Chand Rajendra-Nicolucci: Big data is coming to college campuses

Tuesday, September 3, 2019 - 2:25pm


The Michigan Daily

A university is many things to its students — an employer, a healthcare provider, a landlord, a gym, a restaurant, a library, an educator, an ISP, a creditor and more. Each of these roles has data associated with it which, when unified, can provide an extremely detailed picture of students’ lives. In today’s world of big data there is increasing pressure to utilize this information to improve students’ college lives. Will this lead to surveillance, inappropriate interventions and the repackaging of biases? Or could it set the example for how to responsibly and ethically utilize data in the future? 

Just as Amazon, Google and Facebook are creating models to predict our preferences and behavior, universities are increasingly looking to predictive models for help with student performance, well-being and retainment. This is an area called “learning analytics,” and it’s filled with promises to improve student outcomes by analyzing mounds of data. But just as big tech faces a number of issues related to predictive modeling, universities do as well, with perhaps even higher stakes. For example, one could imagine a tool that recommends majors reproducing historical norms, steering women and minorities away from STEM fields. 

Particularly when thinking about recommendations and interventions, schools will need to avoid simply automating historical stereotypes and biases. Additionally, as schools inevitably turn to third-party vendors for help, they will need to be rigorous about access to student data. Companies’ security and authorization practices should be thoroughly vetted and monitored to avoid letting student data fall into the wrong hands. Already, hackers are increasingly targeting educational institutions for their valuable information.

Universities will also have to consider what they collect. To date, higher education has largely avoided the surveillance craze found in K-12 education, but that could easily change. A system could be constructed using data such as visits to the gym, class attendance and browsing history to surface potential “threats” in the student body. Surveillance in a more benign form is already found in many of the “learning management systems” used by schools. These systems often track minutiae such as keystrokes, time spent on assignments, and distance scrolled in the name of improving student performance. Additionally, schools are embedding trackers in recruitment emails that monitor things like whether prospective students clicked any links and how long they read to help gauge “demonstrated interest” in hopes of increasing enrollment rates. These forms of surveillance are the tools of a system that Shoshana Zuboff, an author and retired Harvard Business School professor, has termed “surveillance capitalism.” In something like a “learning management system,” users are monitored for data to fuel the creation of predictive models that are used with subtle cues and rewards to “tune, herd, and condition” behavior. Zuboff argues that this undermines autonomy by encouraging conformity and pushing users toward the system designer’s most desirable outcomes. If universities implement a form of “surveillance capitalism” on their campuses, they could threaten students’ personal and intellectual development by stunting the exploration and growth that is crucial to higher education.

For big data to be sustainable, Mark Zuckerberg's mantra of “move fast and break things” will need to be discarded in favor of discussion, user-centered design and input from a diverse array of voices. Universities are well positioned to take this approach because of the knowledge, inquiry and discussion that is at the core of higher education. By tapping into the multidisciplinary expertise found on their campuses and including experts from non-technical fields such as history, sociology and anthropology, universities can ensure an array of perspectives are considered when designing technical systems. Additionally, universities are more likely to take a user-centric approach because the student, ideally, is whom they aim to serve. Though there may be external pressure from parents, donors and politicians, profit is (typically) not at the center of a university, making user (student) voices more likely to be heard.

Universities are in a unique position. With their wealth of interdisciplinary expertise, tradition of discussion and responsibility to students, not shareholders, they are as well-positioned as any organization to get big data right. If schools commit to transparency, discussion and student input, they can create an environment where students are educated about, and involved in, the utilization of their data. A healthy cycle of proposal, discussion and continuous consideration could emerge. Students, exposed to a transparent development process that respects autonomy and privacy while also considering sociocultural realities, could be empowered to demand the same type of treatment in the rest of their digital lives. Organizations such as non-profits, governments and even businesses would have a model for how to harness big data responsibly and ethically. By empowering students and creating a proven alternative to the big data practices of the tech giants, universities could challenge the inevitability of our every digital interaction being tracked, packaged and sold. Higher education bills itself as an intellectual leader that enables scholars and students to change the world — here’s an opportunity to live up to it.

Chand Rajendra-Nicolucci can be reached at