“How could you personalize your music if your Pandora data was lost every time you logged in?”
The question came up during a great conversation with Matthew Rascoff of Wireless Generation about personalizing the educational experience. He was making the point that information about students is lost every time they change schools, making it impossible to use data for personalizing learning the way we personalize our music, video, and reading experiences using Pandora, TiVo, and Kindle.
Personalization is the holy grail of education technology, but it can’t be achieved without mechanisms for rich data about each student’s learning. And that data must be persistently stored and appropriately accessible. Matthew neatly turns the traditional metaphor of a “digital locker” on its head by replacing it with the “data backpack” — a container that goes everywhere the student goes.
Matthew’s insights led me to Wireless Generation’s white paper on “An American Examination System.” The paper outlines a platform for using technology to collect and store data on individual student performance. With this rich data following students, teachers will have the data available to know what students need to work on or learn next. The data can become the basis for “adaptive mass personalization.”
What I particularly like about this vision is that it is designed with nicely partitioned layers: beginning with the technology and protocols for data collection and sharing, moving to a format for encoding a hypothesis about an optimal set of learning trajectories, moving to an automated and rich layer of high cognitive demand tests, and supported by grading mechanisms that are automated where possible, and draw on networks of graders where appropriate.
The idea is that by using technology, standardized tests can be graded by any teacher anywhere, not just a given student’s classroom teacher. With this sort of expert crowd-sourcing, the community of teachers can become much more consistent in their expectations, and minimize the variation in grades that comes from the subjective nature of evaluating creative work.
Because the layers of functionality are segregated, this vision represents a meaningful platform approach to what I consider to be the absolute core issues for reinventing education: testing and assessment.
Standardization is what makes a platform powerful. Railroads are a platform for the transportation of goods and people because all tracks are the same size and all trains can ride on any track. This means there’s an the opportunity for innovation regarding different kinds and models of trains, speed and efficiency improvements, business models for transporting goods, etc. Over time the train tracks and systems have come to support another layer of standardization — containers — that allow goods to be efficiently transported across railways, trucks, and ships. If that second layer of standardization had happened at the same time as the first, the resulting standards would have been far less efficient for the mass movement of goods than those that exist today. If business model competition didn’t happen, the economic environment that made container standardization so powerful might never have developed.
When technology products seek to become platforms, they often look to lock up the value chain from top to bottom by standardizing every layer according to their own product’s protocols. They “open” the top API layer to entice others to choose their platform. I’ve often heard this informally referred to as the “build it and they will come” strategy by those who are skeptical of it, and a “vertical integration” strategy by those who are fans.
In industry this plays out in the debate of “open” vs. “closed.” Is the iPhone closed because of the many layers of the platform Apple controls, or open because of the breadth of innovators who can try their products on top of Apple’s API’s? Android embraces the idea that an architecture standardized at a lower layer will breed more innovation and eventually lead to global domination over the siloed Apple stack. The question of which layer to standardize in a cell phone platform is an open one and the economic implications of choosing the right layer are vast.
In the area of education technology, this question of where to standardize is even more critical. Whereas in industry, competing platforms will eventually answer the question of which approach generates the most innovation and success, in education, early decisions about platforms will become institutionalized and almost impossible to change for decades. This is a critical point to consider as the US Department of Education, governors, and chief state school officers all are investing heavily in a vision of the transformation of schooling through technology.
If standardization happens too far down the stack (say, below the level of student data flow protocols), the market will continue to be fragmented and there will be significant barriers for education innovations to reach students at scale. Innovators will essentially be selling school to school and district by district. If standardization happens too far up the stack, areas that are still immature in the field will become cast in concrete and innovation in those areas will be unlikely.
With a platform structure for education, the key question becomes what to standardize in order to enable wide adoption of innovations, and what things must be deliberately and specifically not standardized in order to allow new innovations to flourish.
Using assessment platforms as an example, the Wireless Generation White Paper highlights a specific theory for students’ learning trajectories in math, yet at the same time provides a structure they refer to as a “honeycomb” for specifying that theory. If the field were mature in its understanding of learning trajectories, the suggested trajectories could be a useful place to standardize. But given that competing learning trajectories are still emerging, the honeycomb structure instead provides a well-defined place for critical innovation.
An example of the honeycomb structure discussed in “An American Examination System” (PDF).
Platforms are valuable precisely because of what they do and do not standardize. Widespread standardization is difficult to accomplish, but critical to realizing the potential of technology in reinventing education. Wireless Generation has an intriguing framework for such a platform (which is probably why it was just acquired by News Corp. for $360 million). With the resources of News Corp. joined with the innovative platform approaches of Wireless Generation, standardization at the right layers is a real and compelling possibility. It will be fascinating to watch this marriage unfold.