1. Using Data to Personalize Education

    A couple of weeks ago I was in Las Vegas for the annual Sloan Consortium Emerging Technologies Conference.  And, I have to say, it was fascinating experience.  The theme of the conference was (probably no surprise to those following educational technology trends over the last number of years) the personalization of the educational experience.

    I went to the conference with a particular focus: to explore data analytics and the work coming out of colleges and universities to advance this field.  I know that data analytics hold great promise for K-12 education, and yet I’ve been wrestling with their place within the independent school environment.  For generations now, we have told families: “we know your son/daughter; we know how they learn; we will give them the personalized attention and support that they need to succeed.”  And yet, we can all admit that even with our low student-teacher ratios, our advising systems, and our tight-knit communities, we do still have students who “fall through the cracks” every so often, despite our best efforts.  I went to Vegas to see if there was anything that we could be doing to ensure that students don’t fall through the cracks, but more importantly, to see if there was anything more that we can do to predict which students might be on the verge of falling through the cracks in our classes and prevent that from occurring.  The answer seems to be both yes and no.

    One of the most interesting and anticipated sessions of the conference was on the preliminary findings of the “Predictive Analytics Reporting Framework.”  This project, funded by the Gates Foundation, was formed to try to determine whether there were predictive analytics across schools that could help determine student success in online courses — that is: are there any data points across many schools that can help predict the success of students before success or failure becomes apparent.  The goals of the project are noble: to figure out ways to increase graduation rates, lower drop-out rates, and help students succeed.  The findings so far: while each school might be able to identify predictive analytics for their campuses (and many do), the study has yet to find predictive measures valid across all campuses.  That said, it was also clear to the researchers that different learning environments create important variables on the institutional level.  In other words, student success is greatly impacted by the academic environment created by schools (I am sure a few of you are thinking, “well you didn’t have to go to Las Vegas to understand that, Brad”).  And yet, within schools data points can be developed to understand when a student is going to falter before failure occurs so that there can be an intervention from the start.

    My primary take-aways from the conference focused on what advantages and challenges independent schools have in thinking about using data more effectively to enhance student learning outcomes and greater personalize their educational experience:

    • It was clear to me that independent schools start off in this world of personalization with one incredibly important advantage: we are mission-driven, purpose driven, and an increasingly large number of us are process driven.  We create academic environments that are likely unique to our geographic reach, and, we know, reach a particular set of students well.  Therefore, if we continue to genuinely talk with prospective applicants about fit our environments and communities, and work to make good matches during the admissions process, we have a great advantage in helping students achieve success.
    • We also have done an excellent job of sharing qualitative data in our schools for years.  My experience in independent schools has been that through advisory systems, class deans, classroom teachers, tight-knit communities, and even comment writing, we regularly share information about our students (particularly those at the top or bottom of a class) and try to act on that information quickly.
    • We do not do a good job of using or understanding the quantitative data that we have, or thinking about what data we should be collecting.  It seems to me that this work starts with identifying the key variables for academic success within our own communities.  And, remember, according to at least the preliminary findings from the Gates funded project, in each of our communities, these variable will be different.  This makes sense, right?  In some schools, Model UN might be the most intensive experience that a student can have, whereas in other schools Model UN is a small club.  In some schools, chorus is just a class that students take, whereas in others chorus is constant performances, competitions, and events.  In some schools, AP Art History is the toughest course at the school, whereas at others, it is just another course.  Our schools have different variables.  And, those of us who have been around the same schools for long enough know innately what those variables are (for example, when I was at Holton-Arms, I knew never to allow an advisee to take AP Art History and AP Biology at the same time).  But, we never use the data to back it up.  We never knew what the expected impact would be of a student taking a given schedule, or participating in the fall play, or playing varsity soccer.  But, we have that data.  We know what the GPA of our students is when they participate in certain major events, teams, performances, etc..  We can cross-reference the data, look at historical trends, and make better informed decisions.  We could be helping students and their families be able to make better informed decisions during the advising process, and pre-identify students who will need academic support because of the combination of course work, activities, clubs, and extracurriculars of which they plan on being a part.  Drilling down another level, we might be able to identify particular stress points within our school calendars where students across the board find challenges, and then act accordingly.  And, we might be able to use admissions data more effectively to identify students coming into our program who need more support.

    I know that many of us in the independent school community don’t particularly like to dive into the realm of data.  It’s not why we got into teaching and education, and why we particularly like independent schools focus on the personal and relationships.  And yet, we can all agree that our primary motivation for being in education is to help students learn and grow.  It does seem (and many schools are proving) that data can be part of the equation helping us to get to that goal.  Independent schools have the huge advantage that we are already good at the personal and relationships, by adding some use of data to the equation, we can likely get to the truly personal faster and more effectively than other schools.