A group of Dartmouth computer science faculty and students have been using a smartphone app to track student behavior and predict their academic performance:

The StudentLife app that ran on students’ phones automatically measured the following human behaviors 24/7 without any user interaction:
•bed time, wake up time and sleep duration
•the number of conversations and duration of each conversation per day
•physical activity (walking, sitting, running, standing)
•where they were located and who long they stayed there (i.e., dorm, class, party, gym)
•the number of people around a student through the day
•outdoor and indoor (in campus buildings) mobility
•stress level through the day, across the week and term
•positive affect (how good they felt about themselves)
•eating habits (where and when they ate)
•app usage
•in-situ comments on campus and national events: dimension protest, cancelled classes; Boston bombing.

The collected data provides a number of insights into the lives of Dartmouth students — results that I think would be paralleled if a similar study were performed on Williams students. Attracting the most attention (e.g., this NPR story) so far are their findings (both surprising and unsurprising) about GPA, published in “SmartGPA: How Smartphones Can Assess and Predict Academic Performance of College Students,” and to be presented at an upcoming academic conference on ubiquitous computing:

our results suggest that students who change their night time socializing durations later in the term performed better, compared to those who change their night time socializing earlier in the term. Additionally, students who decrease their evening socializing durations during the term perform better, compared to students who increase their evening socializing durations during the term. We suspect that these students may be preparing for their examinations and focusing on other tasks during the evening (e.g., studying), which could contribute to the observed decreases in ambient conversation duration…

[S]tudents with longer average study durations had higher GPAs at the end of the term, compared to students with shorter study durations. This finding is consistent
with research that found academic-related skills (e.g., study skills and habits) to be associated with higher GPAs. Our results extend this work by going beyond self-reported
study habits to show that unobtrusively measured studying habits (e.g., via WiFi and GPS) can also predict student performance. In contrast to previous research, we did not find class attendance to be a significant predictor of performance, and we did not observe simple correlations between class attendance and GPAs as other studies have suggested.

This study has significant limits — the data set is 30 students, over a 10-week period, and the paper doesn’t describe how those students were recruited, and may or may not suffer from problems related to participants’ awareness that their smart phones were tracking their movements, conversations, etc. Yet it’s interesting and novel research that I hope to look at more closely in the future.

If smart phone tracking of student behavior can be used to predict the likelihood of academic success, might Williams and other schools wrestling with how to help at-risk students succeed find a way to use such tracking in real-time? Students would have to be willing to surrender their privacy, but being able to detect changes in behavior and activity for at-risk students could enable early interventions by a support structure, whether peer-based or institutional, that could yield tremendous benefits.

Print  •  Email