In 2016, leaked email exchanges from Mount St. Mary’s University revealed then-president Simon Newman’s plan to inflate retention stats by encouraging students with a higher likelihood of dropping out to drop out early. It was a move that reignited the debate about the role of data and intervention in higher education. Newman took to The Washington Post to defend his retention plan as being in the interest of students who might be happier at a local, less expensive institution.
That same academic year, Lone Star College-Montgomery was taking a different approach. Led by vice president of student affairs Michael Chavez, the college was using a predictive analytics platform from Civitas Learning to inform a targeted nudge campaign focused on increasing retention rates. After finding that their first-time, part-time students were seeing double-digit retention rate increases after switching to full time class schedules, Lone Star College-Montgomery implemented nudging interventions. These guided students towards orientation programs, student success courses, and initial advising sessions—increasing the number of first time college students taking more than 12 credits by 31 percent.
These examples are two sides of the same coin: two institutions employing the use of data-informed nudging—the behavioral economics practice of implementing minor design changes that affect individual behavior—to address retention and student success. But only one of these examples represents an ethical use of nudges. Where Lone Star Community College shows us the way that higher education can use data and nudge-based intervention to encourage student success, Newman’s actions at Mount St. Mary’s University show us potential pitfalls colleges and universities can encounter when they are not practicing the deliberate and ethical use of data driven solutions.
Watch the EdSurge on-demand webinar, “How Analytics Can Support Student Success in Higher Ed,” sponsored by Salesforce.org.
It makes sense that colleges are turning to principles of behavioral economics for ideas on how to promote student success. Higher education is in the middle of a price crisis, where diminishing resources have made it difficult for colleges to operate without increasing tuition. As the cost of pursuing a higher education rises, there has been more focus on student success and the consequences of incurring debt and not obtaining a degree. Nudge-based solutions are minor changes, often with small price tags, that can result in big returns in student retention. Furthermore, colleges and universities can use nudges to connect students to underutilized institutional resources, maximizing efficiency.
But nudges require an understanding of how and why people make the decisions they do. As such, many colleges and universities are also employing predictive analytics tools to target their nudge interventions. These tools are using existing data to determine patterns of successful behavior, forecast future outcomes, and inform interventions that put students on the path to success.
Simply having predictive analytics in place does not guarantee that the college is using those predictive tools and deploying nudges and other interventions ethically. Two reports from New America, The Promise and Peril of Predictive Analytics in Higher Education and Predictive Analytics in Higher Education: Five Guiding Practices for Ethical Use, make clear that there are dangers to depending on data alone to inform interventions. In fact, we concluded that designing interventions that are deployed based on predictive analytics is one of the most important factors for making sure these systems are used ethically. And even something as simple as nudging is no exception. Nudges can result in big increases in student success but there are also dangers of using them unethically if colleges don’t focus on design. Some of the dangers are:
1. Decision-making vs. direction-giving
Good nudges guide people toward decisions that work in their long-term benefit. But they should always maintain a person’s freedom of choice and encourage active thought. Nudges should not be declarative. For example, if a college wants to encourage students to stay on track to graduate, they might make it easier to register for classes. But they shouldn’t simply dictate a schedule. Preserving students’ choice of which classes to take is an important ethical consideration.
2. Economic phishing
Good nudges improve the welfare of those being nudged, not those doing the nudging. For instance, a college decides to move to flat-rate tuition charging for 15 credits a semester in an effort to get more students to graduate on time. But what if the college fails to communicate that new default to students? Or doesn’t include an opt out for students who can’t swing 15 credits? They could simply be charging students more for the same class load, which benefits the university but hurts the student.
3. Restricting choice based on bias
One way that nudges can work is by simplifying choices for students. Behavioral psychology has shown that people are easily overwhelmed by too many choices and it can be helpful to simplify the options. But colleges have to be careful not to present simplified choices in a biased way. For instance, if colleges have a system that suggests a few possible majors to students based on data about their preferences and academic performance, they need to ensure that the system is not recommending lower paid, less rigorous majors to historically underserved groups.
4. Structuring nudges
Nudges should be transparent; the person being nudged should know exactly what decisions they are being nudged towards. In an effort to preserve freedom of choice, it should also be easy to opt out of nudges. Lastly, colleges also need to consider that the messages communicated through nudges do not affect all students in the same way. What may come off as a helpful hint to one student may be interpreted by another student as a sign that they don’t belong in college. Messages that are not carefully crafted can have particularly damaging effects on first generation and low-income students.
5. Nudges as the sole solution
Nudges should be supplementary. While nudges represent a low cost way to have a major positive impact on students, colleges still need to support other retention and student success efforts. Nudges without additional supports won’t make significant improvements in completion for low income students.
Using predictive analytics and nudges can have a big positive impacts on student success. But colleges and universities need to use these promising tools with ethics in mind; what happened at Mount St. Mary is an extreme example of how the use of these systems can go wrong. Designing and implementing interventions with care can prevent a nudge from becoming a shove.
