Functional data analysis of college students’ sleep patterns and their relationships with academic performance and social networks: A four-year longitudinal study
College students who earn higher grades sleep nearly half an hour longer each night on average, and the size of that sleep advantage fluctuates dramatically across undergraduate years, according to a new longitudinal study published in PLOS One.
Researchers at Temple University tracked 76 undergraduates over four full academic years using Fitbit devices to capture nightly sleep duration, then compared those records against grade point averages and social network data. The study, part of the NetHealth Project, generated 61,225 daily observations in total. On average, each additional point in GPA was associated with 27.4 extra minutes of sleep per night.
But that number masks a more complex story. The relationship between sleep and academic performance followed a U-shaped pattern across the college years, the researchers found. The strongest association appeared during freshman year, when each GPA point corresponded to 54 minutes more sleep. That link weakened substantially by junior year, dropping to just 5 minutes per GPA point, before recovering to 48 minutes per GPA point during senior year.
The method itself is noteworthy. The researchers used functional data analysis, or FDA, a statistical approach that treats sleep as a continuous, dynamic process rather than taking periodic snapshots. Traditional discrete-time methods can miss the evolving relationship between sleep and outcomes as students move through different phases of their college careers.
What they found
Sleep duration increased significantly across undergraduate years overall, though the pattern was punctuated by seasonal fluctuations. Students slept more during academic terms than between them, and the data revealed a gradual upward trend in nightly sleep from freshman through senior year.
The GPA-sleep link was strongest early and late in college, with a marked dip in the middle. The researchers noted that freshman year represents a period of major transition, when students are adapting to new schedules, social environments, and academic demands. Small differences in sleep quantity during this adjustment phase may have outsized effects on academic performance. By junior year, students may have stabilized their routines, reducing the marginal benefit of extra sleep. The rebound in senior year may reflect a period when academic pressure returns in the form of capstone projects, graduate school applications, and career planning.
The study also examined social network size and its relationship with sleep patterns. On this measure, the findings were null. Students with larger social networks tended to sleep slightly less, but the association did not reach statistical significance. The researchers concluded that social network size, at least as measured in this cohort, does not have a meaningful independent relationship with sleep duration.
Why it matters
The findings carry practical implications for the timing of sleep interventions on college campuses. If the sleep-performance link is strongest during freshman and senior years, those may be the most impactful windows for programs aimed at improving student sleep habits. A sleep hygiene workshop delivered at the start of freshman year, for instance, could yield a larger return than the same intervention delivered midway through college.
The use of functional data analysis also represents a methodological advance for sleep research. Many existing studies rely on self-reported sleep data collected at one or two time points, which may miss the continuous, changing nature of sleep patterns. Objective wearable data combined with longitudinal FDA methods could help researchers identify precisely when sleep matters most for specific outcomes.
Limitations
The study has several important limitations. The sample was relatively small at 76 students, all drawn from a single university, which limits generalizability. The research relied on Fitbit devices, which estimate sleep duration through movement and heart rate data and may not be as accurate as polysomnography, the gold standard for sleep measurement. Social network data captured the size of students’ networks but could not account for the quality of those relationships or their emotional impact, both of which may influence sleep. Finally, the observational design cannot establish causality between sleep and academic performance.
Bottom line
Sleep duration is meaningfully tied to academic performance in college, but the relationship is not static. It changes across the undergraduate years in a way that conventional research methods would likely miss. The strongest returns on extra sleep appear during the transition into college and the transition out, suggesting that freshmen and seniors stand to benefit most from sleep-focused interventions. Wearable devices and functional data analysis offer a path toward more precise, individualized understanding of how sleep shapes student outcomes.
Source
Yao Zhao, Haoyu Zhou. “Functional data analysis of college students’ sleep patterns and their relationships with academic performance and social networks: A four-year longitudinal study.” PLOS One 21(7): e0351120 (2026). DOI: 10.1371/journal.pone.0351120. PMID: 42384669.