Vision for Learning Analytics at MSU

Concentric blue, green, salmon and red circles.
By Jeff Grabill, Associate Provost for Teaching, Learning, and Technology, and Director of the Hub

Our vision for learning analytics is to connect institutional and classroom insights to support caring relationships that facilitate learning and eliminate roadblocks to student success

I recently had the opportunity to speak at Colorado State University as part of the Information Science and Technology Center’s distinguished lecture series. I spoke about learning analytics and attempted to connect classroom practice to institutional practice. I want to do the same here. My argument is that a coherent, strategic approach to learning analytics needs to knit together more granular analytics for learning and broader, institutionally focused forms of intelligence. To do so allows us to privilege the classroom and the relationships that are foundational to learning. But this only works if guided by values that ensure that we are focused on student-centric outcomes. The challenge named here is one of the key opportunities for higher education.

Our Approach to Learning Analytics at MSU

We recently began to take this challenge seriously at MSU. Using the Hub for Innovation in Learning and Technology as a location to begin, our current efforts are grounded in the institution’s larger student success initiative. Student success at MSU begins with the understanding that every admitted student has the capacity to graduate in a timely manner. The Learning Analytics Group is in fact a cross-organizational collaboration and builds on existing, previous efforts by many of the collaborators. The Group knits together our Office of Planning and Budgets, Information Technology Services, the Registrar’s Office, the Associate Provost for Undergraduate Education’s Office (APUE), and the Hub for Innovation in Learning and Technology. This Group is led by a faculty member and administrator with joint appointments in APUE and the Hub, Mark Largent (largent@msu.edu).

To maximize the number of students who successfully complete the degrees they choose to pursue at MSU, the Learning Analytics Group examines the institution’s policies, practices, and norms that may undermine student success.

In other words, our goals are to use analytics to

  • Uncover unintended barriers to student success
  • Challenge the myths on which our curricula, our policies, and our practices are based
  • Identify successful interventions, including removing barriers

Each of these goals is animated by a commitment to understand which students face particular challenges and for whom particular interventions work.

We sometimes illustrate the goals of learning analytics and their underlying values by telling stories. For example, we talk about the remarkable and polarizing achievements of Robert Moses, the “master builder” of 20th century New York, particularly the bridges he constructed over the parkways that led to nearby beaches. Those bridges were designed to allow the passage of cars but not busses, explicitly encoding a set of class and race biases to silently and permanently prevent some citizens from accessing public beaches. Sometimes we talk about curbs, which are designed to move water into proper drainage at the boundary of a road. But curbs also can prevent people with mobility issues from finding their way. Curbs aren’t designed to be barriers to mobility, yet they are.

These stories-as-metaphors are effective. They upset people. Some object to thinking of aspects of their university as like the bridges of Robert Moses. They insist that we aren’t intentionally trying to hurt individuals or certain populations of students. That may be true. But intentional or not, we have constructed a number of low bridges for our students. Some are physical. Others can be found in policy. Still others in the curriculum, in how we teach, in how we advise, and in the climate and culture of our institutions. Some we can see, and some, critically, are nearly invisible. Our current learning analytics efforts are focused squarely on identifying the bridges and curbs that get in the way of our students. Our efforts are focused on changing the institution to meet our students where they are and helping them to get to where they want to go.

To the Classroom: Underpersonalized Teaching

There are many ways to locate a learning analytics conversation in classroom contexts, but personalized learning/underpersonalized teaching is probably the best example. Most conversations about personalized or adaptive learning focus on technologies and on providing students with analytically derived interventions of various kinds.

As I’ve noted in the past, the most interesting conversations about personalization or adaptivity happen when we shift from students to teachers and from the use of analytics to facilitate human-to-computer interactions to the use of analytics to facilitate human-to-human interactions. As Michael Feldstein notes, personalized learning “is a set of things that you do, not a set of things that you buy.” The proper focus is pedagogical practice, not technologies. Feldstein and his colleague Phil Hill are correct on two fronts: (1) personalized learning is a set of teaching practices, and (2) the phenomenon it addresses is better thought of as “underpersonalized teaching.”

Analytics and the technologies that provide them are essential, however, if we are to honestly move away from underpersonalized teaching. The most effective personalized teaching asks teachers to see, understand, and act on the information provided to them in a timely and effective fashion.

Put differently, the hard work of personalized learning is in the course and pedagogical reforms that must accompany the move from underpersonalized to personalized teaching. But teachers are not accustomed to having regular formative assessment information from their students at a time when they can act on it. Because of this, teachers aren’t sure what to do with the information when they first see it. Yet, even when teachers start to focus on the fact that this information exists and that they can use it in a systematic way, they often resist using it. To do so means to change the way one teaches. It means teaching differently.  

I get it. It feels uncomfortable. Perhaps most significantly, it challenges our identities as teachers because it changes what we attend to, how we spend our time, and our deeply-ingrained notions of what teachers do. Yet we should dwell in this uncomfortable space in order to make classrooms more impactful and, yes, more caring. Learning analytics can help us identify which students need our attention at any given moment (and which ones don’t). They can help us see thinking and, sometimes, learning. They can provide us with a pattern of engagement that helps us understand the students in our class. Given that the ideal educational intervention has an N of 1, analysis at the classroom level can sometimes help us find that one student so that we can speak with her. As much as it pains me as a teacher to admit, the classroom—my classroom—is sometimes a bridge, sometimes a curb. We can construct with students the forms of mobility they need to pass under bridges and over curbs if we can see them more clearly.

What Holds this Together

A values framework grounded in relationships, focused on student learning and success, and concerned with shaping institutions to value both might just hold together a learning analytics approach that is broad and deep, abstract and concrete. Such a values framework might be constructed around a set of statements like these:

  • Focus on changing institutions, not changing students to meet the needs of institutions.
  • Consider the obligation to use learning analytics to inform and drive necessary changes, from the classroom to the larger institution.
  • Operate, first, on the notion that description may be more important than prediction and that “small data” is just as powerful as “big data.”
  • Commit to evidence-based pedagogical practice not the features of technologies as the driver for learning design decisions.
  • Commit to valuing and supporting innovations in teaching and learning and to supporting colleagues who do the hard work of institutional and personal change.

We are the institution. Collectively faculty, students, administrators, and staff have the agency necessary to create better versions of ourselves. Learning analytics can be powerful if directed at institutional change and at strengthening relationships with students.

Share: