Part V of this week’s Five Principles for Smarter Data Systems series–a guest post from Ben Boer, Senior Policy Associate at Advance Illinois:

Education Sector’s five principles for use of data re-imagine the relationship between data and education. Today, too much good data lives in silos — in state agencies, in districts, in schools, even at vendors. The data in a particular silo serves only the silo’s owner. Obviously, this prevents data from working together to create opportunities for analysis and synthesis. Critically, it prevents much good data from reaching the classroom to help students learn and teachers teach.

Data should be linked to students, enabling innovative new curriculum and allowing for the development of systems that can truly support teachers and administrators, along with parents, students, and community members. That data has to be more user-friendly than a spreadsheet – a truly Learner-Centered, Cross-Institutional system which provides secure but Open access to the Right data to drive instruction. But how do we move from the current closed system, with too little data living in inaccessible silos, to a robust hub of data that be accessed not only by teachers, students and parents, but also those outside the system who are interested in developing applications that can leverage this data? Can we take advantage of the moment when many states are re-thinking their assessment strategy? Can the current collection of technology systems truly handle the multifold increase in data that our systems could soon be producing?

• First, we must consider how assessments can be used to capture data at the instructional level – not just for informing instruction in the moment, but providing data to analyze instruction in more detail

• Second, we must build a technology infrastructure that can support the collection of far more data than we collect currently

The first step is to focus on how implementation of the assessment system impacts the collection of data. Too often, assessments are developed without  the needs of data analysts in mind. Assessments aren’t vertically aligned, course codes aren’t standardized, etc.  There’s little point to collecting cross-institutional data if we can’t see it across institutions.   Optimally, the data and assessment systems would create a feedback loop,  in which assessment captures more data and the data provides for the implementation of more innovative assessment.

Behind this potential vision of a data-assessment partnership is a comprehensive technology strategy that allows the central collection of that data. Few states have sufficient infrastructure to collect the amount of data that will be generated over the next 10 years. Many states do not yet collect transcript information, in particular grades per class, but soon states may feel compelled to go beyond this and collect grades by standard or formative assessment. This order of magnitude increase in data may mean that the current infrastructure needs to be rebuilt, not merely renovated. That’s before we start talking about eTextbooks and click-stream data.  Right now, it can take months to scrub data when it arrives from districts; that process will need to be reduced to hours.

Why do we need all this data? Will it really help our education system? Too often in education we are unable to answer the question of what really drives change. Much of this is due to the limited amount of data we have about actual instruction. Simple experiments such as teacher logs have shown that when we do have more detailed data we can do more powerful analysis. Like other fields, education is making the transition from art to science. Data makes that transition possible.