Nightly sleep time and GPA in first-years
Motivation
For many reasons, it’s important for students to succeed during their first year of college. Success during the first year affects whether students will stay to complete the rest of their degree, and college GPA seems to predict future success in jobs. Universities often invest large amounts of time and money to help students adapt to college during their first year, through special programs, tutoring, orientation sessions, and so on.
Besides programs offered by universities, we can also ask what students can do to improve their chances. One potential factor is sleep. College students often have poor sleep habits, staying up late and sleeping short hours, and a great deal of research suggests that lack of sleep can harm cognitive performance. It seems reasonable, then, to study how sleep habits in college students are related to their grade point averages (GPAs).
To assess this, this study recruited first-year students at three universities:
- Carnegie Mellon University (CMU), a STEM-focused private university
- The University of Washington (UW), a large public university
- Notre Dame University (ND), a private Catholic university
Each student participant was given a device to track their sleep patterns for a month early in their spring academic term. The researchers then obtained grade data from the university registrars, along with demographic information about the students. The demographic information included other factors that might be related to the students’ GPAs, such as whether they were first-generation students whose parents never attended college, so these factors could be controlled for.
Data
There were a total of 634 participants in this study, who each received a Fitbit to track their sleep and physical activity. Students’ GPAs were collected from the university registrar. Each data row represents one participant.
Using the Fitbit data, the researchers extracted sleep “episodes”, which are simply time periods where the Fitbit reported the participant as sleeping. They only counted episodes at least 20 minutes long, separated by at least 5 minutes from other episodes. Gaps smaller than 5 minutes were ignored, so if a student slept for several hours, woke for a few minutes, and then slept again, this was counted as a single episode.
The longest sleep episode between noon of one day and noon of the next was considered the “main” episode; anything else was considered “daytime sleep”. For example, if a student slept from 4pm to 5pm, then again from 2am to 8am, the 2-8am window was the “main” episode (6 hours long), and the 4-5pm window was counted as 1 hour of daytime sleep. The beginning of the main episode was counted as the student’s “bedtime”, so in this example, the student’s bedtime was 2am.
The data is divided into five separate studies, by university and semester:
| Study | University | Semester |
|---|---|---|
| 1 | Carnegie Mellon University | Spring 2018 |
| 2 | University of Washington | Spring 2018 |
| 3 | University of Washington | Spring 2019 |
| 4 | Notre Dame University | Spring 2016 |
| 5 | Carnegie Mellon University | Spring 2017 |
Data preview
cmu-sleep.csv
Variable descriptions
| Variable | Description |
|---|---|
| subject_id | Unique ID of the subject |
| study | Study number, as shown in the table above |
| cohort | Codename of the cohort that the subject belongs to |
| demo_race | Binary label for underrepresented and non-underrepresented students (underrepresented = 0, non-underpresented = 1). Students were considered underrepresented if either parent was Black, Hispanic or Latino, Native American, or Pacific Islander. Students were non-underrepresented if neither parent was from an underrepresented category (i.e., both parents had White or Asian ancestry). |
| demo_gender | Gender of the subject (male = 0, female = 1), as reported by their institution. |
| demo_firstgen | First-generation status (non-first gen = 0, first-gen = 1). Students were considered first-generation if neither parent completed any college (i.e., high school diploma or less) |
| bedtime_mssd | Mean successive squared difference of bedtime. This measures bedtime variability, and is calculated as the average of the squared difference of bedtime on consecutive nights. For example, if we had four successive nights of sleep, we calculated bedtime MSSD by computing the average of (night 2 bedtime – night 1 bedtime)^2, (night 3 bedtime – night 2 bedtime)^2, and (night 4 bedtime – night 3 bedtime)^2. Times are in units of hours. Compared to calculating standard deviation, this looks at change from day to day, not change relative to an overall mean |
| TotalSleepTime | Average time in bed (the difference between wake time and bedtime) minus the length of total awake/restlessness in the main sleep episode, in minutes |
| midpoint_sleep | Average midpoint of bedtime and wake time, in minutes after 11 pm (for example, 364 is 5:04 am) |
| frac_nights_with_data | Fraction of nights with captured data for the subject. (Some students may not have worn their Fitbits every night, or the battery may have died.) |
| daytime_sleep | Average sleep time outside of the range of the main sleep episode, including short naps or sleep that occurred during the daytime, in minutes |
| cum_gpa | Cumulative GPA (out of 4.0), for semesters before the one being studied. (Since these students are first-years during their spring semester, this is usually just their fall GPA. UW has quarters, so this includes both fall and winter quarters at UW.) |
| term_gpa | End-of-term GPA (out of 4.0) for the semester being studied |
| term_units | Number of course units carried in the term |
| Zterm_units_ZofZ | Because each university counts units differently, each student’s units were Z-scored relative to the mean and standard deviation of all students in their study cohort. This score represented the student’s load relative to the average amount of units. The Z-scores for the cohorts were then combined and Z-scored again. 0 represents an average load, while positive values are above-average loads and negative values are below average. |
Questions
- Create plots comparing the term GPAs by gender, race, and first generation status, which may all be related to academic success. Then, conduct one-way ANOVA analyses to see whether GPA differs across different groups.
- Create a boxplot comparing the average total sleep time across schools.
- Find the average total sleep time for each school. Conduct a One-Way ANOVA analysis to see whether total sleep time differs across different schools.
- Explore the linear relationship between total sleep time and term GPA. Is the relationship significant, controlling for cumulative GPA?
- A student from CMU has 360 hours of total sleep time. What is his/her predicted GPA?
References
Creswell, J. D., Tumminia, M. J., Price, S., Sefidgar, Y., Cohen, S., Ren, Y., Brown, J., Dey, A. K., Dutcher, J. M., Villalba, D., Mankoff, J., Xu, X., Creswell, K., Doryab, A., Mattingly, S., Striegel, A., Hachen, D., Martinez, G., & Lovett, M. C. (2023). Nightly sleep duration predicts grade point average in the first year of college. Proceedings of the National Academy of Sciences of the United States of America, 120(8), e2209123120. https://doi.org/10.1073/pnas.2209123120.