The issue

In our work to improve higher education as in other domains, it’s not enough just to innovate.

Campuses across the country are littered with the failed remnants of excellent innovative projects that died because there was never genuine buy in, real financial and institutional support, or a hospitable policy environment. Even those pilots that survive the end of a grant or an administrative initiative typically remain small and are not often replicated elsewhere. The problem, of course, is that pilots—even those with clear evidence of success—rarely, if ever, replace the long-established structures they are designed to improve.

But to move that elusive achievement needle—for a college, a state, our nation—change at scale is what matters: Improvements must benefit the vast majority of our students.

Unfortunately, our pursuit of scale rarely has a clear road map—more often, it feels like a search for gold at the end of the rainbow. Scholar Archibald Cochrane wryly observed that scaling innovations was like working in a crematorium: So much goes in, and so very little comes out.

Scaling up a promising innovation is difficult under any circumstances. In higher education, it is especially challenging because of decentralized decision-making, antiquated incentive systems, and increasingly unpredictable funding challenges.

Given these inherent complications, it’s arguable that the basic premise of “scaling up”—that one starts with small pilot projects, and then grows the numbers of colleges or individuals served—is untenable. An alternative might be to work at scale.

The context

Currently, the Charles A. Dana Center at the University of Texas at Austin, in partnership with the Texas Association of Community Colleges, is developing a new approach to developmental mathematics education known as the New Mathways Project (NMP).

In this work, we are pioneering three integrated strategies to work at scale from the beginning—first in Texas. Below, I outline the three big ideas that inform our work to design innovations and initiatives for scale from the get-go.

The big ideas

1. Design the innovation for scale.

Let me illustrate this strategy with a story about Thomas Edison and his invention of the electric light bulb as chronicled by a 2001 study from Hargadon and Douglas. In the late 1800s, gas light was in fashion. It offered yellow, flickering, dim light of about 12 watts. It required a lampshade to protect it from drafts that might blow it out, and it posed a fire danger. Engineers piped gas under the roads and into homes and offices. The utility companies that produced gas had monopoly power in the market and exerted their influence to generate a favorable regulatory environment for themselves. A fleet of lamplighters and utility workers existed to maintain and service gas lights.

Edison’s genius lay not only in his technological advances to provide cleaner, cheaper, safer light, but also in his recognition that simply making available rational improvements to the existing lighting system was no guarantee that this reform would achieve broad acceptance. He understood that such innovations confront a complex society with established norms and cultures, ways of working and labor forces, formal regulations, and informal ways of doing things. Understanding how this context could help—or hinder—the spread of his innovation led him to adapt the innovation to seamlessly integrate into 1880s life.

To wit: He dimmed his bulbs to just 13 watts, and he covered his electric lights with lampshades, not because electric lights were at risk from drafts, but so that they were indistinguishable from gas lights. Rather than running his power lines above the roads—which he knew was more technologically efficient—he ran them under the roads to take advantage of the existing labor force and to operate within existing gas regulations so that new regulation would not be necessary for electricity. In sum, in rolling out the new bulbs, he designed for scale from the beginning by looking at the whole system and minimizing the cost of the transition in terms of financial costs and human-effort expenditures. He identified the non-vital elements of his design to adapt (for example, lamp shades), and he adapted them in ways that allowed him to make modifications to improve technical efficiency over time (for example, power lines now run above the street).

Edison’s principle of designing for scale has informed the New Mathways Project in a number of ways, including our approach to course approval, transfer credit, and advising. The NMP offers a faster, more rigorous, more contextualized approach to developmental mathematics; it aligns gateway mathematics courses to the skills students need to be successful in their programs of study.

One specific example of how the NMP design has evolved to enable scale from the outset is that it offers underprepared students a common first course for the NMP’s three different curricular pathways—statistics, quantitative reasoning, and STEM. This strategy addresses one of the project’s primary implementation challenges: getting the right student in the right path at the right time. A great many students have not identified a program of study when they start college, and very high student-advisor ratios rarely allow for the intensive support students need to make that choice. By creating a common starting point for the pathways and pairing the first mathematics course with a co-requisite student success course that includes explicit instruction on career exploration and degree planning, the NMP provides students with a highly supportive environment for a full semester, during which they can make an informed choice about their academic and career program and then select the math path aligned to that choice.

2. Design the initiative for scale.

Once one develops a promising innovation, the initiative to implement it at scale can sometimes fail to transfer the innovation from one institution to the next. While there are many reasons an innovation can fail to scale, one we see frequently in higher education is that the conditions under which pilot colleges succeed are often very different from the conditions experienced by other colleges that have been on the sidelines.

Conditions contributing to pilot success may include special grant funds to support the work, outside technical assistance, and the opportunity to define and “own” the development and implementation of the innovation. It is foolish to presume that new colleges will spontaneously begin scaling an innovation for which they do not have external funding, technical support, or opportunity to participate in development and implementation.

It is natural to initiate an innovation with small-scale pilots. Innovations bring uncertainty, and it is quite reasonable to expect that financial and human resources may limit how much new, high-intensity work can happen at a large number of colleges simultaneously. Yet innovations are more likely to scale if most of the target population is involved from the beginning. I posit that we can resolve this tension between resources available and numbers of stakeholders engaged by planting the seeds of scale at all target colleges from the outset, and by creating multiple levels of engagement that enable some colleges to work intensively from the start and others to ramp up engagement as evaluation data, stakeholder support, and resources become available.

Through the Dana Center’s partnership with the Texas Association of Community Colleges, all 50 community college districts in the state have signed on to participate in the NMP. Each college, through an application and enrollment process, selected their desired level of participation in the initiative among three tiered levels of engagement. We are working with nine co-development partners to create NMP courses and implementation resources that colleges will implement starting this fall (August 2013). A second group of colleges are choosing to participate as active learning sites; each will pair with co-development partners and begin implementing the courses (with support from the co-development colleges) in year two. Finally, the remaining Texas community colleges are choosing to engage as capacity-building sites and are laying the groundwork now (by hosting opportunities for faculty to begin learning about the NMP) for implementation to kick off in years three or four. Importantly, we structured the NMP initiative to enable all these Texas community colleges to provide input and feedback on the NMP resources now, as we develop them.

3. Seek permission to scale.

Institutions of higher education are, by design, multilevel systems with high degrees of autonomy at each level. Faculty own what happens in classrooms and are responsible for maintaining the integrity of their courses. Presidents ensure the institutional mission is carried out, and manage the institution’s operations and budgets. Trustees, advisors, legislators, state agency staff, and institutional researchers: Each has their own domain of authority and responsibility. And understandably, each tends to innovate within that domain. Thus, in higher education, one reason that we see so few innovations (especially classroom-level innovations) go to scale is that articulating action across this multitude of domains is enormously challenging.

In top-down approaches to scale, those who choose new innovations are often not the end users. Without shared ownership, innovations can become domesticated—that is, through lack of understanding, resources, or will to change, the innovation can lose the qualities intrinsic to its efficacy—so that practice barely changes at the street level. Alternatively, bottom-up approaches to scale rarely reach the power centers at other levels of the system to ensure broad support and accountability for institutional change. There are also a host of perceived barriers that may sound something like, “So and so would never let us do that.”

To address both real and perceived implementation challenges in the NMP, we are engaging all levels of the system in what we call cycles of mutual permission giving. We started by working at the state level with mathematics faculty, introducing the NMP concepts, gauging faculty interest, and soliciting feedback on who else in their systems we should involve and what else would need to happen to pull off change on the order of magnitude of the NMP.

We then talked to presidents and chief academic officers to convey the interest of their faculty and their suggestions for coordinated action at their levels. Knowing their faculty wouldn’t immediately revolt if they started exploring curriculum reform inspired presidents to explore the NMP further and put policy and resource considerations on the table for discussion. We then shared this information with the faculty. This kind of cycle led all 50 presidents and chancellors to agree to raise their Texas Association of Community College dues, providing seed money for the development of the NMP courses and implementation supports.

Notably, coordinating work across all levels of the system includes not only official system actors but the external players that further legitimize our professional action and define our professional identities, namely, our field’s major associations concerned with mathematics in two-and four-year institutions, community college advocacy networks, and developmental education associations. We need to ask these entities for their permission and invite them into the work. We have found that a willingness not to oppose the innovation is essential, but formal endorsements or partnerships are best.

Coda: Edison famously said, “Genius is 1 percent inspiration and 99 percent perspiration.” And so it is with scaling. Coming up with the idea, relatively speaking, is the easy part. Designing innovations for scale, designing initiatives for scale, and getting permission across levels of higher education ecosystems is the real work.