Some social problems are so complex that they feel impossible to solve. That seeming intractability arises from our inability to understand systems in all their complexity—to identify the scope of a problem, and the short- and long-term benefits of one solution versus another. If we look at a problem like gun violence, for example, we see myriad causes all tangled up: the Second Amendment, the gun lobby, political fundraising, and lackluster voting on gun issues, to name just a few. More often than not, when we try to untangle the mess, we either throw up our hands in defeat or myopically focus on whatever narrow problem we have a chance of affecting. Neither leads to progress; instead, they fuel discontent and apathy.

But at their core, complex systems are about connections: Who depends on whom? And what causes what? One way to understand these connections is through structural network analysis, which looks at the basic structures and patterns of a system and asks how any one root cause of a problem affects another. Structural network analysis teaches that the secret to simplifying complex systems is to recognize that while everything is connected, those connections are not random. And the more connected a node is to other nodes, the greater the chance that positively affecting that first node will improve the system as a whole.

So what if there were a way to model, map, analyze, and therefore make progress on big, complex social problems? What if we approach these complex systems as networks and focus on gleaning insights from their structure about the dynamics at play within them?

We’re already proving it’s possible, starting with STEM education in the United States. With 4 million teachers and 50 million students attending 100,000 public and public charter schools, and with federal, state, and local laws and regulations layered on top of deep cultural expectations and norms, the American education system is thorny, vast, and uniquely complex. As with any complex problem, there are numerous issues at play. One of the most fundamental is the shortage of teachers in America’s classrooms, especially those qualified to teach STEM subjects. As a recent New York Times article put it, “A country’s proficiency in science, technology, engineering and mathematics is vital in generating economic growth, advancing scientific innovation and creating good jobs.” Yet America lacks the number of qualified teachers it needs to pass along these skills and mindsets.

100Kin10 used a blend of tools to map the STEM education ecosystem and identify the grand challenges facing STEM teaching today. (Image by Eric Berlow, 100Kin10)

Six years ago, 100Kin10, a nationwide effort focused on addressing why it’s so hard to get and keep great teachers, especially in STEM, set out to add 100,000 excellent STEM teachers to America’s K-12 classrooms by 2021. To reach the 100,000 goal, 100Kin10 mobilized hundreds of independent organizations, including universities, nonprofits, and government agencies. Each of these organization made a commitment to help achieve the goal, and the resulting network is on track to hit the 100,000 STEM teachers mark on time, by 2021.

But would just increasing the number of teachers be enough? Infusing 100,000 teachers into the nation’s schools would certainly make a difference for this generation of students. But without addressing the underlying reasons that not enough people want to become teachers or stay in the classroom, especially in STEM, it would do little to help future generations. Without figuring out how to address those underlying challenges—things like the low prestige of teaching, few opportunities for career advancement short of leaving the classroom, and little time for collaboration or professional development during the workday—there would need to be a 100Kin10-like effort every 10 years, running in endless, Sisyphean cycles.

Bringing in network science

The first step was to name the core problem. In other countries, at other times, it wasn’t so hard to get and keep great teachers, in STEM or any other field. What is true now doesn’t have to be. We realized we had engineered our way to a broken status quo, which meant we could engineer our way out of it.

But as Albert Einstein once famously quipped, “We can’t solve problems by using the same kind of thinking used when creating them.” We knew we needed to examine—from every angle—why it’s so hard to get and keep great teachers, especially in STEM. Without seeing the whole, those of us working on the problem were inevitably missing important elements, likely explaining why million-dollar programs so rarely resulted in field-level solutions.

But seeing the whole picture is hard: Neuroscientists, biologists, and behavioral economists have shown that we’re evolutionarily hard-wired to confirm perspectives we already hold and ignore those that differ. Scientists call these phenomena “confirmation bias” and “change blindness,” and everyone experiences them to some extent. Just as ecologists have to understand all the components of an ecosystem before they can implement a new conservation initiative, we had to understand all the components of the STEM teacher system before we could determine how best to attract people to and retain them in the profession.

We started by mining the expertise of our diverse network. Using good old-fashioned methods like interviews, workshops, and surveys, we polled thousands of teachers, principals, professors, mentor teachers, and others, asking, “Why is it so hard to get and keep great teachers, especially in STEM?” Taking inspiration from the Toyota 5 Why’s, we kept asking “why” until we hit bedrock root causes, and no new major reasons appeared. This generated a list of more than 100 root causes standing in the way of a healthy STEM teaching ecosystem, including lack of access to culturally relevant STEM curricula, the low number of teacher preparation faculty with expertise in elementary STEM, the time for teachers to collaborate during the school day, and school leaders’ responsibility for creating a positive work environment.

We grouped these obstacles into seven “grand challenges” to STEM teaching today:

  1. The perceived lack of prestige of the profession
  2. Inconsistent teacher preparation for the classroom
  3. Lack of support for STEM in elementary schools
  4. Limited professional growth throughout a teaching career
  5. Few opportunities for leadership and experimentation in the classroom
  6. Little or no access to high-quality instructional materials
  7. Lack of support among school leaders and parents for science, technology, and engineering, in particular

To examine the interplay between these challenges, we adapted a blend of tools, including system mapping and network analysis, which are more common in fields like ecology. We partnered with Jeff Mohr at the data visualization company Kumu and Brandon Barnett, director of corporate strategy at Intel, to customize a game in which players consider 2 of the 100 root causes, selected at random, and suggest how one affects the other. We then asked: Would improving the first make the second better, worse, or keep it the same? We turned to our network again, and more than 750 teachers, principals, university professors, deans, and others played. We were looking not for people’s preconceived ideas of importance, but rather for their unadorned understanding of consequences, in order to identify the root causes that would have a positive effect on the most other root causes. To share a timely example of a question, “If the number of people who perceive PK-12 STEM teaching as an intellectually rigorous career increases, what happens to the teachers’ salaries?”

When ecologists do a similar exercise in the natural world, they call the highest-leverage nodes the “keystone species”—the lynchpin to everything working right. As we set out to find our own keystone species, it was the first such full-scale effort for a social-sector problem.

In just a few months, we amassed more than 35,000 votes on what affects what, showing us how the US education system’s problems are interconnected. This process allowed us to see STEM education’s network structure, including all the system’s problems and the inter-dependencies between them, and to identify STEM education’s keystone species.

Examining the STEM education ecosystem

In STEM education, like in all ecosystems, everything is connected, so no one program will successfully address why it is so hard to get and keep great teachers, especially in STEM. But everything is not equally connected, so addressing some issues will have a more far-reaching effect than addressing others. This might seem intuitive, yet it rarely guides how we try to solve social challenges.

To our surprise, we found that each of the grand challenges has at least one root cause with outsized potential to improve the larger system and move closer to getting and keeping great STEM teachers. Top opportunities include:

  • Prestige: scholarships or loan forgiveness for STEM undergraduates who become STEM teachers
  • Teacher preparation: statewide tracking of STEM teacher supply and demand
  • Elementary STEM teachers: teacher preparation faculty with specific expertise in elementary STEM
  • Professional growth: greater teacher collaboration and professional development during the school day
  • Teacher leadership: school leaders who create positive work environments
  • Instructional materials: districts’ identification of high-quality engineering curriculum
  • Valuing science, tech, and engineering: a greater number and range of STEM courses required in high schools

This network approach to problem-mapping allowed us to synthesize thousands of diverse and expert perspectives into a coherent and actionable whole, a map of our collective understanding of consequences. The map points to a new model for social change: Rather than address symptoms, we can use the network insights to diagnose and treat root causes. Rather than hundreds of separate projects working independently, we can leverage the commitment, trust, and learning of the network to focus on the highest-leverage problems.

For instance, each of the 280 leading organizations that are part of 100Kin10, across the education, nonprofit, and corporate spheres (organizations such as AT&T, the National Science Teachers Association, UTeach at the University of Texas, and the federal Department of Energy) indicated which of the root causes they focused on. We made that data public so that any organization could find allies and collaborate. We also commissioned research by leading scholars and doctoral students and identified places where we were already making progress, creating a public clearinghouse of ideas and solutions for everyone to use.

And in 2018, 100 individuals self-organized into “project teams” to work on 13 different root causes, including the number of teachers prepared to meet the unique challenges of rural schools (solution: supporting STEM education in rural communities) and the number of people perceiving STEM teaching as an intellectually rigorous career (solution: improving professors’ perception of the STEM teaching profession). These teams work together in six-month sprints toward a specific goal that will contribute to solving the underlying root cause. The rural team is currently drafting a toolkit on engaging rural communities in STEM, complete with successful stories of STEM learning in rural contexts. The perceptions team is producing a booklet with evidence-based suggestions for how to enlist undergraduate professors (surprisingly powerful influencers on STEM undergrads) to encourage STEM teaching.

100Kin10 is building on pioneering work by others who have taken steps to apply ecology and network science to social problems. We the Data, a collaboration between Intel Labs and a group of TED fellows, presciently employed a similar approach to the question of personal data and privacy, and the Hewlett Foundation partnered with Kumu to map Congressional gridlock. Structural network analysis has also been successfully applied to understand the resilience of communities to climate change. In each of these cases, the groups working on these issues knew they had to build a shared understanding of the problem and map its ecosystem before they could begin making meaningful change.  

This approach—examining problems as networks, revealing the 360-degree view, recognizing that all connections are not created equal, and focusing on the specific areas with the greatest potential to drive real change—is ripe for addressing our most complex challenges. And when we do, we take networked impact to a whole new level, because the multi-faceted challenges we face require a network that can take action in a way that addresses complexity. It’s the only shot we have at making lasting change.