Growing your innovative programs isn't just about adding more users. A new framework requires educators to stretch old notions of what scaling is and how it can be successful.

ScaleYOU'VE PROBABLY SEEN IT happen in your own school district. Somebody comes up with a great idea that has a dramatic impact on student success, and suddenly other teachers, other grades, other schools want to adopt it too. It could be a new game for engaging kids in math, or a video podcasting project for reading and writing. Maybe it's a new way to teach teachers how to integrate technology into their lessons. Yet, when that idea is transferred to a new environment-- a different classroom or school-- it doesn't produce the same success. That's the challenge inherent in scaling.

And it's a challenge that poses more trouble for education than it does for other fields, according to Christopher Dede, a professor of learning technologies at the Harvard Graduate School of Education.

The difference between education and other industries is that "every educational situation is different in a way that is more profound than every fast-food restaurant being different," says Dede, who is co-editor of the 2005 book Scaling Up Success: Lessons Learned From Technology-Based Educational Improvement (Jossey-Bass). "If there's a new way to fry French fries, you can adopt it at a whole variety of fast-food places. But in education, the range of variation is much higher." Those variations include such factors as how the teacher prepares content, what pedagogical style is used, and even classroom or school culture.

The common approach to scaling, says Dede, is to jump in and say, "Let's go out and find more money, recruit more participants, hire more people. Let's just keep doing the same thing, bigger and bigger." That, he observes, "tends to fail, and fail badly." Dede says some innovations just aren't built for appropriating by other institutions. "There are many, many programs out there that are of high quality, [but] there are very few that operate at scale, that aren't just a little hothouse program that only affect a few people-- and really are incapable of doing more than that."

"The real measure of whether something has scaled is basically whether it scales. It's not whether it follows the [five] dimensions. It's whether it manages to have a huge impact."

Got Scale?

Finding the most effective way to transfer a successful program to other environments is the aim of a current research project sponsored by Microsoft that focuses specifically on how to apply technology to scaling education innovations.

The project is an outgrowth of the company's US Partners in Learning (PIL) program, begun in 2003 as a five-year, $250 million global project to build public/private partnerships for the purpose of meeting three goals: universal digital literacy, a stronger workforce, and improved quality of life. Because of its success, the program was re-upped for another five years in 2008. In 2005, to learn more about what it takes to scale up successful educational initiatives, the PIL team dedicated a portion of its funding to a select group of organizations that had programs it deemed ready to be scaled through the use of technology.

Allyson Knox, Microsoft's academic program manager, brought Dede, a former graduate professor of hers at Harvard, in to work as a consultant with initiatives chosen for what was named the Mid-Tier Project.

As grant manager for PIL, Knox got used to hearing the same pitch from applicants: "I have this great program, I want to write a grant proposal, and I'm going to prove that I can scale or replicate this to x number of programs throughout the US." Then, they would submit plans, she says, that would be focused strictly on replication numbers: If we get five teachers trained in this, then each of them can train five more teachers, and that's how we'll grow this thing! Fixating on numbers is the wrong approach, Knox says, because you have to factor in steps that must be taken to address the different circumstances and challenges faced by users in different environments. Simply equating scaling with multiplying is a mistake.

Dede and Knox saw the need to educate grantees on the concept of scaling. As part of that effort, the team adopted, adapted, and began disseminating a framework first put forth in research published by Cynthia Coburn (see Links), an assistant professor at the University of California-Berkeley Graduate School of Education. Coburn proposed a definition of scale that went beyond growth in numbers to include depth, sustainability, spread, and shift. Dede and Knox, along with research consultant Saul Rockman, president of the San Francisco- based consultancy Rockman Et Al, added a fifth aspect to the framework: evolution (see "Redefining Scaling").

"Often, the first way people measure scale comes back to the number of students or teachers who have been trained," Knox says. "They don't look at all five dimensions, which would help them achieve greater impact and scale." That isn't to say that trying to scale a project will necessarily fail if all five dimensions aren't considered. "The real measure of whether something has scaled is basically whether it scales," Dede says. "It's not whether it follows the dimensions. It's whether it manages to have a huge impact."

Redefining Scaling

ScaleThe usual way to think about scaling a program is to consider how it can be expanded to reach more people. However, that may actually inhibit the program's potential to continue to work effectively. Harvard University's Christopher Dede and Microsoft's Allyson Knox have found that scaling encompasses five dimensions.

SPREAD: Diffusion of the innovation to a large number of users; the traditional way of thinking about scale.

DEPTH: Producing deep, transformative, consequential changes in instructional practice, leading to improved educational outcomes for students.

SUSTAINABILITY: Maintaining changes in practice over a substantial period.

SHIFT: What happens when districts, schools, and teachers assume ownership of the innovation and spread its impact.

EVOLUTION: The ongoing revision of the innovation by those adapting it.

Spread and Depth

One program whose grant application won support from PIL was the Teacher Leadership Project (TLP), a professional development program launched by Northwest Educational Service District (NWESD) 189, one of nine educational service districts in the state of Washington that provide administrative and instructional support services to districts and schools.

TLP grew out of an initiative begun in 1997, when NWESD, using a modest grant from what was then the William H. Gates Foundation, pulled together a group of 27 teachers statewide for a special week of summer training that would show them how to integrate new types of technology into student-centered learning, with a heavy emphasis on using the technology in support of collaborative, project-based exercises. After the week was up, participants continued getting together every couple of months to maintain the learning, sharing, and collaborating. That initial undertaking led NWESD to create the Teacher Leadership Project, and those original participants became trainers themselves, training 150 teachers in the same collaborative lessons during the summer of 1998 and another 250 in 1999. By 2000, NWESD was confident enough in the program to appeal for a larger grant. The Bill & Melinda Gates Foundation, renamed that same year after absorbing the William H. Gates Foundation and the Gates Learning Foundation, came forward with $45 million to support a major expansion: training 1,000 teachers a year for three years throughout Washington.

Scaling up the program involved hiring "a lot of people" and managing "a lot of logistics around organizing"-- in other words, doing more, more, more-- according to Becky Firth, who as director of NWESD's Technology Leadership Center oversees the Teacher Leadership Project. Firth, a former English teacher, was one of the original 27 educators chosen by NWESD when it launched the project in 1997. To reach teachers statewide, TLP ran four consecutive weeks of training in nine different locations each summer. Following the workshop, the same participants were brought together three more times during the year for Friday/Saturday sessions, with the cost of substitutes covered by the grant money.

By the time the funding had run its course, 3,000 teachers in the state had gone through the training. "A Decade of Reform: A Summary of Research Findings in Classroom, School, and District Effectiveness in Washington State," a 2003 report from the Washington School Research Center at Seattle Pacific University, hailed the program, stating, "Evaluation results have shown that many participants have changed their classroom instruction, sometimes substantially."

Firth says that TLP's success prompted districts in other states to begin contracting with TLP to provide training for their staffs. But one factor made the arrangement untenable: Flying instructors across the country four times a year to conduct training was too costly. "We knew financially it wasn't going to work," Firth says.

The program needed to explore a different way to scale in order to reach educators outside of Washington. The solution: Add an online component.

This desire to use the web as a scaling tool won over Microsoft's Mid-Tier Project, which stepped up with $426,000 in funding to help TLP develop a hybrid program. The transition, however, from a wholly face-to-face model would require focusing on additional dimensions of scaling, on top of the traditional notion of spread: Depth and evolution would also come into play.

To ensure depth-- meaning the kind of profound changes in instructional practice that lead to improved educational outcomes for students-- Firth tapped the expertise of teachers who had been through the program and gone on to become trainers. The consensus: The face-to-face summer workshops had to stay. "That puts everyone on the same footing," Firth says. "Everyone has the same experience."

The initial workshops were retained, but the follow-up two-day sessions were moved online. Plus, a new web-based portal was developed where teachers could maintain and add to lesson plans they were developing in the program, share them with other teachers, and exchange feedback. To prevent dropouts, the program added online mentors to check in on participants.

Under the hybrid format, Firth says, TLP has been able to conduct face-to-face workshops in several states, including Wisconsin, Tennessee, Texas, Florida, Minnesota, and New York, but any further expansion of the program will have to wait. "I'd love to see thousands of teachers trained across the country every year," she says. "But we're thinking small-scale right now, maybe 100."

With the online portion of the training program successfully developed, TLP is now creating standalone training modules to add to its online content. That's an example of evolution in scaling-- rethinking core beliefs about the innovation.

The biggest question for Firth is whether this new hybrid model will ensure the final scaling dimension: sustainability-- maintaining changes over a long period. That's because significant changes to classroom instruction "aren't going to happen in a year; it's going to take two or three years," she says. "Our one year [of training] is still better than the typical three or four days of professional development a teacher gets. But it takes time to see that overall change in the classroom."

A Sixth Dimension: Emotion

ScaleIntroducing new ways of working can be challenging when scaling a project. "People are eager to change when they feel they're struggling," says Harvard University professor of learning technologies Christopher Dede. "When they feel they're operating at a very high level of quality, it's much harder emotionally to say, 'To have a big footprint, I have to do things differently, even though it seems risky to me to move away from my perfect method of operating.'"

In fact, he adds, that was one of the biggest surprises for him: the emotional dimension of scale. "Asking people to take something that was highly successful and modify it-- that's often the hardest thing for a person or organization to do."

Shift and Evolution

A second professional development initiative supported by the Mid-Tier Project provides an example of how scaling can be less about spread than it is about shift and evolution-- adapting what you already have. The initiative, the 21st Century Learning Project, was developed by the Alabama Best Practices Center. The center already provides professional development through a program called Powerful Conversations. Four times a year, educators meet to share ideas on improving teaching and learning in the state. But, according to the center's executive vice president, Cathy Gassenheimer, there was a sense that Alabama educators needed to learn more about how to use Web 2.0 collaboration tools in the classroom. "Students outside school have so many tools," she says. "Unless they're able to use them in school, we're going to lose them."

Not wanting to pull teachers out of school any more than it already was doing for Powerful Conversations' quarterly events, the center explored digital alternatives. With $426,000 in funding from the Mid-Tier Project, the center launched the 21st Century Learning Project in 2005, recruiting 20 schools to participate in a virtual learning community built around an online curriculum called "Keeping up With the Net Generation." Each school was allowed five participants, of which at least one had to be a principal, an assistant principal, or an instructional leader.

During the project's first year, four teams at a time met online after school for two-hour webinars, using the collaboration tool Elluminate Live! to hear presentations and have discussions. Follow-up was done via Tapped In, an online learning community of education professionals, as well as through in-person quarterly meetings. Also, the center selected 10 "fellows" to receive extra training in order to coach in schools.

By 2007, the Microsoft grant had expired and schools needed to self-finance their participation. The tab came to $7,500 apiece. Participants were cut from 20 schools to 10, because, Gassenheimer explains, "we wanted to make sure we could provide them with a whole lot of support."

But in 2008, the project received a $150,000 grant from the state, which meant schools would not have to pay their own way. In fact, each school got $2,500 from that fund to apply to expenses related to professional development. Plus, schools would no longer be limited to five participants; seats for each after-school webinar became first come, first served. Meanwhile, the project is now recording its sessions and posting them on a wiki for broader consumption. The subjects covered are becoming more topical, not just explaining the basics of Web 2.0. This year, says Gassenheimer, participants are learning "how to '21st-centurize' math and science lessons at elementary levels."

TO READ Christopher Dede and Allyson Knox's report, "Using Technology to Scale Up Innovations," which examines the scaling of the two professional development programs described in this article, visit here.

Scaling the 21st Century Learning Project has required frequent tweaking. A year after the 2005 launch, the idea of the fellows was dropped. "What we didn't calculate is, if you're teaching full time, it's hard to have time not only to coach your peers, but also to provide support to other schools," Gassenheimer explains. Plus, the program now kicks off with a face-to-face meeting, and the Tapped In service has been replaced by an invitation-only wiki site on the social networking service Ning.

The project continues to evolve as it is adapted by school districts. One participant, Talladega County Schools, added a tiered mentoring system, school visits by mentors, and "Friday Five," a weekly e-mailed newsletter with advice and tips. "When they called to see if we'd mind if they use our materials, we said, 'No, please, use it-- and let us learn how you're using it," says Gassenheimer. Now the center is bringing some of Talledega's ideas into its version of the training.

What the Mid-Tier Project has done through its support for the 21st Century Learning Project and the Teacher Leadership Project is show the importance of a multidimensional scaling framework when considering how to expand an innovation in education. "It's crucial [for] reflecting on ways not just to do the same things bigger and bigger, but to do something that's related to what you're doing but still different in some important ways," Dede says. "The framework is saying, 'It's more complicated than you think it is.' Stepping back and recognizing that complexity is a really important first step."

If you would like more information on scaling technology initiatives, please visit our website at Enter the keyword scaling.

Dian Schaffhauser is a freelance writer based in Nevada City, CA.

This article originally appeared in the 06/01/2009 issue of THE Journal.