracts may require that certain conditions are met to maintain affiliation. We see a pattern where programs with a research basis and health focus emphasize fidelity.
Here we assess 20 cases that used an affiliate pathway for scale up, showing both the advantages and challenges of this model of growth for programs and their impact. That 20 of our 45 cases pursue scale up through affiliate pathways reflects their prevalence in our search results.

SCALING BY AFFILIATION: NATIONAL ASSOCIATION FOR URBAN DEBATE LEAGUES
Organized competitive debate (also known as forensics) has been a high school activity in the U.S. for more than 50 years, with tens of thousands of students learning to study and research complex public policy topics, master the logic of rhetoric, argumentation and public speaking, and become effective school leaders. While debate programs in the U.S. have flourished in private and suburban public high schools, many urban schools had not hosted debate teams, partly because of resource shortages. To bring the benefits of organized debate to urban and minority youth, the National Association for Urban Debate Leagues (NAUDL) launched its first Urban Debate League in Chicago in 2002. Results have been promising. Effectiveness in organized debate is highly correlated with academic achievement: 90% of urban debaters graduate on time, 85% enroll in college, 84% are students of color, and 72% are from low-income families. During the 2016 calendar year alone, 9,479 urban debaters delivered 168,000 speeches in 555 urban schools in 22 cities.
A league is structured so that multiple high schools, and sometimes middle schools, send debate teams to compete with each other. Each year debaters throughout the country debate a single complex policy question, or resolution, for an entire year. Coached by teachers, debaters conduct extensive research on the resolution and develop arguments for and against it. They hone their arguments in afterschool practices and compete at weekend tournaments. At each tournament two-person teams participate in a series of 75-minute debates. Each team alternates sides, arguing for the resolution in one round and against it in the next.
Debate leagues require committed teachers as coaches in each school, along with students in multiple grades so that a "pipeline" of up and coming students is developed and younger students can learn from older students. A league also requires partnerships with local schools districts and school boards, as well as with the community. "When we start a new league, we meet with members of the local school district and school board. We recruit them. We need a strong local presence for their organization to grow and maintain," said Linda Listrom, Executive Director of NAUDL. "We negotiate for a school-based after-school program. And we discuss if the district will pay coaches a stipend, provide bus transportation and a few other items. The extent of commitment can vary." When a district commits, Listrom and her staff assist in finding the right person to chair a board of directors, someone who can help in recruiting additional directors for the league who are well-connected in the local community and will commit to accessing resources in the community to support the league and the debaters. Having a league board with deep roots in the community is vital because of how conditions can change with school districts. "You can't just be funded by the school district," explained Listrom. "It's a problem because if the school has a budget crisis there is no way for the program to survive." The national office provides leagues with seed money and technical assistance until they become their own 501(c)(3). It can take a couple of years to get to this point. Once established, the league becomes an affiliate with the national network.
A key challenge for Listrom and the NAUDL is the nature of an affiliated set of independent organizations. "You can't be bossy," she said. "But if you don't boss them, how do you get them to do what you want? I'm using carrots not sticks." Part of what works well is the leagues having leeway to experiment. "We see innovative things that come from the leagues, some of which other leagues have adopted," said Listrom.
Urban Debate Leagues have a key source of support that further diversifies and leverages the headquarters and franchise leagues: The Citi Foundation, which supports Urban Debate Leagues by encouraging expansion in cities where Citi Bank and Citigroup have their largest "When we start a new league, we meet with members of the local school district and school board. We recruit them. We need a strong local presence for their organization to grow and maintain." operations as well as establishment of leagues in urban areas where Citi has a large presence but debate leagues did not exist. The Citi Foundation had a Pathway to Progress initiative for low-income urban youth for which Urban Debate Leagues was a good fit. In this way, Citi serves as a national partner to supply Citi volunteers (Citi employees who are required to volunteer a few hours per month in community service) who then can serve as judges for league tournaments, which can each involve 700-1200 debaters. The partnership with a national corporate foundation relieves some of the staffing pressure of moving into and then expanding in large urban areas like Miami.
Affiliate pathways work because the developer or lead organization identifies a local affiliate (such as a school district in the case of Urban Debate Leagues) that then can use its social capital with local business and nonprofit organizations to build up a local presence and enthusiasm for the program. In this way, the national program becomes a local program. Not all social programs using an affiliate pathway to scale up embed themselves as deeply in a community as do Urban Debate Leagues, but not all social programs require such an extensive local infrastructure. And unlike the affiliation process for Urban Debate Leagues, some programs that pursue an affiliate pathway are more reactive than proactive by only initiating the affiliate process when a prospective affiliate contacts them.

Program Characteristics
Twenty of the programs we studied used affiliate pathways to scale up (Table 3.1.). Here we review evidence of effectiveness and spread of the programs, and provide an overview of the focus, time to scale and funding of these programs. We conclude this section with examples of how these programs started. We then discuss partnership and fidelity decisions.

Affiliate Pathways
The national program becomes a local program. Prior to attempting scale up, these programs produced evidence of program effectiveness with various degrees of rigor, partly depending on whether they began as research projects based at universities or not. Non-experimental study designs to generate data such as cohort studies, prepost designs, or cross-sectional surveys were used by more than half of the cases. A quarter of the programs measured effectiveness using a randomized controlled trial and about an equal number used qualitative case studies, interviews and observations to provide evidence of effectiveness.
Intergenerational Tutoring did not initially have a rigorous study of effectiveness, although the program had been active for more than 20 years. The lack of a rigorous study design for Intergenerational Tutoring, despite its longevity and popularity as a program offered at Oasis Centers across the country, proved problematic. "This presented a roadblock for its growth," said Oasis Institute President and CEO Marcia Kerz. "As we began seeking funding to scale the program from corporate, government and private sources, it became apparent that our survey feedback, while encouraging and informative about the level of our partners' satisfaction with the program, just wasn't enough." This led to a formal partnership with Maryville University in St. Louis to conduct a study using a pre-test, post-test, control group design.
The developer of The Incredible Years Programs, Carolyn Webster-Stratton, Professor Emeritus from the University of Washington used random assignment to intervention or control groups to measure program effectiveness in multiple studies over 3 decades. Measures included parent and teacher reports, child assessments, home, laboratory and school observations of parent-child, and child social skills, attribution and self-esteem testing. Assessing effectiveness is ongoing for programs offered by Incredible Years. Webster-Stratton said: "My experience scaling up IY [Incredible Years] has taught me that EBP [Evidence-Based Practice] program development must be thought of as an ongoing building process rather than an endpoint. New data will continually emerge to inform real world clinical practice and each unique setting or environment can inform improvements or adaptations to the construction process and further research." Some programs, like Boot Camp for New Dads, relied on surveys and post-workshop evaluations to determine effectiveness. Moneythink assessed student impact through diagnostic and summative tests and surveys of 350 students in their launch year (2012-13) and saw improvements in financial knowledge and attitude across the board. Like programs in the branching pathways, the affiliate pathways programs each had an explicit theory of change model embedded in their work.

EVIDENCE OF SCALE UP
Just as Urban Debate has spread to more implementation sites through organizational affiliation, so too have the other social programs in our study that have taken this pathway to scale up. While eight of these programs have spread to between 11-50 sites, seven others have spread to between 501-1000 sites. For example, Gateway to College is an alternative education program that serves 16-21 year olds who are off-track or out of school. Graduates of the program have better attendance, fewer behavioral problems, improved academic performance, and graduate from high school with an average of 33 college semester credits. With headquarters in Portland, Oregon, the program has scaled up from Portland into a national network.
Another example of scale up via affiliation is Girls on the Run, a program to help build self-and collective efficacy among 8-to 12-year-old girls. Each of more than 200 Girls on the Run local councils is an autonomous affiliate of the national organization. The program developer began in North Carolina with 13 girls. Today, 100,000 volunteers serve 185,000 girls per year in more than 200 councils in all 50 states. More than one million girls have participated in this program.

PROGRAM FOCUS
The 20 affiliate pathway programs vary in their focus. The Campus Kitchens Project focuses on food security and KaBoom! builds playgrounds. Many are focused on learning and are associated with the school day and school age children. For example, Citizen Schools' Expanded Learning Time extends the school day and is designed primarily for middle school students. Other programs focus on high school or college-aged students. Some programs focus on relational support, such as Boot Camp for Dads, which focuses on fathers engaging with their infants, supporting their mates and navigating their transformation into fatherhood.
Health is a focus of several programs as well. Clinical Chapters offers psychotherapy to current and former foster children, Wyman's Teen Outreach Program empowers teens to avoid risky behaviors such as teen pregnancy, and KaBoom! focuses on active play.

TIME TO SCALE UP
The social programs vary considerably in the amount of time a program was offered prior to scaling it up. About a quarter of the programs had been offered in one or just a few sites for 20 or more years before scaling up. Nurse-Family Partnership (NFP) was first tested in 1977 and was carefully expanded to a couple of test sites. After about 20 years of testing and observation, the developer, David Olds, a professor of pediatrics, psychiatry and preventive medicine at the University of Colorado Denver, was satisfied that NFP was ready for replication, and scale up began. Across the 20 programs in affiliate pathways, the average time from program origination to the beginning of scale up was 11-12 years.

SOURCES OF FINANCIAL SUPPORT
Foundation and individual contributions were the primary, and sometimes only, identifiable source of funding for programs when they were first developed. During the scale up period, funding became more complex -different sources and more of them. Nearly all lead organizations were receiving funds from foundations. Foundations, however, were not a primary source of support for programs that were receiving federal funds. We observed increases in the use of fees for service, corporate support, and individual contributions. The majority of organizations at the point of scale

PROGRAM ORIGINS
The social programs in the affiliate pathways have differing points of origin. Some started from the personal interests of one or a small group of people, similar to those in branching pathways. Several started as university based pilots or interventions that eventually became a spin-off organization. A couple were purchased. These origin stories illustrate the diversity of programs and organizations found in affiliate pathways.
Girls on the Run was founded by Molly Barker. As interest in the program grew, Barker tried to meet the demand herself, thinking that only she could accurately deliver the program lessons and activities. She was coaching at multiple sites and trying without much success to make it as a for-profit venture supported by tuition fees. Soon with advice from others she realized that the program had the potential to impact many more girls if it were configured as a 501(c)(3) nonprofit organization and could receive grants. Shortly afterwards, the organization and program began to rapidly grow. Professional staff were recruited and the organization's centralized structure began to change to decentralized councils. The program was standardized: Each council, its own 501(c)(3), taught the same material and used the Girls on the Run brand name, but they were able to seek their own ways of generating funds.
The basis for StreetWise MBA was an action research initiative begun in 2003 by the Boston University School of Management and Sociology Department. Its designers found that established small businesses in low-income communities, particularly women-or minority-owned, were often unable to grow to their full potential due to a lack of business development resources tailored to the needs of urban businesses. Results suggested an imbalance in which existing training programs, technical assistance, and financial resources prioritized startup businesses rather than existing and growing small businesses. Faculty created a 9-month course to address this imbalance. By 2007, the work had spun off as a new nonprofit, Interise, which reinvented the course into a scalable curriculum, the StreetWise MBA. In 2008, Interise was chosen to be the national instructional provider for the U.S. Small Business Administration's Emerging Leaders initiative, thus making StreetWise MBA available to small business owners not just in Massachusetts but in 10 cities across the nation. Today, Interise licenses its curriculum to 18 partner organizations and serves over 1500 small business owners each year in 75 communities throughout the U.S.

The social programs in the affiliate pathways have differing points of origin.
In contrast, the Wyman Center, after delivering the Teen Outreach Program (TOP) in St. Louis for 8 years, purchased the TOP copyright from a longtime collaborator, Cornerstone Consulting, in 2005. They developed an expansion plan with the goal of reaching 10 new partners a year. Within just a couple of months, however, they were faced with exponential demand as the U. S. Office of Adolescent Health recognized it as a model program. "Our phones started to explode with this new recognition," said Claire Wyneken, president and CEO of the Wyman Center. "At the same time the phones are blowing up, we were building our staff and processes. Conceptually we had a plan. Operationally we had to immediately implement. We were building the ship as it set sail. And it's been a fascinating journey since."

Strategic Choices: Partnerships
The staffing of affiliate pathway social programs includes a national-level staff and many paid and volunteer staff at the local level. Lead organizations tend to be nonprofits, but that is not always the case. Supporting partners and implementing partners are essential to scale up efforts. Distribution partners are not present. Lead partners typically seek out core partners, but lead partners are sometimes sought out by implementing partners. Sometimes they already knew their partners from previous encounters, and a couple met through a third party.

LEAD PARTNERS
The lead partner organizations we studied were mostly nonprofits other than colleges and universities. These organizations vary in complexity. Well-established nonprofits such as Citizen Schools, Jumpstart and KaBoom! had income of more than $20 million in fiscal year 2015 and employed more than 100 staff each. But these organizations started off small and lean. "Initially, my predecessor didn't foresee a national program," said Steven M. Rothstein, CEO of Citizen Schools, a nonprofit that partners with high-need urban middle schools and school districts to narrow achievement and opportunity gaps by expanding the learning day through the late afternoon. "It began just as an afterschool program in Boston," said Rothstein. Strong relationships with school districts and preliminary outcome data attracted the attention of program officers in private foundations, which then led to contacts with and the involvement of corporations that offered apprenticeships and the organization and program expanded.
Other organizations are less complex, employing fewer people and having leaner staffing models. Girls on the Run, founded in 2000, had $4 million in income in FY 2015 and 29 employees. Moneythink had income of around $1 million and 10 employees. The smallest nonprofit was Boot Camp for Dads, operated by the New Father's Foundation. This foundation is led by five family members and in 2012 reported revenue of about $50,000 (though the organization has begun Boot Camp for Moms, and is going online with program content for mobile phones). A couple of organizations are for-profit, such as The Incredible Years.
A few programs did have colleges or universities as lead partners. For example, Reclaiming Futures helps young people in trouble with drugs, alcohol, and crime. The initiative began in 2001, with $21 million from RWJF for 10 pilot sites to create a six-step model that promotes new standards of care and opportunities in juvenile justice by restructuring how agencies and organizations work together. This work continues and is housed in the Regional Research Institute for Human Services of the School of Social Work at Portland State University. So a unit such as a center or institute at a university leads the program within the context of the educational institution. If and when the program grows too large, it may spin off from the university. Nurse-Family Partnership, The Incredible Years, LifeSkills Training, and StreetWise MBA all began as research projects based in universities and led by academic researchers.
So lead partner organizations typically start off as small, dedicated nonprofits or research projects that reflect the vision for change of their entrepreneurial founders. They grow as organizations insofar as a leader is able and desires to attract programmatic support and successfully scale up his or her program. This then can lead to a demand for training, fidelity checks, and guidance about adaptation.

SUPPORTING PARTNERS
For social programs we examined that follow affiliate pathways to scale up, supporting partner organizations are essential. None of these lead partners provide all of the required strategic planning, funding for scale up, infrastructure such as trainers and workshops and evaluation expertise themselves. Even after one accounts for the local affiliates that will serve as implementing partners for a social program, the national and regional teams still need support.
Nearly all of the leaders we interviewed identified core partners who provided either funding or consultative support. Funders were especially prevalent supporting partners and some leaders mentioned more than one foundation as a key partner. Funders were often private foundations such as The Wallace Foundation, The Edna McConnell Clark Foundation, and the Citi Foundation. Federal agencies that provided funding were sometimes identified as core partners.
About half of the program leaders identified core support partners that provide particular expertise, such as organizational development, communication and evaluation consulting and services. "The reality of funders demanding greater emphasis on measurable outcomes combined with a growing need by schools for additional help, led us to partner with Maryville University in St. Louis," said Marcia Kerz, President and CEO with The OASIS Institute. The Wyman Center worked closely with Cornerstone, a consulting group that had operated the Teen Outreach Program (TOP), and also several researchers, including some at Mathematica Policy Research.

Funders were especially prevalent supporting partners.
Corporations that provide volunteers for programs are sometimes mentioned as core partners, such as with Home Depot for KaBOOM! and Fidelity Investments for Citizen Schools.
The cast of supporting partners that assist a social program is dynamic. Over time, partners come and go as their own organizational priorities and interests change. Communities That Care, an approach for using digital tools to provide prevention science to youth for the reduction of violence, substance abuse and delinquency, initially had a key scale up partner in the U.S. National Institute on Drug Abuse, a role that then transitioned to the health book publisher Channing Bete and then to the U.S. Substance Abuse and Mental Health Services Administration (SAMHSA).

IMPLEMENTING PARTNERS
Partnerships we looked at that pursue scale up through affiliate pathways often do so by seeking out potential organizational affiliates, training staff of those that are interested, and then supporting program delivery by those implementing partner organizations. Not surprisingly, more than three-quarters of the program leaders we interviewed identified implementing partners as core partners. Similar to the branching pathway, specific implementers were seldom mentioned and reference was to a generic type of implementing partner such as schools or community groups. For example, Gateway to College has prioritized "two-year colleges," Citizen Schools works with "middle schools," and KaBOOM! identified "neighborhood groups that want to build urban playgrounds" as core partners.
Although lead partners in branching and affiliate pathways identified similar core implementing partners (e.g., schools, colleges), there is a difference. In an affiliate pathway contracts are established with specific implementing organizations in each region or location to offer the program. Somewhere between the lead partner and these generic implementation partners are "middle men" that purchase or license or who have an agreement with the lead partner to offer the program. These "middle men" may not be individually identified, but they are present in these partnership constellations and they may have to prove that they can implement the program successfully.
Federal agencies were sometimes identified as implementing partners. LifeSkills Training (LST) is a drug and behavioral problem prevention classroom curriculum for middle school students that teaches personal self-management skills, social skills, and resistance skills. Implementers are teachers, counselors and social workers who teach intermittent lessons over three years. SAMHSA was identified as an implementing partner for LST though implementation support is indirect. SAMHSA oversees the National Registry of Evidence-based Programs and Practices (NREPP) that awarded LST top ratings for the quality of materials, training and support structure. This recognition facilitated dissemination of LST and encouraged implementation across communities.
While most social programs in our study that take affiliate pathways to scale up target change in individual behavior, like LST, some programs target change in the delivery systems that serve those individuals. Reclaiming Futures operates in the same problem domain as does LST-substance abuse, violence-but by convening institutional stakeholders in communities including Chicago, Santa Cruz and Seattle, in finding better ways to integrate and streamline and improve the operation of the justice system for youth. Judges, probation officers, substance abuse treatment professionals, guidance counselors and social workers convene to change organizational, city, county and sometimes state policies. For Reclaiming Futures, the core partner is the implementing partner, too, which is a community coalition that convenes to address systemic problems and solutions.

HOW KEY PARTNERS MET
When we asked leaders of these social programs how they met their core scale up partner organizations, about half were described as meeting through a proactive search where the lead partner looked for a partner. Fairly common was a lead partner proactively looking for funding that led to a core support partnership with a foundation. Jumpstart applied to the Corporation for National and Community Service to gain access to AmeriCorps members. Lead partners also went in search of implementation partners. StreetWise MBA leaders at Interise found partners by going on speaking tours and presenting at conferences, workshops and panels. Now, included in Interise's three-year growth plan is a data-driven analysis of target states and "partnership verticals" (anchor institutions, government agencies, business associations and economic development organizations). Interise then decides the best way to introduce StreetWise MBA to these targets. Through its membership in a number of related professional organizations, Interise leadership continues to cultivate potential partners, a process that continues through visits and meetings with them in their own locales.
Somewhat less frequently, core partner organizations sought out lead partners, had previous ties, or meet through a third party. The Incredible Years founder, Carolyn Webster-Stratton, said that partners regularly contact her organization "because of research publications, the SAM-HSA Clearinghouse, etc. Sometimes counties contact us because they've just chosen us off an approved list of evidence-based practices." Girls on the Run identified New Balance as a key partner. New Balance leaders had read about them in Runners World and called and said, "Let us help you." At The Campus Kitchens Project, their partnership with the Sodexo Foundation stems from a previous relationship of the foundation with CKP's parent organization DC Central Kitchen. The National Association for Urban Debate Leagues partnered with Citi Foundation because a member of the League board knew a treasurer at Citi Foundation who arranged a meeting. Gateway to College made a connection to the Bill & Melinda Gates Foundation via Jobs for the Future which knew about Gates' interest in helping struggling high school students get to college. This is social capital at work. Leaders we interviewed did not describe meeting core partners through happenstance.
Reinvention is common and lead partners engage with implementers though decision making is still largely with the lead organization and core partners. In an affiliate approach, a lead partner trains, advises, may rate and review and correct, and may analyze differential achievements across locations and try to understand why those results are occurring so that staff at modest performing sites can learn from one's best implementers. But in working with affiliates, one thing a lead partner does not do is implement a program. With an affiliate approach, to a degree, implementation is out of the lead partner's control.

REINVENTION
We observed many minor changes to these programs as they had been taken to scale. But even minor changes to a program can affect its chances to reach and benefit more people.
Several programs of study reduced costs by changing from a paid staffing model to the use of volunteers for implementation. For example, Jumpstart teamed up with AmeriCorps and various universities to gain volunteers and also teamed up with Head Start so they could share space for implementation.
Training shifted from face-to-face trainings at implementers' worksites, or face-to-face group-based trainings that many implementers could attend, to online training, which also reduces costs. Clinical Chapters and Communities That Care both moved their training online so they could reach more implementers at reduced cost, and also so they could reinforce the core components of their programs. Wyman's Teen Outreach Program included online elements to training and technical assistance so that it could become a basis for a learning collaborative as well as to reduce costs.
Perhaps no affiliate pathway program reinvented itself more radically than KaBOOM! when its leadership decided to make all of their planning materials available online for anyone to access, adopting a do-it-yourself dissemination approach to scale up. The result was that in 2009 KaBOOM! saw the construction of 1,600 playgrounds-almost the number of KaBOOM! playgrounds built over its first 14 years combined. KaBOOM! staff members are still available to assist in planning and construction, but since freely offering its materials online, KaBOOM!-influenced playgrounds outnumber KaBOOM!-built playgrounds 10 to 1.
A couple of programs expanded their audience. Year Up was reinvented in order to make the program part of a community college associates degree, and to reduce its costs. Urban Debate expanded into middle schools in Atlanta, Baltimore, Chicago, Minnesota and New Jersey where they already had high schools participating.
With an affiliate approach, to a degree, implementation is out of the lead partner's control.

More so than branching pathways to scale, affiliate routes imply partnering.
To what extent did partners that pursued scale up via an affiliate pathway work collaboratively on reinventing their program prior to scale up?

INTERACTIVE REINVENTION
When reinvention occurred in this pathway, the focus was on the program with implementing partners typically providing feedback that the lead partner then used to modify the programs. For example, Citizen Schools was an afterschool program for 15 years with an optional part of the day for students in need. In planning for scale up, leaders of the Boston area program reasoned that if the program could be required as a part of a longer learning day, then the program could have greater effect and its sustainability would be strengthened. Leaders from Citizen Schools involved implementing schools in this discussion, and they agreed. Citizen Schools became a part of their curriculum.
Communities that Care was developed by J. David Hawkins and Richard F. Catalano, both faculty at the University of Washington. Ownership of the program was transferred to Channing Bete, a publishing company from 2001-2005 and then to the federal government's Substance Abuse and Mental Health Services Administration (SAMHSA) from 2005-2014. The program is now back at the University of Washington in the Center for Communities that Care. "Communities were implementing the program without technical assistance and training, at least not from us, and there was little measurement of fidelity, adaptation or outcomes," said David Hawkins. "So when the program came back to the university, we knew we had to make some changes. We refined the implementation materials so that the curriculum is available online. The training system is web based with all key points on 2-to 5-minute videos with a diversity of people providing the theoretical foundation and testimonials." These changes were informed by what the program staff were hearing and observing from implementers and also from core partners such as SAMHSA. These changes are part of a new scaling strategy.
Core supporting partners are also involved in reinvention. The Center for Financial Services Innovation is a core support partner for Moneythink. "They have helped us with many technical issues, with structuring Moneythink and helping us to rethink its innovative features, and in building and carrying out plans for measuring the impact of our work," said founder of Moneythink, Ted Gonder. Program changes have helped Moneythink to scale their program.

Changes were informed by what the program staff were hearing and observing from implementers and also from core partners.
StreetWise Steps is the licensed version of StreetWise MBA, a nine-month certificate program that provides small business education to small business owners in inner cities. Organizations licensing StreetWise Steps include the New York City Department of Small Business Services, the University of Pittsburgh's Institute of Entrepreneurial Excellence, the Greater Dallas Hispanic Chamber of Commerce, and the City of Portland. While licensees do sign a license agreement "with a nice big fine" if they violate the intellectual property clause, said CEO J. Jean Horstman, "we are participative. More and more people recognize that we're different from most license holders. We've built a network with a large number of nodes, and encourage the sharing of competencies across the network. There is a lot of time spent informally building relationships so that licensees in cities can best help their small business owners." So the learning network that StreetWise Steps has put into place evolves as new licensees join and as new information about competencies is shared in the network.
StreetWise MBA encourages its licensees to apply their own organizational name to the Street-Wise program, effectively rebranding it locality by locality. "This makes fidelity to our model highly attractive to them," said CEO Horstman. "Because we subordinate our brand, and support theirs locally, the recognition for a highly successful program as evidenced by impact data which are influenced by program fidelity accrues all to them. By eschewing our brand, and building theirs, they are keen to deliver StreetWise MBA to our standards."

ADAPTATION
For social programs that pursue scale up through affiliates, fidelity is a serious matter. But lead and supporting partners deal with adaptation by implementers and threats to fidelity to differing degrees. Part of this difference can be explained by whether program developers are researchers or practitioners. And programs can be more or less amenable to adaptation in the field. Here we assess what the 20 affiliate pathway lead partners did to ensure that the delivered quality of their programs did not suffer as they reached more and more sites or more individuals at each site. We assess lead partners in terms of how much fidelity they sought from implementers, since some social programs work best when there is a high degree of fidelity and other programs work best when adaptations to local conditions are encouraged, and also in terms of the extent to which lead partners monitor what it is that implementers are doing in the field.

PRESCRIPTIVE FIDELITY
About half of the programs following affiliate pathways to scale up sought high fidelity and created systems to ensure its achievement. One such program is Nurse-Family Partnership (NFP), begun by Professor David Olds, Director of the Prevention Research Center the University of Colorado.
Olds believed that if low-income women pregnant with their first child were routinely visited by a nurse in the home during the pregnancy and until the child was age two, both mother and child would benefit. Olds oversaw multiple privately and federally-funded randomized, controlled trials of NFP that provided evidence of the program's effectiveness, and triggering a decision to scale up NFP. Demand quickly outstripped the capacity of the university department to respond so Olds established a spin-off NFP National Service Office in 2003. Now about 32,000 families are served by the program annually (253,230 ever served) in 42 locations, with the paid involvement of about 1,850 nurses.
An affiliate pathway works for NFP because it allows for the National Service Office to educate and support nurses in the delivery of the model and monitor the delivery of the program. This ensures fidelity to NFP's 18 program model elements, while also encouraging nurses and implementing agencies to make suggestions to overcome impediments to effective program implementation. These suggestions are then jointly considered by the National Service Office and the Prevention Research Center staff for rolling out small-scale tests of whether adaptations can improve upon the basic program, or not. In these two ways the program both carefully retains effectiveness and potentially improves over time.
Another example of an academic program developer who has learned a great deal about fidelity and its importance for her program's successful deployment is Carolyn Webster-Stratton, developer of The Incredible Years Programs. Over the years, she kept track of barriers to effective implementation she encountered such as inadequate funding for program managers in affiliate agencies, lack of logistical support, and poor clinician support for training and consultation. She then designed ways to help agencies overcome these barriers. This included the development of an organizational readiness assessment, guidance for assuring standardized quality training for group facilitators led by certified trainers, principles-based training workshops with release time for clinicians, ongoing live and videotape feedback about clinical performance, development of peer support networks within implementing agencies, and adherence to dosage, order of content presented, and clear session protocols. Group facilitators were encouraged to complete program certification in order to assure fidelity program delivery. Another early childhood program, Jumpstart, had a similar emphasis and set of controls to ensure fidelity, partly through regionally deployed quality control experts.

RESPONSIVE ADAPTATION
Eight of the 20 programs scaling up via affiliate pathways are well-characterized by responsive adaptation. They allowed for or encouraged adaptation, and had created means of regularly monitoring implementation to learn what was going on.
The Campus Kitchens Project (CKP) affiliates on campuses are given latitude to apply a standard program model differently in each community according to the opportunities that student leaders themselves identify and pursue as long as it is in response to a needs analysis. This degree of customization means that a Campus Kitchen may emphasize one or more of four model components in a service activity, and implement that component in a unique way as long as all four model components are addressed.
Like NFP, the vast majority of CKP workers are in the field helping clients. Both programs stress the principles of their approach and require implementers to carry out activities that reflect each principle. But CKP is much more flexible in encouraging, not just allowing, its implementers to adapt to local conditions. NFP stresses fidelity while at the same time inviting suggestions from nurses about how to improve the program. CKP stresses food recovery, meal preparation, transportation and student engagement while celebrating creative approaches of its student teams. The different orientation toward fidelity between these two cases is partly explained by their genesis: NFP was an academic intervention by university researchers; CKP was an extension of a local social enterprise providing meals and culinary job training.
The Teen Outreach Program, headquartered at the Wyman Center, balances adaptation and fidelity. "People come to TOP and they are surprised," said CEO Claire Wyneken. "They're used to precision delivery-do this in this way with no variation. But what the research told us was that a youth development framework is different. That service learning, youth voice and choice inside the program and strong youth development are all critical to the outcome. There can be variance but it should be planful. So this surprises others and challenges us. Here is what you must do and here is where you can be flexible." At the same time, Wyman is monitoring implementation. "We created structures so that data, certification and training and TA support fidelity. We have a monthly call with partners and we really try to build genuine relationships so they know our role isn't to hit them with a fidelity stick but to help them succeed." Still, Wyneken noted that earlier efforts to achieve scale did so with less rigor toward fidelity. Wyman has adjusted this balance so that current TOP scaling efforts have a stronger focus on monitoring implementation.

EXPECTANT FIDELITY
Expectant fidelity means that implementing partners are discouraged from adapting a program but little if any monitoring is occurring to know what those implementers are doing. Just two of the 20 affiliate pathway programs took this approach. This situation might characterize lead and supporting partners that are too busy to monitor implementation, or lead partners that do not believe that adaptation is likely, or situations when implementers are so expert that little if any monitoring is necessary. Clinical Chapters operates in the latter way. Licensed therapists provide professional evidence-supported psychotherapy to foster youth in one-on-one clinical sessions. Fidelity to professional and ethical standards is assured so monitoring is not emphasized.
Affiliate Pathways "There can be variance but it should be planful. So this surprises others and challenges us."

INDEPENDENT ADAPTATION
When it moved to free online access to its materials for any neighborhood interested in building a playground, KaBOOM! placed little emphasis on either fidelity or monitoring. Moneythink also emphasized the potential of adaptation without much monitoring of services delivered and implemented. "We realized we couldn't scale up our chapter model forever without incurring huge costs," said CEO Ted Gonder. "This realization moved us to think about how to change our basic service model. As a result, we unbundled our software and curriculum and impact tracking methods so we can provide them to other organizations that want to do financial education but may only need some of the components we have to offer, not all of them." This decision at Moneythink reflects both its orientation to the importance of flexibility in implementation, and also demonstrates considerable leadership-led program reinvention.

Distribution Network Pathways
A distribution network pathway involves a lead partner working with a distribution partner that has an existing network of implementing partners. This pathway can resemble a supply chain: One organization supplies the social program, a distribution partner delivers it through implementing partners, and support partners facilitate development and implementation.
Programs we studied in this pathway moved more quickly to scale up than did programs in other pathways. A couple of these programs were designed to scale up from the start. Universities and/ or research-based programs are prominent in this set of programs. Lead partners work closely with distribution partners, and supporting partner organizations play vital roles. Implementing partner organizations are not mentioned as often by lead partners as being core to scale up efforts. Reinvention is common and often done with the participation of core partner organizations, not just by lead partners. We also find examples of integrative reinvention, where the organizations as well as their programs change. Extensive guidance is provided to implementers so that field-based adaptations to programs will enhance rather than detract from the achievement of positive results. Implementation activity is not always monitored.

SCALING BY DISTRIBUTION NETWORKS: POWER SCHOLARS ACADEMY
Power Scholars Academy (PSA) is a six-week summer program designed to reduce the loss in student math and reading learning that commonly occurs during summer months while school is out, and to foster physical and social-emotional growth. PSA is a partnership between the YMCA of the USA (Y-USA) and the nonprofit Building Educated Leaders for Life (BELL). PSA was created from BELL's evidence-based summer learning model, which BELL and Y-USA reinvented so that the program could reach and benefit more children throughout the country. The program was originally called Y-BELL, but the partners changed the name to Power Scholars Academy to help create a "neutral space" in which the two organizations could work as equal partners. Y-USA is the national organization that ties together 2700 local YMCAs. As the distribution partner, Y-USA provides reach into disadvantaged communities and taps into its local Y's existing relationships with schools that become the implementing sites for PSA.
For an entire year, BELL and Y-USA negotiated and explored the creation of the partnership by which they now operate. This has helped to produce a trusting, successful bond between what have become two highly interdependent entities. "Team work makes the dream work," said Y-USA's Mary King. Staff from both organizations commented that people who come to meetings sometimes get confused about which people are from BELL and which are from Y-USA. This doesn't appear to be the result of any lack of clarity in roles, but rather is an indication of how close BELL, the lead partner, and Y-USA, the distribution partner, have become, due both to their shared mission and the successes that have come through working together.
PSA has supporting partners, too. Rebecca Kelley, then with Y-USA, and BELL's Lauren Gilbert, said Bridgespan Consulting helped "us to figure out if we wanted to partner". The Wallace Foundation was an active sponsor of PSA in its early stages and provided funding for assessment, evaluation and consulting expertise. The Wallace Foundation's investment in PSA was an outgrowth of and complement to other investments in program developers to build their infrastructure and evidence base to further scale programs.
While a high priority has been placed on PSA fidelity, in practice local implementers have made significant adaptations so that the program can succeed at the local level. BELL staff describe a difference between program elements that are "fixed" and those that are "fixed-flexible," with the latter being open to adaptation by local sites with approval from BELL and Y-USA. This is important particularly for staff who may have had some experience with prior YMCA programs when "freely adapting to make it work" was the program philosophy. At the same time, fidelity to dosage is reinforced in the program's communication with parents by stressing the importance of attendance.
Working together has changed the culture and operations of both organizations. Through the partnership, BELL's leaders have recognized the value they bring to a partnership through codifying and packaging information and program details for external partners. For Y-USA, their engagement with BELL has helped them to (1) work more systematically to assess their interest and capacity to pilot and ultimately scale and sustain evidence based programs, (2) better align internal technical assistance across the three areas of Y-USA focus (healthy living, social responsibility, and youth development), (3) explore the regional and state infrastructure supporting YMCA program delivery and advocacy, and (4)  This case demonstrates interesting aspects of how a program developer can work with a distribution network organization. These two organizations saw eye to eye on the basis of realizing a shared purpose. They established trust. This high degree of organizational bonding meant that questions about reinvention and adaptation could be successfully addressed. The reputation of the Y-USA national office with the local Ys was critical for convincing a select few YMCAs to demo the program as a field test. This was accomplished, in part, by institutionalizing annual summer visits where YMCA executive directors visit other associations that have launched PSA.

Program Characteristics
In this chapter we look at 13 cases that used distribution network pathways. These programs are listed in Table 4.1. Below, we review evidence of effectiveness and spread of the programs that met inclusion criteria. We also provide an overview of the focus, time to scale and funding sources found across this set of social programs.

EVIDENCE OF EFFECTIVENESS
Of the 13 social programs using distribution network pathways, many had been assessed in one or more randomized controlled trials. For example, the YMCA's Diabetes Prevention Program (YMCA's DPP) is based on a lifestyle intervention originally developed by the National Institutes of Health and the Centers for Disease Control and Prevention and shown through randomized controlled trials to reduce weight and delay or prevent the onset of type 2 diabetes. The effectiveness of the modified program was evaluated in a matched-pair, group randomized intervention trial involving two YMCA facilities in greater Indianapolis. The modified program was found to be effective and was considerably less costly than the original DPP intervention.
BAM-Sports Edition is another program that had rigorous assessment of effectiveness. BAM-Sports Edition focuses on developing skills related to emotional regulation, control of stress response, improved social-information processing, interpersonal problem solving, goal setting and attainment, and personal integrity. Faculty from the University of Chicago conducted a randomized field experiment in the Chicago Public Schools by assigning 2,740 disadvantaged males in grades 7-10 to BAM-Sports Edition or to a control group. Participation in BAM-Sports Edition increased the number of days students were present and grade point averages, effects that may translate to a 10 to 23 percent increase in graduation rates relative to the control group. The intervention was also shown to reduce violent-crime arrests during the program year.
Other programs had evidence of effectiveness obtained from a quantitative non-experimental study, such as cohort studies, pre/post measurement and/or cross sectional surveys. For example, Money Matters conducted pre/post measures of students attending the program to identify changes in their financial behaviors, such as saving money, opening savings and checking accounts and developing budgets. No program relied solely on anecdote as evidence of effectiveness though many complemented research evidence with narratives. We also found a theory of change embedded in each program. Perhaps no program in our study experienced scale up as dramatically as Housing First, a program offering permanent, affordable housing as quickly as possible for individuals and families experiencing homelessness in order to transform their lives. Once housing is secured, supportive services and connections are extended to program participants. The program has been spectacularly successful. After its debut in New York City in 1992 as Pathways to Housing, affiliates were gradually established, led by CEO Sam Tsemberis, who had a faculty appointment at Columbia University. With strong evidence of effectiveness came policy adoption by the U.S. Department of Housing and Urban Development, and the Veteran's Administration to offer housing through the program to all returning veterans. "There are so many great programs out there but few of them do the research," said Tsemberis. "About six years ago, the VA realized that it had about 70,000 homeless vets and they had to do something. We had the strongest evidence." After conducting a large Housing First randomized controlled trial, in 2014 the Canadian government was the first country to adopt the model as their national policy to end homelessness. By that time, many communities across the U.S. had followed suit. "It just took off," said Tsemberis. "The health minister of Quebec talked to the minister of health in France. Then Denmark followed and Holland followed. So the trajectory has radically changed."

PROGRAM FOCUS
The topical focus of programs ranged from public health, literacy and numeracy, character building, to academic achievement. More generally, these programs focus on education, as did those programs in branching and affiliate pathways. Educational programs targeted pre-K children; others such as Responsive Advocacy for Life and Learning in Youth (RALLY) focused on elementary school students or like Diplomas Now, on middle and high school students. Adults were at least partly the focus of several of these 13 social programs, often with a focus on their health, including Active Living Every Day (ALED), CATCH Health Habits, the YMCA's DPP, Housing First, and Safe Surgery.

TIME TO SCALE UP
As shown in Figure 4.1., only three of the 13 programs had been in operation for more than 10 years prior to scale up, and several started scaling up within three years from when they began. On average, scale up started within 7-8 years for the distribution network pathway programs.
Housing First began scaling up around 10 years after it was launched in 1992. In comparison, Climate Matters, a program developed by scientists at George Mason University and Climate Central, began in 2010 with a single pilot program led by one weathercaster. By 2012, 10 weathercasters had implemented the program. By the end of 2014, nearly 200 had done so.

Distribution Network Pathways
With strong evidence of effectiveness came policy adoption.

SOURCES OF FINANCIAL SUPPORT
While all 13 programs rely primarily on external funds, they vary in the sources of these funds.
Foundations were sources of financial support for all 13 both prior to and during scale up efforts. Federal funding was prevalent among those programs initiated at universities or by research organizations, again both prior to and during scale up efforts. Nonprofits that initiated programs were sometimes supported by individual contributors. Fee for services was rarely a source of funding prior to or during scale up. Funding does appear to become more complex in terms of number and type of sources as the social program scales up.

PROGRAM ORIGINS
As described later in this chapter, many of the programs originated as research projects either in universities or in nonprofits with a research mission. For example, Ed Maibach, Director of the Center for Climate Change Communication at George Mason University, wrote a research proposal to explore the development of Climate Matters that the National Science Foundation (NSF) funded. The results of this research suggested that the prospects for encouraging and enabling TV weathercasters to embrace a new role, that of local climate educator, were quite good. These findings led to a second NSF grant involving George Mason University, nongovernmental organizations, professional societies and government agencies, to systematically plan the development and scale-up of Climate Matters. Maibach assembled a diverse partnership of organizations with complementary skills including content experts for the science of climate change (Climate Central), social scientists who study how humans process information (at George Mason), and communication practitioners (via professional associations) who know how to get useful information into public dialogue. Climate Central, one of the core partners, has started to make Climate Matters "a cornerstone of their organization", said Maibach. "They are successfully raising significant amounts of funding from philanthropic sources to support their work."

Strategic Choices: Partnerships
Lead partners in distribution network pathways identified on average more core partners than did leaders associated with branching or affiliate pathways. We note that these lead partners often had a research mission. All lead partners identified a distribution partner, a role that is specific to distribution network pathways. Supporting partners were also frequently mentioned. Implementing partners were not called out frequently as being core to these efforts. Partners frequently knew each other prior to establishing these partnerships.

LEAD PARTNERS
Most of these 13 social programs have their basis in research. For many, the lead partner, the organization taking the program to scale, is associated with a university or a nonprofit with a research mission. The lead partner organizations for Diplomas Now, Climate Matters, RALLY and Safe Surgery are affiliated with universities. Active Living Every Day (ALED), a program to integrate physical activity into everyday living, was developed by behavioral scientists at the Cooper Institute in Dallas, Texas and funded repeatedly by the National Institutes of Health (NIH). AfterSchool KidzLit, a reading enrichment program, was designed by the Center for the Collaborative Classroom (CCC), formerly Development Studies Center whose mission is to conduct research and develop school-based and afterschool programs that help children achieve academically and develop socially. CCC believes that academic and social skills are equally important in learning.
Several lead partner organizations without a research mission are scaling programs that originated in research settings. The YMCA's DPP originated as a research project funded by NIH and then by CDC, with later piloting done by the YMCA of Greater Indianapolis. A similar trajectory occurred

Distribution Network Pathways
Many of the programs originated as research projects either in universities or in nonprofits with a research mission.
with CATCH Healthy Habits. In its initial form, CATCH was the largest school-based health promotion study ever conducted in the U.S. The OASIS Institute reinvented the program and rebranded it, and distributed it to OASIS Centers across the country.

DISTRIBUTION PARTNERS
Distribution partners were mentioned as being core to scale up efforts by each leader we interviewed in this pathway. The distribution partners in this study were sometimes national nonprofits like Y-USA, City Year, and Boys & Girls Clubs of America. These large distribution partners tie together many localities. For example, Boys & Girls Clubs of America serves youth in more than 4,000 communities. Some distribution partners are regional nonprofits such as the South Carolina Hospital Association. The South Carolina Hospital Association is made up of some 100 member hospitals and health systems in South Carolina. "The South Carolina Hospital Association has been our most important partner and a huge lab," said Bill Berry, Chief Medical Officer and Director, Safe Surgery Program. "They are helping us to test how to spread in a larger than one-off world." Safe Surgery may have such a close relationship with this hospital association because it is their first distribution partner.
The American Meteorological Association and the National Weather Association are distribution partners for Climate Matters. "The associations are invaluable partners," said Ed Maibach, George Mason University. "When considering how to scale up, we consciously included associations as explicit partners in our work. These organizations represent weathercasters and weathercasters trust these organizations." The researchers also reasoned that meteorologists would accord high credibility to messages about and invitations to try Climate Matters if the associations were involved.
Diplomas Now has two distribution partners. The lead partner, Talent Development Secondary, at Johns Hopkins University, wanted a way to scale up its school-based intervention for high schoolers. Bob Balfanz, of Talent Development Secondary, knew of complementary school interventions being offered by two large distribution partners, City Year and Communities in Schools. Each had their own network of cities and school district partnerships. By fusing their programs into Diplomas Now, the three organizations could work together in a larger number of schools that are low-performing. City Year is an AmeriCorps program whose 3,000 members are trained to work in schools that are underserved to help students get on track and stay on track to graduate. City Year has operations in 26 cities in the U.S. and emphasizes social emotional development, academic support, and afterschool and extended learning. Communities In Schools is the nation's largest dropout prevention organization with a presence in 2,400 public schools through nearly "They are helping us to test how to spread in a larger than one-off world." 200 local affiliates in 27 states and the District of Columbia. Its emphasis is on the most difficult students who need case management and individualized or small group counselling. The footprints of these two distribution partners determined where and to what extent Diplomas Now could scale since both organizations had local bases of operations in cities, with local directors, staff and existing relationships with school districts. The team considered the 13 cities where City Year had existing relationships with school districts overlapping with Communities In Schools, ultimately focusing on middle and high schools with high poverty rates that serve 57,000 students. Diplomas Now submitted a proposal to the U.S. Department of Education's Investing in Innovation (I3) Program with the objective of validating the combined Diplomas Now approach. A resulting $30 million award in 2010 for Diplomas Now from the I3 Program was the largest I3 validation award as of that time.
While most distribution partners were nonprofits, a couple of them are private organizations that provide publishing, products and services. For example, Nemours BrightStart!'s distribution partner is Kaplan Early Learning Company. Kaplan is a leading international provider of products and services that enhance children's learning. Nemours owns the intellectual property to Nemours BrightStart! but Kaplan produces the products and sells them.
What we did not find in our study were universities as distribution partners. This is not surprising. Universities face well-documented challenges in scaling innovations because the tasks required to do so exist outside of faculty role expectations and faculty reward structures.

SUPPORTING PARTNERS
Lead partners in this pathway almost always identify at least one core supporting partner, similar to the other two pathways. Foundations are especially prevalent as supporting partners. In several instances, a foundation provided funds and then stepped away. But in other instances a foundation was an active supporting partner described as providing consultation, suggesting unexplored market niches, making new business connections, and supporting evaluation or research about a program as it scaled.
A. J. Watson with Youth Guidance describes the Edna McConnell Clark Foundation (EMCF) as working fairly closely with Youth Guidance staff on planning for scale up of BAM-Sports Edition both in the Chicago area and nationally, and in helping Youth Guidance to refine its lines of communication with the University of Chicago Crime Lab, which was a funding and evaluation partner of BAM-Sports Edition. "EMCF has been the most important funder-partner. Their support has been transformative. They offered multi-year support to help us grow to scale. They have been involved in much more than writing a check; they are like a venture capital firm. They help us connect with experts for strategy support, financing and consultation around implementation science," said Watson. EMCF has also helped BAM-Sports Edition to establish a set of metrics to measure their achievements. As another example of an expanded role of foundations, the Robert Wood Johnson Foundation incorporated ALED into their Active for Life set of translational programs that led to expanding the ALED target population to include older adults.
Consultants were sometimes identified as key supporting partners. Bridgespan, a nonprofit consulting group, provided support to all three organizations affiliated with Diplomas Now, as well as other organizations. Consultancies were important partners for PSA. Bridgespan helped Y-USA to identify BELL as a program to work with as the Y-USA rolled out a number of programs as part of their achievement gap initiative. La Piana Consulting developed PSA's first business plan and the subsequent market research and revisions to the plan were supported by FSG, a consulting group. In reference to FSG, Rebecca Kelley, formerly with Y-USA and PSA, said, "They are on calls together and really participate and help us make this easier and more intuitive for local implementation."

IMPLEMENTING PARTNERS
Implementing partners were mentioned just a couple of times as core partners by leaders from these social programs. The focus in distribution network pathways is on distribution and supporting partners. When implementing partners were mentioned, it was because they played some kind of a special role, that is, a role beyond implementing the social program. Boston schools were mentioned by Gil Noam, Founder of RALLY, as a demonstration site. "RALLY still exists in Boston to provide what we call "test kitchens" for further experimentation with program models," said Noam.
Of course, without the implementing partner organizations, none of these social programs will spread. The emphasis of leaders on the network organizations that tie together the actual implementing partner organizations may be an indication of how efficient this type of pathway can be; a scale up lead partner works with the distribution organization who, in turn, does much of the actual work of interacting with the implementing sites.

HOW KEY PARTNERS MET
Partners in the distribution network pathway often had prior knowledge of each other. This knowledge was sometimes rather general. The Schwab Family Foundation knew about the expertise of the Boys & Girls Clubs of America when they were looking for a partner for Money Matters, but they were not well versed on what this large network could do for them. At other times prior knowledge was more specific, such as between Youth Guidance and World Sport Chicago in creating BAM-Sports Edition. This partnership came about, at least in part, because the head of World Sport Chicago was a member of the Also common were instances of a lead partner seeking out a scaling partner primarily because they were seeking funds from the future core partner. Ed Maibach at George Mason University applied to NSF for funding for what became Climate Matters, the Cooper Institute approached RWJF for funds for Active Living Every Day and staff at Youth Guidance responded to a call for proposals from the Crime Lab (University of Chicago) to develop BAM-Sports Edition.
Less common were partnerships formed through serendipity or intervention by third parties. But there are instances of these ways of meeting. When describing how Ariadne Labs came to partner with the South Carolina Hospital Association, Bill Berry said it was "entirely serendipitous." Atul Gawande was approached by the South Carolina Hospital Association to speak, and said he'd do it but then challenged them to implement Safe Surgery in their state.

Strategic Choices: Fidelity
Here we present our findings concerning both reinvention and adaptation of the 13 social programs scaling using a distribution network pathway.

REINVENTION
These 13 programs were nearly always reinvented prior to scaling up. Reinventions to a program may be time consuming and involve a degree of risk, but respondents indicated few reservations or concerns about the changes they made. Reinvention among these 13 social programs appears to flow from the partnerships, and is seen as a means to an end -to improve scalability and allow the program to have greater societal impact.
The most common form of reinvention we observed was in the delivery of programs which is expected as developers are coming to rely on their distribution partners to communicate and often times deliver the programs in this pathway. For example, PSA shifted the delivery of the original program to Y-USA; RALLY shifted to being delivered by City Year; and Nemours BrightStart! changed from being delivered by Nemours to Kaplan Early Learning.
In several cases, the purpose or focus of the program was modified. When the Center for the Collaborative Classroom (CCC) created AfterSchool KidzLit, the audience was children and the focus of

Distribution Network Pathways
Reinvention is seen as a means to an end -to improve scalability and allow the program to have greater societal impact. the program was integrating social development and cognitive skill development. After working for many years across the country, CCC leadership realized that how you teach is as important as what you teach. From this learning they began focusing on improving and supporting facilitation skills with the idea that their work would improve teacher practice as well as child learning. The difference in focus between BAM, the original program, and BAM-Sports Edition, the scale up program, is a sports component that occurs afterschool.
Reinvention of the audience occurred, too. The original CATCH program was designed as a physical activity program for school children and relied on teachers in schools to deliver the program. When OASIS decided to scale up CATCH via its members who are older adults, the program name was changed to CATCH Healthy Habits. This meant a corresponding change in the target audience to include older adults: Research has proven that older adults experience meaningful improvements in their own physical and mental health by helping young children.
Sometimes reinventions led to reduced costs. YMCA's DPP leadership modified the original program from one-to-one meetings in which a client met with a medical provider, to a group format led by nonmedical YMCA personnel that reduced costs from about $2700 to $300 per patient with comparable outcomes. OASIS' decision to have adult volunteers serve as instructors in place of teachers reduced its cost of delivery as well.

Observations across the programs in this pathway suggest that partners tended to engage in integrative or interactive reinvention with just a couple of instances of contractual reinvention.
Core partners working very closely together-integration-appears most prevalent, but not all lead partners need or perhaps want this type of close collaboration.

INTEGRATIVE REINVENTION
The example at the beginning of this chapter, about PSA, exemplifies integrative reinvention. BELL and Y-USA jointly reinvented BELL's summer learning loss prevention program into PSA. Both of their operations have changed as a result. Integration of this type also occurred with CATCH Health Habits.
For some programs, the lead and distribution partners jointly reinvented the program by fusing their organizations. AfterSchool KidzLit is one of several programs developed by the Center for the Collaborative Classroom. They have been working in schools since 1980; their work in afterschool programming began in 1995. By partnering with national afterschool providers and CBO's across the country they learned what was important and needed. By field-testing and piloting with these partners, they were able to develop materials that were clear and easy to use for afterschool leaders. It was from these partnerships that KidzLit leaders realized that their niche was afterschool, not "more school" (the program and its content had to look and feel different than school). The collaborative experience changed the CCC.

A slightly different type of integrative reinvention took place between RALLY and City Year. City
Year developed a Whole School-Whole Child program that applied the RALLY model, although City Year does not use the RALLY name. Rather, the developer of RALLY, Gil Noam, and staff at Harvard's Program in Education, Afterschool & Resiliency provided comprehensive training for City Year school-based staff who manage the AmeriCorps members who then serve as tutors and mentors in public schools, many located in Boston. RALLY, then, has been integrated into City Year, an organization that has the capacity to impact more students because of their expansive network.

INTERACTIVE REINVENTION
Often, the relationship between the lead partner and a distribution partner involves close collaboration and shared goals as they work together to reinvent a program, but these changes don't require commensurate changes in the organizations. What's distinctive about interactive reinvention is the balance between shared activities and independence.
This balance is well represented by Diplomas Now. A key to the effective functioning of the Diplomas Now national partnership was an agreement initially that Talent Development Secondary, Communities in Schools and City Year could keep their core business with only modest change to accommodate the new partnership. "We considered creating a new organization-a fourth organization-but we settled on something like an inter-agency team in the military that bridges the Army, Navy and Air Force," said Bob Balfanz, Co-Founder. "We each brought a unique piece to the new model so there wasn't conflict or turf battles. Each of us could do 85 percent of our own model as long as we collaborated on 15 percent as Diplomas Now." The partnership has shared practices and structures that have evolved over time. This includes a national leadership structure with weekly collaboration conference calls and periodic in-person retreats to focus on in-depth analysis of the current state of the model and strategic planning around sustainability, growth, and model enhancements over time.
Youth Guidance and World Sport Chicago worked together to develop a response to a competition from the University of Chicago Crime Lab for promising ideas about how to reduce crime and violence among Chicago youth. Their response was BAM-Sports Edition which merged their efforts but required little change in how each organization operates.
The Schwab Family Foundation and Boys & Girls Clubs of America jointly developed Money Matters and George Mason and Climate Central jointly developed Climate Matters. Both programs were designed for scale up from their start which is unusual among the programs we studied. This type of "co-invention" -where lead and distribution partners design programs with scale in mind-is deserving of additional focus.

Distribution Network Pathways
"We each brought a unique piece to the new model so there wasn't conflict or turf battles." Instances of a lead partner determining how or if modifications will occur by contracting reinvention also occur among these 13 partnerships. Contractual reinvention occurred when publishing companies were the distribution partners and their role was to produce materials and to then sell the materials. For example, Nemours contracted with Kaplan Early Learning to publish and sell Nemours BrightStart! Laura Bailet, Executive Director of Nemours BrightStart!, said that Nemours owns the intellectual property-the content of the program-and Kaplan brings the packaging and marketing expertise. In discussing the contract between The Cooper Institute and Human Kinetics to scale up ALED, Michele Guerra, who had been Active Living Partners Director at Human Kinetics, said that fidelity was very important to both The Cooper Institute and to Human Kinetics. Guerra did not see the contract as unusual because "academics sign contracts all the time with publishers. What was different was this was a larger than normal effort. It centrally involved a foundation [RWJF] that had reservations about a commercial publisher also providing facilitator training. The effort was considered unusual by Cooper, but not so much at Human Kinetics because we'd already been doing this sort of thing."

ADAPTATION
Most of the lead organizations we studied that are scaling a program via a distribution network pathway anticipate adaptation. They provide guidance to limit or direct reinvention in the form of scripted curricula, training manuals, websites, webinars, coaching and evaluation support. Still, this guidance is largely voluntary and the programs may lack fidelity in some locations, or exceed documented impacts from early pre-scale up tests.

INDEPENDENT ADAPTATION
Many lead partners anticipate, and some encourage, adaptation so that the program will fit well with local contexts. Climate Matters, for example, supports tailoring by each television meteorologist by providing them localized data and analysis that shows the ways that the climate is changing in their viewership area. AfterSchool KidzLit is guided by an interest in having implementers take ownership of the program. Neither Climate Matters nor AfterSchool KidzLit determine how these programs will be implemented. While both engage in ongoing evaluation, individual users are not required to report on outcomes nor is training or certification required for use.
Nemours BrightStart!'s curriculum provides a range of activities and approaches that allows for tailoring the curriculum to the needs and interests of a community of learners. It also provides a comprehensive set of tools to guide the adaptation of their curriculum. Laura Bailet encourages evaluation as a fidelity check Many lead partners anticipate, and some encourage, adaptation so that the program will fit well with local contexts.
and "offers to do a year of data collection, and help the sites with data-based decision making." But sites are not obligated to engage in evaluation. Nemours's distribution partner, Kaplan Early Learning, also offers free online implementation and training webinars and you can purchase tailored training. Like evaluation, however, training is not required. Schools, parents and others who purchase the Nemours BrightStart! materials from Kaplan Early Learning are free to choose how they use them. Bailet recognizes this may facilitate a variety of implementation scenarios. "We really encourage measurement at the site level to help instructors learn and improve. But what we've found is that if you do half this program, it still can work. We want everyone to be with us for a live training, but it's not always possible. The instruction and way that the program gets used may be "scary", but it's still far better than what those kids were getting prior," she said.
A belief in adaptation is central to scaling up of the RALLY afterschool program. Gil Noam, RALLY's director, talks and writes about the importance of local context, charismatic leaders, motivated staff, a mission targeted to local needs, family involvement, and youth engagement in making a program a successful. Noam argues that rather than scaling up a standardized model, programs should be customized to address the needs of local youth. As Noam explains, "People do not want another Starbucks-they want the old mom-and-pop shops." As part of this customization, Noam suggests using data-driven programming where an evaluation feedback loop helps guide management decisions, aids in the construction of activities that engage youth in learning, and improves program services and supports. He supports flexibility but encourages sites to conduct evaluation to ensure and improve the quality of the program. City Year is a distribution partner for the core components of RALLY. While tailoring continues to be a focus and Noam and his team train the City Year AmeriCorps volunteers, RALLY leadership will not be evaluating or collecting data directly from City Year sites. Neither is City Year evaluating the scale up effort.

RESPONSIVE ADAPTATION
A smaller number of lead partners encourage adaptive behavior by implementing staff but they monitor implementation or require specific training for implementation as a means to control or guide adaptation.
Sometimes, monitoring is achieved through required training prior to implementation. ALED allows the flexibility of being offered independently or in conjunction with existing community-based physical activity programs. It can also be delivered in a variety of ways that match each participant's learning style, readiness to change, self-confidence, and lifestyle. However, providers of the program are required to complete an online facilitator course, participate in either an in-person or web-based training workshop, and pass an online exam. In addition, ALED has an online tracking system to use in evaluating programs to assess fidelity. Required training and an evaluation that provides feedback to the lead partner are ways that ALED, like other programs in this study, guide and monitor adaptation behavior by implementers.
Diplomas Now is also characterized by monitored implementation. Jeff Jablow, Senior Vice President at City Year, said, "We think there's a level of variation built into the model. Dedicated staffing, early warning indicator meetings, local funding and support -all these things play out with some variance." Diplomas Now also provides training and onsite coaching and guidance. They also require regular measurement of social climate and quantitative monitoring of student progress, which allows the lead and distribution partners to monitor the progress at each site and to step in with corrections or to alter the program. Schools are required to commit to this level of assessment initially in order to offer Diplomas Now.

PRESCRIPTIVE FIDELITY
A couple of programs scaling up via distribution networks require greater degrees of fidelity and discourage or disallow adaptation. For these evidence-based programs, the lead or distribution partner closely monitors implementation to ensure that fidelity is maintained. This is the case for the YMCA's DPP. Jonathan Lever, Executive Vice-President with Y-USA, said that the curriculum, which is approved by the CDC, "is implemented in the same way at sites." Participating Ys are made aware they cannot change the curriculum, the order of sessions, size of the classes, or other core elements (one YMCA's DPP program manager waves a two-inch, three-ring binder in the air to remind Ys that they must not deviate from what's "in the binder" as they deliver the program). Lifestyle Coaches complete training and certification courses annually to meet the requirements necessary to be an effective group facilitator. The Y-USA uses checklists, site visits, and real-time reporting to quickly identify where fidelity is lacking and how to support affiliates in taking rapid corrective actions. Documents about the YMCA's DPP suggest that fidelity management requires a major culture shift for YMCAs that are accustomed to choosing programs and adapting them as they see fit.
CATCH Healthy Habits also requires fidelity and monitors site delivery to ensure fidelity. Like the YMCA's DPP, CATCH is a health-based program. Peter Holtgrave, with the distribution partner OASIS Institute, says, "The curriculum is scripted. It is evidence based. We don't want to veer away from the true intent. The volunteers are allowed freedom of role choice and implementation place. For example, some team members like to switch roles; one week leading the interactive lesson and the next leading the physical activity. But the curriculum is the curriculum." All volunteer meeting facilitators attend training that includes the rationale, background and structure of the program. Facilitators are required to be certified as Active Generations/CATCH trainers. The training offers a chance for volunteers to practice the physical activity portion of the program and ask questions about lessons and implementation. They sometimes help collect evaluation data using the SOFIT tool used to assess physical education classes.

Conclusion
To our knowledge no prior study has assessed a larger number of effective social programs that have successfully scaled up in the United States than this report, and none expressly explored how pathways, partnerships and fidelity relate to scale up. We found very little cross citing among investigators in education, youth development, and health, let alone other substantive areas such as substance abuse, community psychology, communication and nursing. Since we find scale up occurring and being written about in each of these areas, we offer the current report as a step in an interdisciplinary direction.
Our study is necessarily exploratory given the pre-paradigmatic stage of research about social program scale up. This study is also based on a purposive sample so results should not be generalized to a larger population of social programs that have scaled up. Within our sample of 45, however, we did find themes and patterns that warrant discussion and further investigation. We discuss these themes and patterns below in the hope that they will propel this field forward.

WHAT WE DID
We set out to build on recent work about approaches for organizing the scale up of social programs. We did this through an exploratory comparative case study of 45 social programs. We focused on trying to better understand three strategic choices that were common among the scale up efforts we examined, as depicted in Figure 5.1. As suggested by Figure 5.1, these decisions are related and may work in concert. A choice of pathway may affect fidelity and partner choice. Making the decision to go the distribution network route for rapid scale up is also a decision that the program in question can fare pretty well on its own, with little if any support from the developer or national host organization. This may be the case because the program is simple, or robust in its effects, or very well codified for staff new to the program to readily understand and deliver it effectively. A choice of partners early on may influence the pathway pursued and a team's orientation to fidelity. If a team is committed to a particular partner organization, then bending to that organization's way of growing and providing support to implementers may be desirable. And the desire or requirement for fidelity may lead to a decision about which partners make the most sense and which pathway will best facilitate the replicated achievement of fidelity. All three of these strategic considerations are choices made by lead partners, sometimes in conjunction with supporting and distribution partners, and sometimes with the agency and input of implementing partners-those local organizational staff who provide direct services to students, patients, and community residents. We defined scale up as increases in quantity, as in the number of sites or the number of beneficiaries served, and the maintenance of quality, as in later iterations of the program continuing to provide comparable benefit to participants. We did not assess program effectiveness (others had done that). Our focus was on what partnerships did to retain and in some cases, improve effectiveness and how the pathway chosen was related to both delivery of an effective program and partnership formation and operation.
We used expert informants, database searching and web searches to identify prospective social programs. Databases were the Catalogue of Nonprofit Literature, Issue Lab, Education Resources Information Center and ProQuest Social Sciences. Together these sources generated an initial set of 105 social programs that we reduced to 79 on the basis of relevancy. We then subjected each of these 79 to a set of inclusion criteria (Chapter One), resulting in the final sample of 45 social programs that we included in this study.
We retrieved documents from websites, news releases, evaluations, published journal articles, book chapters, final reports, foundation publications, corporate reports, government reports and publications, and other third party documents. Team members used a document coding protocol to conduct a content analysis of the documents for each social program. To address questions that the written records did not contain answers to as well as to update our information, we conducted personal interviews with 100 program leaders, advisors and staff.

WHAT WE LEARNED ABOUT PATHWAYS
Pathways to scale have varying labels, such as expansion, replication and collaboration (Management Systems International, 2012). Gabriel (2014) writes about scaling routes of several types. We drew on such descriptions and also on the business literature to identify three general means by which social programs reach more sites and benefit more people: Branching, affiliate, and distribution network pathways, with the acknowledgement that this is not a comprehensive listing of all pathways to scale.

USE OF PATHWAYS IS DYNAMIC
Although our focus was on a specific scale up time period for each social program, we found that many social programs had cycled from one path to another, intended to change their approach to scale, or were using multiple pathways at the same time. Our results suggest that a branching pathway may be a gateway path; i.e., a preliminary way forward that allows a lead organization to test a social program under somewhat controlled circumstances, develop its own organizational identity and build up the organization's capacity, and do things on its own time schedule. Once this foundation is established and as unmet need is identified that the organization is unable to respond to, the lead organization may transition the program to another pathway that can better achieve organizational objectives. Shifts from one pathway to another may result from changing environmental conditions external to the partnerships, such as policy change or funding availability. Designation as an evidence-based practice by a federal funding agency can quickly move a social program from a branching to an affiliate pathway because increased demand from jurisdictions or organizations due to funding availability will place pressure on a lead partner organization to meet that demand. Expert consultation from private foundations can have a similar stimulus effect. In addition, as programs are being implemented in the field and developers receive feedback about how implementation is proceeding and the beneficiary experience, they may make additional decisions to create variants of a program that are better fits with implementing organizations, by activating alternative pathways. A single lead organization, then, may have variants of a program scaling in different pathways at the same time.

INITIAL ORGANIZING INFLUENCES PATHWAY CHOICE
In this study, social programs scaling using a branching pathway had their origins in practice. These programs largely started from the personal interest of one or a small group of leaders who developed a nonprofit that grew to accommodate the scaling of the program. We often found that the organization and the program initially shared the same name. Leaders of these organizations and programs tend to see themselves, at least initially, as uniquely qualified to take their program to scale and, thus, a branching pathway is a good fit for the organization and the leader. How these organizations are initially funded also influences pathway decisions. As one respondent stated, "Why would you give a program away? That's often the only way you can attract funding." Organizations that are dependent on a single program to generate funding for the organization may be less inclined to share control and access to the program since they are dependent on it for funding and identity.
We found that social programs scaling in a distribution network had a different origin. The majority of these programs began as research projects within universities or research-focused organizations; few started as practice based programs. The organizations leading the programs typically pre-date the development of the program and have programmatic and leadership responsibilities which extend beyond the program of study. Further, the structure of universities and the norms of research favor program development and testing, not time spent "doing" scale up. Funding is also an important consideration. Research-based programs are often initially funded by research grants with some subsidization from the lead organization. The intent of many of these research grants is to test an intervention, not to bring it to market. In addition, when the developer is affiliated with a university, her salary may not be dependent on funding for the program. If the program ends, she, and perhaps her research team, will simply transition to another project, as she teaches courses, advices students or works on other grant funded projects, without a disruption in income.
So, the point of origin is important as initial organizing and funding structures may be more, or less, compatible with specific pathways.

PROGRAM MATURITY VARIES BY PATHWAY
In our set of programs, we saw differences in the maturity of organizations and programs using branching versus distribution network pathways. Many of the social programs we studied that use a distribution network pathway tended to be newer, with scale up occurring quickly relative to other program trajectories. A few of the programs that are scaling via a distribution network pathway were designed with scale up in mind, as a plan, from the very beginning. Newer program partnerships may be choosing distribution network pathways due to a new way of thinking about partnerships and collaboration. Much has been written in the past 10 years about scaling up using organizational networks. This emphasis is certainly reflective of investments by foundations to intentionally foster social impact by using organizational networks. It may also be reflective of changes associated with the explosion of social media. More and more individuals and organizations are working in virtual networks and able to quickly establish new connections. This way of working and living may predispose people to collaborative problem solving and the sharing of solutions. Whether such platforms can effectively support the achievement of scale up depth and program ownership (Coburn, 2003) is yet to be determined.

PROGRAM CHARACTERISTICS VARY WITHIN PATHWAYS
We anticipated that highly complex innovations with many moving parts, multiple implementers and that are context-sensitive would scale using branching pathways. They did. But we also found complex innovations scaling using affiliate and distribution network pathways, especially when program leadership and staff had operationalized a way to carefully train and monitor the performance of staff. We also anticipated that social program developers who wanted high degrees of fidelity would choose branching pathways. Again, they did. But so did some developers who were using affiliate and, to a lesser degree, distribution network pathways. Differences within a pathway were sometimes as pronounced as difference between pathways. This is most obvious within affiliate pathways where programs vary in complexity and leaders vary in the extent to which they control adaptation -about half of the programs were seeking prescriptive fidelity and about an equal number were supporting responsive adaptation. It is possible that the characteristics of the program play a secondary role in determining a pathway in the programs we studied; other factors, such as organizing structures, leadership decisions and partner selection may play a more deterministic role.

WHAT WE LEARNED ABOUT PARTNERSHIPS
Though it is common to speak of lead partners or social entrepreneurs as individuals, in practice we found partnership constellations that draw their membership from multiple organizations and for multiple purposes. The prevalence of partners and the critical roles that they play suggest that the scaling of social programs is unlikely to occur solo. Program scale up appears to require multiple core partners and, for some pathways, a great many implementing partners, too. Our findings reinforce the observation made earlier in this report: "Today… it is generally accepted that successful social entrepreneurs are masters at mobilizing alliances of groups and individuals to all work together for a cause" (Bloom & Chatterji, 2009, p. 119).

Conclusion
Newer program partnerships may be choosing distribution network pathways due to a new way of thinking about partnerships and collaboration.

SCALE UP INVOLVES MULTIPLE PARTNERS
Partners are at once constant and different across the pathways we studied. Across all pathways, supporting partners were identified as being very important by the program leaders we interviewed. None of the lead partner organizations had all the resources and knowledge needed for strategic planning, funding, infrastructure such as trainers and workshops, and evaluation expertise. Even after one accounts for the local organizations that serve as implementing partners for a social program, the national and regional teams still need support. Funders were especially prevalent, with some of the leaders identifying up to three core partners all of whom were funders. Funders were often private foundations with regular mention of federal and state agencies. Less frequently mentioned were corporations and corporate foundations.
Lead partners in branching and affiliate pathways nearly always mentioned implementing partners as being very important. Specific implementing partners, such as Oakwood Elementary, were seldom mentioned. Instead, leaders described their implementing partners in general terms ("schools", "community centers"). In contrast, lead partners in distribution network pathways talked about distribution partners as core partners and identified them by proper name. Distribution partners have their own network of organizations and by working with them, lead partners share in the control of implementation. Distribution partners were sometimes large, complex nonprofit organizations with far flung implementation sites, such as the Boys & Girls Clubs of America. Some were membership associations like AARP. In two cases, the distribution partner was a for-profit publishing company. Implementing partners were infrequently mentioned as being core partners by leaders who were scaling up via a distribution network pathway.
That leaders did not always identify funders or consultants or even implementers as core partners does not mean that these stakeholders are not present or that they are unimportant. Such partners may be allies but not be thought of as part of the core team that does much of the leadership and coordinating work of scale up.

PARTNERS PLAY MULTIPLE ROLES
Any one partner can play multiple roles in a partnership. Foundations, in particular, often did more than provide funds. and created new networks of thought partners. Distribution partners also played roles beyond their primary purpose of spreading and supporting the implementation of programs. Some of them were deeply involved in the development and reinvention of social programs.

SOCIAL NETWORKS CONNECT AND ENABLE
The importance of partnerships begs the question of how partners come together. In our results, lead partners in branching pathways tend to find their partners by searching for them. Lead partners in affiliate pathways also primarily find their partners by proactive searching, though they do form partnerships as a result of partners finding them. We find a different pattern emerging from distribution network pathways where lead partners are more likely to rely on previous experience to find a partner, though they also search for them.
Active searching by lead partners can be effective if they know where to look. Knowing where to look suggests the importance of being embedded in social and professional networks that span organizations of various types as a means of accessing partners.
Being embedded in such a network increases the likelihood of a third party introduction, or of a lead partner being "discovered" by a future supporting, implementing or distribution partner. It also increases the likelihood of happenstance -of discovering a partner when you don't expect to. Being connected in a social or professional network doesn't just mean you're in touch with others and know what's going on; it also means that relationships can be used to access resources. This is the real power of social networks: being a member in them enables an organization to get what it needs to scale up a program.
We found program developers who feel isolated professionally. They wanted to meet others who could help and advise them about scale up. These interviewees wondered aloud about how they might better learn from and share ideas with others who are in the business of scale up. There may be a need to assist in creating relationships among potential partners and to invite social program developers into those relationships. Greater embeddedness in social networks may also help to reduce the uncertainty of how well suited potential partners are for each other. is new and, we submit, an insightful contribution to the literature about program fidelity. Our results show that core partners worked to find a balance that is right for their program, a balance between maintaining fidelity while promoting a degree of flexibility to reflect the unique circumstances of delivering a program to beneficiaries.

REINVENTION IS COMMON
Reinvention is a normal part of the development of nearly all of the social programs we studied.
As they debut and test and grow their programs, developers reinvent, trying to find a fit between their program and their early collaborators and the implementers at the few sites where a program is initially tried. In some cases, the juncture between "pre-scale up" and scale up is clear, and a conscious effort to reinvent is made to improve the likelihood of a positive response by future adopting sites. In other cases, the ramping up is not a sudden step up but rather a continual and gradual increase in the number of sites, organizations and beneficiaries exposed to a program. For these partnerships, reinvention may occur a bit at a time, again and again.
Many of the organizations hosting the programs we studied describe themselves as learning organizations. Part of what they're continuously learning is how to ensure that the program is effective for its intended beneficiaries. Reinvention was sometimes a seemingly minor modification, but even minor changes to a program can demonstrably affect its chances to reach and benefit more people.
What did leadership do to programs that were considerably reinvented? Our results show that they often changed their delivery mode, changed or expanded the audience or beneficiaries, or altered their focus-and many of these changes were made to reduce the cost of the program or to increase its advantages to implementers.

REINVENTION IS COLLABORATIVE AND SOMETIMES TRANSFORMATIONAL
Although final decisions about programmatic or scale up changes reside with a lead organization in branching pathways, our results show that their decisions to reinvent were influenced by funders and consultants. Regarding affiliate pathways, partners often engaged in interactive redesign where partners work together to jointly reinvent. We found several examples of integrative reinvention among core partners in the distribution network pathway and, to a lesser extent, in partnerships taking affiliate pathways. Integrative reinvention took different forms and had different outcomes. Some integration brought organizations close together and we observed a blending of goals. In a few, the organizations merged and lost organizational distinctions. The result of integrative reinvention was sometimes very significant, both programmatically and organizationally. Contractual Reinvention is a normal part of the development of most of the social programs we studied. reinvention -where reinvention is controlled via contracts and formal agreements -was found in affiliate and distribution networks and not present in branching pathways.
When we observed partners working together to purposely reinvent programs and make operational changes to their own organizations, they tended to use a distribution network pathway. Partners do not figure prominently in reinvention decisions made in branching pathways; the lead organization makes the decisions though these decisions are informed by core partners.

IMPLEMENTATION GUIDANCE IS AVAILABLE
Each lead partner in this study provided guidance to implementers about how to best deliver their program. Programs had implementation guides, tutorials, and trainers available. Online resources are plentiful and include training videos and compelling first-person stories from implementers and beneficiaries. Many of the programs hosted regular meetings and/or annual conferences where implementers gather to share and learn with each other. Often, this support was freely available. Many programs also had personal and tailored training available for a fee.

ADAPTATION IS OFTEN EXPECTED AND SOMETIMES ENCOURAGED
Among our study's programs, results show that most lead partners believe adaptation to be required for effective and continued implementation of their program. Most leaders had personal experience with field conditions; they had worked in or studied clinics or community centers or schools at some point in their lives. They knew that even the best program could fail if put into place in an ill-suited environment.
The point of adaptation is to adjust a program so that it best suits a particular delivery organization and/or its clientele. We found that when adaptation was expected or encouraged, sometimes certain components of a program were prescribed while others were only suggested, or could be implemented in one of a number of ways. When the latter was the case, program staff would often communicate multiple examples of implementation carried out in different ways as a menu of examples for new implementers. Being rigid about certain program components but flexible about others was a common orientation to fidelity by program leaders.
Across the three pathways, we found lead partners who anticipated adaptation. Anticipation was most common among the social programs scaling via affiliates and distribution network pathways. As mentioned earlier, many of these programs were collaboratively designed with distribution partners who were well aware of differences among implementation sites and knew from working with their sites that adaptations would occur.

Conclusion
Most lead partners believe adaptation to be required for effective and continued implementation of their program.

MONITORING FOR CONTROL AND LEARNING VARIES
Across the branching and affiliate programs are many examples of the use of contracts, memorandums of understanding, formal agreements, data reporting requirements, dashboards, ongoing evaluations and reviews, required trainings and meetings to monitor and control the implementation process. This is true for programs in which adaptations are encouraged and especially so for programs when adaptations are not encouraged. Affiliate pathway programs often require the implementer (who is also the affiliate) to earn the right to offer a program. That right can be rescinded if adherence to protocol is not sustained.
The importance of monitoring shifts when we look at programs scaling up using a distribution network pathway. Among these partnerships, just more than half anticipated or encouraged site-specific adaptation while not continuously monitoring for implementation. These are programs that tend to have greater collaboration in design and reinvention, so perhaps monitoring is less necessary. But this may also suggest a trade-off for some programs that use a distribution network approach -the developer may need to allow others to address issues of implementation fidelity, or rely on training and support materials and YouTube clips of correct delivery to minimize variation in implementation. Because we did not analyze field-based implementation and program outcomes, we cannot conclude that a loosening of fidelity equated with a loss of effectiveness, though it can result in such for complex and detailed programs (Dusenbury, Brannigan, Falco & Hansen, 2003;Mowbray, Holter, Teague & Bybee, 2003;Blakely et al., 1987;Green, Ottoson, Garcia & Hiatt, 2009).

Moving the Field Forward
Opportunities exist for scale up partnership teams to learn from researchers, and of course, as has been our personal experience in this study, for researchers to learn from partnership leaders about what practitioners have tried, what has worked, and what is unfolding next. Before we end by drawing attention to four directions that we believe could move this field forward, we draw attention to a general observation about the current state of the field: Clearly and without doubt, scale up of social programs in the U.S. is occurring in the domains of education, youth development and health. The phenomenon is real. The social programs propelled by scale up partnerships that we studied combat some of society's most important problems with aggressive change agendas. Climate change, diabetes, illiteracy, infant mortality, learning loss, violence, low esteem, hunger, homelessness, social exclusion, and low expectations are big, systemic social problems. These cases of scale up are ready examples of how social entrepreneurs and a variety of organizations can work together to affect change. These are real success stories.
The importance of monitoring shifts when we look at programs scaling up using a distribution network pathway.
Based on what we have learned, we offer several suggestions here that may help scale up partners to further improve decision making processes about social programs.
First, scale up partnerships might emphasize design activities of several types. Pre-scale up design activities can be conducted to estimate population need, program readiness, lead partner readiness to provide access to a program and support implementation, and implementing organizations' capacity and motivation to do a good job with a program (Dearing, Smith, Larson & Estabrooks, 2013). Tools and metrics that assess these topics exist and could be used to help assess which of a set of programs is best suited to scale up and which aspects of specific programs require reinvention to improve their odds of being accepted. Similarly, strong evidence exists about how evidence-based implementation techniques can be taught and learned (Miller, Sorensen, Selzer & Brigham, 2006). Is applied work like this necessary for scale up? The literature about the diffusion of innovations is clear that social justice is commonly a problem in what spreads to whom, and when (Buchanan, Cole & Keohane, 2011). Wealthy organizations in wealthy communities where capacity is high tend to be early adopters of new effective programs (Bhatti, Olsen & Pedersen, 2010). Poor organizations and poor communities adopt later, with inconsistent uptake and less success in implementation. Overtime, divergent patterns of the haves and the have nots emerge (Rogers, 2003). These general tendencies can be counteracted through designing for diffusion (Green, Gottlieb, & Parcel, 1991). Design activities represent a systematic means of providing social programs to those communities that need them the most.
Second, the 45 cases reported here make little systematic use of the concept of social influence. There are now a number of tools for identifying which individuals, organizations and communities are opinion leaders among their peers. Social influence-people and organizations looking to each other for cues about new ideas, products and services-has always been a driving force in the diffusion of innovations (Bohlmann, Calantone & Zhao, 2010). Whether one is trying to convince school districts, community centers, or CEOs to try a new program, knowing who influences whom can accelerate the rate of scale up because, in many instances, later adopters imitate earlier adopters. Social influence, when used by scale up practitioners, can also deepen the commitment to implement and sustain a worthy program once it has been adopted by implementing organizations in communities. This knowledge is well-codified and clearly applicable to social programs once they are ready to scale, for example, from jurisdiction to jurisdiction in pursuit of policy stream funding (Karch, 2007;Mintrom, 2000).
Third, a number of interviewees told us that they had an information deficiency and lacked the connections necessary to learn from their peers. Scale up processes, barriers and facilitators can be rather similar across topical areas like health and education. Yet, many nonprofits are topically focused and may not work outside of their fields or geographic area. Thus it is challenging to know who may have experience with particular pathways, partners or aspects of fidelity. The limited availability and accessibility of an interdisciplinary knowledge sharing system for decision making holds this field back. There are promising and innovative networks for entrepreneurs such as The Hive and for civic-minded entrepreneurs such as Code for America. Could a coordinated and useful system of resources link together social program leaders in education, youth development and health?
Finally, private foundations that see investments in the spread of effective social programs and resultant social impacts as relevant to their portfolios have an opportunity here, perhaps as conveners since federal agencies and corporate foundations are engaged in this area, too. As our results show, several funders have repeatedly worked alongside one another in common cause to scale up social programs. Thus there is precedent for private foundations to consider further collective action in preparing the next generation of scale up practitioners and researchers. weathercasters receive content customized to their locality with readymade climate data and graphics.
climatecentral.org/what-we-do/our-programs/climatematters Clinical Chapters. A program offered by A Home Within with the goal of bringing a stable caring relationship into the lives of foster youth. A network of qualified therapists agree to see a foster child, teen or young adult in weekly, pro-bono therapy. Senior clinicians provide consultation, and volunteer psychotherapists have access to ongoing education and training, an expanded referral base, and a network of colleagues.

ahomewithin.org
College Possible. A college readiness and college support program serving young people from low-income communities. College Possible high school graduates are coached during their junior and senior years of high school in key aspects of preparing for college. They also receive coaching support as they transition to college and all the way through college graduation. AmeriCorp members guide students through a set curriculum in high school and provide coaching in college.

collegepossible.org
College Summit. A comprehensive college access program that helps high school students from low-income communities to and through college. Includes training rising senior students as influential peer leaders, providing school year curriculum for 9-12 grade, and analyzing data to improve college enrollment goals.

collegesummit.org
Communities that Care. This process focused innovation guides communities through a proven five-phase change process that prevents youth problems before they begin. The process begins with a youth survey to identify a community's risks and strengths. Based on these data, Communities that Care helps communities select and implement tested and effective prevention programs and amplify programs already working.

communitiesthatcare.net
Diplomas Now. A comprehensive whole-school reform intervention that identifies students at risk for dropping out and works to eliminate the problems that lead to dropping out, including poor attendance, poor behavior or course failure in English or math. Each student at risk works with an adult who has the capacity to help the student improve achievement during and after the school day. An early warning indicators system identifies struggling students for intensive assistance.

diplomasnow.org
Earth Force Process. A six-step instructional model delivered via a professional development team that works with local educators and groups of students on an environmentally-focused service-learning project of their choice. Students work together to design and implement a project that explores root causes of a community problem and takes action to improve it. earthforce.org Experience Corps. An intervention that engages older adults as literacy tutors to children in grades K-3 with the goal of helping children to read on grade level by the end of third grade and to keep older adults physically and mentally active. After a small-scale start, AARP purchased the program and made modifications to broaden its applicability, and now supports expansion through their national network.
aarp.org/experience-corps Gateway to College. Reconnects high school dropouts with their education. Students complete their high school diploma requirements at community and technical colleges while earning credits toward an associate's degree or certificate. The program is administered by the Gateway to College National Network. Gateway programs are supported by training, technical assistance, coaching, and program evaluation provided by national network staff.

gatewaytocollege.org
Girls on the Run. Physical activity and positive youth development program for girls in 3rd-8th grade. Life skills are taught through interactive lessons and running games. The program culminates with the girls being physically and emotionally prepared to complete a celebratory 5k running event. The goal of the program is to unleash confidence through accomplishment while establishing a lifetime appreciation of health and fitness.

girlsontherun.org
Higher Achievement. A year-round after-school and summer program designed to close the opportunity gap for middle school youth in at-risk communities. The program consists of 1) Afterschool Academy that includes core subjects as well as seminars, field trips, community service; 2) Summer Academy which includes coursework and electives, as well as a three-day trip to a college; and 3) High School placement services which help scholars advance to top academic highs schools.

higherachievement.org
Housing First. A model program that provides housing in order to then address a variety of mental health, substance use, and employment needs for adults including returning veterans. For people who have experienced chronic homelessness, there is an expectation that intensive and often specialized services may be needed indefinitely. Program staff work with a wide range of stakeholders including social service agencies, clinics, landlords and employers.

pathwayshousingfirst.org
The Incredible Years -Parents, Teacher and Child Programs. Three interlocking, comprehensive, and developmentally based programs targeting parents, teachers and children. Training programs are guided by developmental theory on the role of multiple interacting risk and protective factors in the development of conduct problems. The three programs promote socio-emotional and academic competence to prevent, reduce, and treat behavioral and emotional problems in young children.

incredibleyears.com
Intergenerational Tutoring. Run by OASIS this program promotes literacy by helping high-need children build the skills they need to read at grade level. Elementary school teachers help identify children in grades K-4 who need help. Then with parental consent, trained OASIS volunteer adult tutors are paired with children. Tutors work one-onone with students each week throughout the school year.
oasisnet.org/National-Programs/Intergenerational-Tutoring Jumpstart. A program that trains college students and community volunteers to serve preschool-age children in low-income neighborhoods with a focus on language and literacy skills. Two-hour sessions take place two days per week and revolve around a core storybook. Early education partners, many of which are Head Start programs, provide space and materials.

jstart.org
KaBOOM! Organizers build a playground during a single day. Six months prior to the build KaBOOM! finds a funder, searches for an appropriate site and considers the economic needs of the neighborhood. Ten weeks before the build, project managers visit the community and speak to children about what types and colors of playground equipment they prefer. Local communities raise 10% of costs.
kaboom.org Lifeskills Training. A substance abuse prevention program to reduce the risks of alcohol, tobacco, drug abuse, and violence by targeting the social and psychological factors that promote substance initiation. Adolescents and young teens are provided the confidence and skills necessary to successfully handle challenging situations.
lifeskillstraining.com/index.php Money Matters. This program teaches teens basic money management skills to help them save, spend and invest wisely. The program consists of a personal finance guide and a facilitator's guide and is offered through a partnership with the Boys and Girls Club of America. bgca.org/whatwedo/EducationCareer/Pages/MoneyMatters.aspx

Moneythink.
A financial education program that uses college age near-peers and a customized curriculum to improve the financial capabilities of high school students. Online networks and technology are integral to the program.

moneythink.org
National Writing Project. Mission is to improve student achievement by improving the teaching of writing and improving learning in the nations' schools. The National Writing Project (NWP) is a nationwide professional development network that serves teachers' across disciplines and at all levels.

nwp.org
Nemours BrightStart! A screening and targeted instruction program to help children achieve reading success. The program offers direct services for young children as well as tools, services and resources to parents, educators, health care professionals and community leaders. Now offered through a national partnership with Kaplan Early Learning Company.
nemours.org/service/health/brightstart.html Nurse-Family Partnership. A national community health program that serves low-income women pregnant with their first child. The program provides professional nurses who make home visits to expectant and young mothers over 30 months. The program improves pregnancy outcomes, improves child health and development, and improves the economic self-sufficiency of the family.

nursefamilypartnership.org
Playworks Coach. Aims to address the physical, emotional and social needs of elementary school children by coordinating play and physical activity programming-throughout the school day-taught from a framework of youth development.

playworks.org
Posse Scholars. Identifies public high school students with extraordinary academic and leadership potential who may be overlooked by traditional college selection processes. Posse extends to these students the opportunity to pursue personal and academic excellence by placing them in supportive, multicultural teams-posses-of 10 students. Partner colleges and universities award scholarships and posse scholars are supported by posse mentors.
possefoundation.org/our-scholars Power Scholars Academy. A six-week summer program designed to tackle summer learning loss in math and reading as well as foster physical and social-emotional growth. The program is a partnership between the YMCA and Building Educated Leaders for Life, and involves small group activities, academic instruction, hands-on enrichment activities, field trips and service projects.
experiencebell.org/category/tags/power-scholars-academy Professional Training Corps. A year-long intensive training program that provides low-income young adults with hands-on skills development, college credits, corporate internships, and support. Classroom activity is augmented with a six month internship with a corporate partner. Students earn college credits and a weekly stipend, and are supported by staff advisors, professional mentors, dedicated social services staff, and a network of community-based partners.
yearup.org RALLY. Encourages learning, promotes social-emotional development, fosters trusting and caring youth-adult relationships, and builds youth resilience and interpersonal skills. The program partners with after school groups and school districts that have the internal capacity to align the RALLY approach with their organizational missions.
hfrp.org/evaluation/the-evaluation-exchange/issue-archive/ current-issue-scaling-impact/the-rally-program-scaling-aninclusive-approach-to-intervention-and-prevention Reading Partners. Aim is to increase student reading proficiency. Reading Partners is a community-driven children's literacy program that recruits and trains volunteers to work one-on-one with students who are at least six months behind in reading.

readingpartners.org
Reclaiming Futures. Designed to impact community and state-level systems by bringing together judges, probation officers, substance abuse treatment professionals, and community members to work together to help teens in the justice system. Participating sites receive technical assistance including a coach, guidance on the 4-module online tool kit, webinars, and assistance with communications.

reclaimingfutures.org
Safe Surgery. This program began with the World Health Organization goal to reduce high numbers of patient complications, injuries, and deaths, not only in modern sophisticated hospitals, but in poorly equipped, understaffed settings as well. The intervention is a checklist of simple, practical and affordable changes that can be executed in any operating room, anywhere.
ariadnelabs.org/areas-of-work/safe-surgery Streetwise Steps MBA. Business education for small business owners in historically underserved and lower income areas through provision of tools, training, and network access. Participants meet bi-weekly in small classes that combine classroom learning with real-world case studies and the participant's business. Industry experts, business peers, and instructors provide help.
interise.org/streetwise-mba Summer Search. Supports students from low-income backgrounds to transform what they believe is possible for themselves and develop the skills they need to become college-educated leaders who give back to their families and communities. Selected students participate in workshops and one-on-one mentoring during the school year and challenging summer trips and are assigned college mentors.
summersearch.org Teen Outreach Program. Designed for teens in 6th to 12th grade and guided by a nine-month curriculum. The program reduces the risk of problem behavior while promoting healthy choices and empowering teens to lead successful lives and build strong communities.

teenoutreachprogram.com
Urban Debate. Prepares low income students of color to succeed in college and in future careers by organizing and supporting competitive debate teams in urban public schools across the country. Students in teams learn skills and develop and practice confidence and self-and collective efficacy in competitive situations. urbandebate.org

WINGS.
Recruits the most challenging students attending the most at-risk public elementary schools from low-income, predominantly African-American families. The program integrates social and emotional learning curriculum into activities that teach students to master self-awareness, emotions, social awareness, healthy relationship skills, and responsible decision-making.
wingsforkids.org YMCA's Diabetes Prevention Program. A lifestyle intervention designed to decrease the incidence of type 2 diabetes among pre-diabetic individuals. Participants meet as a group with a trained YMCA lifestyle coach for 16 core sessions to learn and practice lifestyle skills. During sessions they learn how to make healthier food choices, how to incorporate more physical activity into their daily routine and how to manage a healthy weight. ymca.net/diabetes-prevention