Data, Choice and Control: Does Something Have to Give?

At an auto shop years ago, I saw a sign that I’ll never forget. It said, “We do three kinds of service: fast, cheap, and good. You can have any two of the three.”

I think of that sign every time I talk to a district or state CIO. Because so many of them, when it comes to their goals for how data is collected and used in their schools, feel caught between three conflicting priorities: teacher choice, swelling data, and local control.

Of those three priorities, data is the one nobody wants to sacrifice. They’ve heard the success stories— like Tacoma Public Schools, which credits data-driven decision making as a factor in boosting graduation rates by 27 percent in the last five years. They’re confident in the link between using student data and driving student achievement.

So the conflict tends to come down to choice and control. Education leaders can allow teachers to choose which apps and tools they acquire and use, but vendor agreements and piecemeal IT environments mean the district won’t have control of the data. Or they can insist on local control, but that means working with only certain pre-approved tools and depriving teachers of the freedom to choose the ones they want.

Torn like this, leaders try to make the best of an imperfect situation. They simply can’t do everything they want. Something has to give.

Or does it? This dilemma is common all over the country, so it may seem like just a fact of life. But there are states and districts—thousands, in fact—that are finding a better way.

What’s their secret? In a word: integration. They unite all their disparate data sources, getting them to “speak the same language” and feed into one locally controlled repository. Let’s discuss local control first.

There are two benefits to local control: ease of accessing the data (since it’s all under your digital roof) and assurance that you’re handling it responsibly. The first benefit is pretty straightforward—educators and administrators have a much easier time finding the data they need when it’s all wired in together. The issue of stewardship is a bit more complex.

Many districts cede control of some of their data to vendors. It’s a provision in many licensing agreements, stemming from the fact that the vendor tends to be in the best position, in terms of technology, to oversee the data its product uses. While that doesn’t necessarily put the data at risk—no one’s suggesting that vendors are irresponsible or mischievous when it comes to school data—the idea that a district isn’t the one dictating how student data is stored and secured can be very concerning to the outside eye. So, by integrating all your data sources, you put the district in a technological position to oversee its own data. And in a world of stolen identities and data breaches, parents are a lot less concerned when they know the important decisions governing their kids’ information—all of it—are made by the district.

In a way, achieving that control is the easy part. It’s more or less a byproduct of feeding all your data sources into an operational data store (the central repository I mentioned above). The hard part is getting all those data sources, which weren’t necessarily designed to work together, to connect to that repository and report data in the same format as one another.

You might assume that one product records data just like every other product, but that isn’t necessarily the case. Three products utilizing student information can be like a song, a painting and a poem telling different parts of the same story. Looked at individually, we can see what they’re saying and how they go together. But to get the whole story in one telling, you have to invent a method of merging and compiling songs, paintings and poems into one intelligible form.

In education IT, we call that “interoperability.” And to achieve it, you need to standardize the data. That is, you adopt (or create) a set of data definitions, and get all the products in your IT environment to accept and use those definitions. Using an open and widely adopted data standard tends to be your best bet, because then you establish a common language for data and open the doors to teacher choice at the same time. With an open standard defining how data sources work together in a given school, edtech vendors know exactly how to plug their product into the IT environment so that it gels with everything else.

Interoperability isn’t something you achieve overnight. For a mid- to large-size district, it may take one or two years—comparable to the time and effort it takes to install an SIS or other district-wide system. But the benefit of spending time and effort on interoperability is that it makes future installations much, much quicker and easier. Once a district has achieved interoperability, vendors can integrate new products in a matter of weeks or months. Or maybe even hours, if you’re using an open data standard that the vendor has already worked with.

Basically, districts have the choice of putting effort into integrating their IT environments now and saving lots of effort with every adjustment down the line, or keeping their current setup and expending exorbitant effort over and over again every time they add or adjust anything. The math usually works out in favor of integration.

Josh Klein, CIO of Portland Public Schools—whom I had the pleasure of interviewing for a handbook on this subject—really hit the nail on the head:

No one knows a classroom like the teacher who stands in front of it every day. And Josh wants his district to respond to teachers’ needs, quickly and ably. That’s what integration is all about. And actually, rather than coming at the expense of local control, integration puts his district in the position to maintain it.

So the next time you feel you have to choose between teacher choice and local control, think again. You can choose both.

Troy Wheeler (@EdFiAlliance) is President of the Ed-Fi Alliance