Listening is a Data Skill
When data people think of the skills required for their job, listening probably isn't the first, or even tenth, thing that comes to mind. We tend to fixate on who knows the fanciest modeling techniques or some other technical part of the job.
The single most important lesson I've learned working in public education is that no amount of sheer quantitative ability is sufficient to make a positive impact on kids' lives. For any calculation I do to make that kind of difference, I need to be able to convince teachers, administrators, parents, and a hundred other types of stakeholders of the information's importance, its centrality to schools' mission, and its ability to make the lives of all these stakeholders easier and more effective.
In other words, to do my job the right way, I need to understand the people who I hope to help.
I'm not talking about learning how to say things in jargon so that you can sell people whatever idea you want. The best uses of data in education I've seen start as projects that educator bring to data analysts. The project may not yet be fully developed. Something along the lines of "we want early-warning indicators that help us know when a kid will be off track to graduate, and we want to know years in advance" will do just fine. A good data person should be able to work with that. We should be able to ask the right questions to translate a general wish into hypotheses that we can test.
At the end of the project, we should be able to tell the educator why what we did is what we thought they want and okay with the possibility that the educator disagrees. We have to accept that educators see the world differently than data people, and we may have lost the intended meaning of their original question.
There are ways to mitigate this possibility, with checking in early and often during the project being the most important. The longer I've spent in public education, the better I've been able to "speak the language" and give educators what they need with less fuss. Still, I embrace the back-and-forth with clients. I show them some preliminary results and say, basically, "Is this what you want?" Inevitably, they see some findings as valuable but have questions about others that require more analysis. I head back to the Batcave and produce some more findings, and the cycle starts again.