09/09/2014 11:15 am ET Updated Dec 06, 2017

How Do You Develop a Scalable STEM Education Model That Nurtures Deep Learning?

I was very intrigued by Alex Hernandez's post about scalable K-12 solutions and wanted to share some of our own hard won lessons in trying to build a scalable STEM education solution.

Here are the constraints that the solution needs to be engineered for:

  • It should be scalable i.e. it should be pretty clear what needs to be adjusted to support another million users - and how.
  • It should inspire the learner -- not just drill and kill.
  • It should support deep, meaningful learning (not just teaching to the test) in Physics and Engineering. These are complex fields requiring much support and generally there is a lack of teachers who have been trained in these fields.
  • It should be able to support all types of learners - not just the initiated.
I came to education as an engineer and started to list all the possible variables in the right hand side of this equation. I built a team with some of the most passionate and brilliant folks, each bringing their own very different perspective. Together we tried the following:
  • after-school programs for high school students
  • 2 hour workshops
  • 5 week programs
  • 16- week programs
  • 3 month courses teaching girls to program
  • programs for PreK children
  • programs only for girls
  • programs only for mothers
  • in-school classes
  • programs at science centers
  • programs at museums
  • summer camps
  • after-school programs for middle school students
  • professional development sessions for teachers
  • professional development sessions for librarians, other after-school educators
  • programs for parents
  • programs for engineers, scientists, university students
  • MOOCs
  • mobile apps and games
  • we even published books
That took 7 years and roughly 180,000 hours of work (91 years if it had only been me :)

What we learned was this:

  • Pure technology solutions are scalable, but they do not address the constraints of this problem. They may not be accessible or interesting or inspiring to the uninitiated. But technology can amplify an effect. And it can be harnessed for evaluation and quality control as the solution scales up.
  • People are the main driving force behind any meaningful solution. The hard part is to figure out how to retain them in the system - for sustainability.
  • To enable deep learning, conceptual knowledge gains in Physics and Engineering, it isn't enough to swoop in at middle school and "intervene". You have to start much earlier.
  • You cannot also get away by training just one set of influencers (such as teachers). Children spend approximately 20% of their waking time in school. 
There is no quick solution to complex, messy, social problems that involve people. Just as you cannot expect to become a master swimmer or pianist in a few years, similarly for deep learning you have to be ready to devote significant time, resources and patience.
All these lessons were hard earned and painful because we didn't really want to hear what we were hearing. I was hoping for quick glory. "Engineer figures out education problem that educators have been trying to solve for decades".

Well.. quick solutions are neither meaningful nor deeply satisfying.

What we have developed now (after 91 collective years of work) is the following model that is a combination of a hundred 1% great solutions :):

  • VOLUNTEERS NEED TO SEE PROFESSIONAL GROWTH: We train engineers and scientists to develop open-ended, hands-on design challenges that illustrate core Physics concepts in the work they do day to day. They then support local families to conduct these design challenge in person. The training and teaching experience helps them develop their own job-related skills and thus they remain involved with us over many years. This strategy brings in valuable expert social capital into the system and is able to retain it in the collective system.
  • HARNESS TECHNOLOGY FOR SCALE: We create high quality videos of the scientists and engineers explaining their work and publish these videos and the design challenges they created on an online curriculum platform. This is the scale and efficiency part of the model that is enabled by inserting technology into the system at just the right point. This is also the part that inspires a sense of wonder about the world.
  • EMPOWER STAKEHOLDERS TO CO-INVEST: We lean heavily on a child's main caregiver and champion - her parent. We educate parents to support their children in developing curiosity, creativity, persistence, being able to apply the engineering design process and deepening conceptual knowledge. We train parents on using the online curriculum platform so that they have access to all the knowledge and resources they need. We invite committed parents to co-invest in the program and make it sustainable in their schools (like Habitat for Humanity).
  • We train educators (teachers, librarians, after-school educators and other parents) to use the online curriculum with their students.
  • HARNESS TECHNOLOGY FOR DEPTH:And finally, depth comes from enabling the scientists and engineers to provide feedback to each child's project via the online platform. Thus a parent or educator doesn't need to have all the answers, but can lean on expert mentors to scaffold learning for the child as she grows.

It is definitely not an easy, one-line solution - especially when you consider that we start engaging families as early as PreK. But the main insight was that people provide and enable scale. The need for high quality STEM education is urgent and huge, and STEM experts and parents can fill this gap with the right training and access to knowledge. It can be done at very large scale with the right technological tools.

Lets start.