Abstract
Purpose:
To examine how sleep and physical activity predict body mass index (BMI) in college students.
Design:
Cross-sectional.
Setting:
Medium-sized public university in the Southeastern United States.
Subjects:
386 undergraduate students (245 females; 18-25 years).
Measures:
Surveys included the Pittsburg Sleep Quality Index (PSQI) and Concise Physical Activity Questionnaire (CPAQ). PSQI provided 5 sleep scores: PSQI Global Score, Sleep Quality Factor Score, Sleep Efficiency Factor Score, Sleep Duration, and Habitual Sleep Efficiency. Height and weight measurements were taken to calculate Body Mass Index (BMI).
Analysis:
Correlational analyses were completed first. Linear and moderation regression models using CPAQ as the moderator were used to predict BMI. The Johnson-Neyman technique determined regions of significance where sleep significantly predicted BMI dependent on CPAQ score.
Results:
Sleep Duration significantly predicted BMI (β = -.385, p = .043) while significant interaction terms predicting BMI were found for Global PSQI Score × CPAQ (β = -.103, p = .015) and Sleep Quality Factor Score × CPAQ (β = -.233, p = .013). Johnson-Neyman analyses demonstrated that better sleep quality (measured by Global PSQI and Sleep Quality Factor Scores) predict lower BMI when exercise levels are low and higher BMI when exercise levels are high.
Conclusion:
At low levels of exercise, better sleep quality significantly predicts lower BMI, suggesting that interventions designed to increase sleep quality could promote healthy weight maintenance in college students.
Purpose
Rising rates of obesity and being overweight pose major public health concerns given the numerous negative implications excess weight can have on health. College students are at particular risk for rapid weight gain. Not only do young adults experience more weight gain than other age groups, 1 but behavioral changes that accompany the transition to college can be conducive for gaining weight. 2 Environmental and psychosocial factors contribute to potential weight gain. One obvious factor is exercise where increasing exercise can increase energy expenditure and decrease body weight. Another factor is sleep habits. Poor sleep habits are associated with being overweight. 3 In addition, better sleep is related to improved self-regulation which could help control overeating. 4 Furthermore, there could be an interaction between sleep and physical activity affecting weight gain. 5
The purpose of the current study is to determine how sleep and physical activity predict body mass index (BMI) in college students. We hypothesize that the strength of sleep as a predictor of BMI will differ based on physical activity levels.
Methods
Design
The study design was cross-sectional with participants completing all measures in a single session. Data were collected in a 20 to 30-minute session between 10 AM and 2 PM.
Sample
A total of 386 undergraduate students (245 females, 18.9 ± 1.09) from a medium-sized public university completed the study to earn research credit in an introductory psychology class. Participants used an on-line participant pool to volunteer for and schedule a time to complete the study. All participants completed a screening survey to ensure they were in good mental and physical health. The Institution Review Board approved the protocol and each participant signed an informed consent prior to data collection.
Measures
To assess Body Mass Index (BMI), a commonly used measure to assess body fat, a researcher collected weight and height information for each participant after removal of shoes and coats. Participants completed the Pittsburg Sleep Quality Index (PSQI) and the Concise Physical Activity Questionnaire (CPAQ) using an on-line survey in Qualtrics. Both measures have been shown to be valid 6,7 and access sleep and physical activity, respectively, across the last month. In the current study, Cronbach’s α is .638 for PSQI and .682 for CPAQ.
Analysis
All analyses were completed in SPSS 25. BMI was calculated using the standard formula. PSQI was scored to create 5 sleep variables: global PSQI score (an aggregate sleep quality score), sleep quality factor score (SQ), sleep efficiency factor score (SE), sleep duration factor score (SD), and habitual sleep efficiency score (time asleep divided by total time in bed; HSE). Higher scores for global PSQI, SQ, and SE indicate worse sleep while higher scores on SD and HSE indicate more or better sleep. The CPAQ provided one score with higher numbers indicating more physical activity.
Bivariate Pearson correlations were calculated for the BMI, sleep variables, and CPAQ. Five linear regression models were completed using the sleep variables as predictors for BMI. The PROCESS macro 3.4 for SPSS was used to calculate moderation models predicting BMI using the sleep measures as explanatory variables and CPAQ as the moderator. The explanatory and moderator variables were mean centered and heteroskedasticity-consistent Huber-White standard errors were calculated.
The Johnson-Neyman technique, a validated statistical technique, 8 was used to further examine moderation models with significant moderator beta values. This produced beta values predicting BMI for each sleep measure at specific CPAQ values. The Johnson-Neyman technique provides “regions of significance” where an explanatory variable significantly predicts the response variable, 8 in this case ranges of CPAQ scores where the sleep variables significantly predict BMI.
Results
The means, standard deviations, and Pearson correlations are reported in Table 1. No significant correlations were found between BMI and the sleep variables or CPAQ, although SD approached significance. Correlations among the sleep measures were all significant.
Means, Standard Deviations, and Pearson Correlations for BMI, CPAQ, and Sleep Measures.
N = 386 for all correlations; H. Sleep Efficiency: habitual sleep efficiency.
None of the linear regression models were significant. Only the linear regression model using SD to predict BMI approached significance (β = -.383, p = .053). Table 2 shows the results from the moderation analyses. Only SD significantly predicted BMI as a main effect. However, CPAQ was significant as a moderator variable for global PSQI and SQ. The Johnson-Neyman technique demonstrated that Global PSQI positively predicted BMI when CPAQ was at least 0.99 standard deviations below the sample mean but negatively predicted BMI when CPAQ was at least 2.33 standard deviations above the mean. Similarly, the Johnson-Neyman technique showed that SQ positively predicted BMI when CPAQ was at least 1.40 standard deviations below the mean but negatively predicted BMI when CPAQ was at least 1.60 standard deviations above the mean.
Moderation Regression Analyses predicting BMI.
All explanatory variables are mean centered; H. Sleep Efficiency: habitual sleep efficiency.
Discussion
While sleep was not found to directly predict BMI in the current study, the moderator analyses showed that longer SD significantly predicted lower BMI. In addition, sleep quality, as indicated by global PSQI and SQ, significantly predicted BMI depending on the level of physical activity. At low levels of activity, better sleep quality significantly predicted lower BMI, supporting previous research. 3 However, at high levels of activity, better sleep quality significantly predicted higher BMI. This counterintuitive relationship in students engaging in high levels of activity may be explained by the limitations of the BMI measure. Incongruencies between BMI and body fat have been found in students engaging in high levels of exercise. 9 In addition, better sleep quality is important for muscle recovery and development. 10 This suggests that the current finding does not indicate a relationship between better sleep quality and increased body fat but rather better sleep aiding in the development of healthy, lower-fat mass.
There are several limitations in this study. Although BMI is commonly used as an indication of body fat, it is limited in determining specific body fat percentage. Future studies could use additional measures of body fat such as skinfold calipers or bioelectrical impedance measures. In addition, the current study used a convenience sample of college students. Additional studies with other age groups are needed. A larger number of women completed the study, reflecting the student participant pool of all students taking Introductory Psychology. Future studies could be designed to examine the relationships found here in other populations or populations of equal male and females. Also, the current study used a single data collection point. Future studies could be designed to examine these relationships across time as well as additional covariates that may impact these relationships.
In conclusion, the current study found that increased sleep duration predicts lower BMI in college students supporting the importance of sleep in heathy decision making in students. This study’s finding of physical activity as a moderator of the relationship between sleep quality and BMI in college students indicates that quality sleep may be important both in students who are less physically active as well as those who are more physically active. When working with college students, health professionals could better emphasize the potential protective effects of sleep in terms of better sleep habits as well as contributing to the maintenance of healthy BMI levels.
So What: Implications for Health Promotion Practitioners and Research
What is Already Known on This Topic?
Poor sleep is related to increased BMI. In addition, better sleep is related to increased exercise. However, little research has incorporated the role of physical activity into the relationship between sleep and weight.
What Does This Article Add?
This paper provides empirical support for an interaction of sleep and exercise that affects BMI, mainly that sleep quality predicts BMI differently at different levels of physical activity in college students. Specifically, better sleep quality was associated with lower BMI at low levels of physical activity but associated with higher BMI at high levels of physical activity.
What Are the Implications for Health Promotion Practice and Research?
University health promotion services could use this study’s findings to justify promotion of sleep quality targeted at students who exercise less than their peers. This study encourages further research to investigate the interaction of sleep and physical activity as it relates to BMI in the broader population.
Footnotes
Acknowledgments
We thank Tyler J. Holt, Drew M. Morris, and Emily M. Smith for their assistance with data gathering and data management.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Ethics Approval
Clemson University Office of Research Compliance; IRB number: IRB2017-345.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding was provided by the Clemson University Creative Inquiry and Undergraduate Research program.
