Physical Exercise Helps Reduce Academic Procrastination in College Students

How Physical Exercise and Sleep Quality Drive Student Success

Recent research involving 428 university students in Shanghai highlights a clear link between exercise and reduced academic procrastination. The cross-sectional study examined how self-efficacy acts as a mediator in this dynamic. It also investigated sleep quality as a moderating variable.

Participants completed several standardized questionnaires. These included the International Physical Activity Questionnaire – Short Form (IPAQ-S) and the Academic Procrastination Scale. They also filled out the General Self-Efficacy Scale (GSES) and the Pittsburgh Sleep Quality Index (PSQI).

Data analysis was conducted using the PROCESS macro in SPSS 26.0. The team tested a moderated mediation model while controlling for demographic variables. These covariates included sex, age, academic field, year of study, single-child status, and geographical origin.

The results show that physical activity connects negatively and significantly to academic procrastination. The beta coefficient stood at −0.127, with a 95% confidence interval between −0.223 and −0.031. Self-efficacy partially mediated this relationship, accounting for roughly 27% of the total effect. Meanwhile, sleep quality moderated these links, making them stronger in good sleepers.

Rigorous Statistical Testing and Methodological Validation

Researchers thoroughly tested all statistical assumptions before analyzing the main hypotheses. Raw data was cleaned to fix missing or erroneous entries, such as zero durations or overly fast submissions.

The criterion validity of the Chinese IPAQ-S had been previously confirmed against accelerometer data in similar student groups. Because the questionnaire measures a specific behavior rather than a latent attitudinal construct, standard reliability metrics were not applied. Composite reliability (CR) and average variance extracted (AVE) were theoretically non-applicable.

Instead, internal consistency was evaluated using Cronbach's Alpha. It reached an acceptable value of 0.71 for self-report measures. An exploratory factor analysis supported a single-factor solution. It explained 57.8% of the total variance. The Kaiser-Meyer-Olkin (KMO) measure reached 0.722, and Bartlett's test was statistically significant (chi-square = 672.558, p < 0.05).

The distribution of standardized residuals approximated a normal curve. Skewness and kurtosis remained under absolute thresholds of 3 and 10. Residual independence was confirmed via Durbin–Watson statistics between 1.74 and 2.03.

The Mediating Role of Self-Efficacy

No multicollinearity was detected in the data. Variance inflation factors (VIF) ranged from 1.06 to 4.03, while tolerance values exceeded 0.24. Both stayed well within conventional limits. No influential outliers were identified. Cook's distances remained below 0.09, and leverage values stayed beneath calculated thresholds. Studentized residuals showed no absolute values above 3.

Due to mild heteroscedasticity in Models 1, 3, and 4, robust HC3 standard errors were applied. Bootstrapping with 5,000 resamples generated confidence intervals for subsequent analyses. However, the Breusch–Pagan test indicated homoscedasticity for Model 2 (p = 0.093). When academic procrastination was the dependent variable, the overall model was significant (R = 0.312, R2 = 0.098, F = 5.666, p < 0.05).

The mediation analysis utilized PROCESS Model 4, adjusting for all covariates. Bias-corrected 95% confidence intervals were built using 5,000 bootstrap resamples.

The overall association between physical activity and procrastination remained significant after adjustments (beta = −0.174, bootstrap standard error = 0.049, 95% CI [−0.272, −0.077]). When adding self-efficacy as a mediator, the direct effect of physical activity remained significant (beta = −0.127, standard error = 0.049, 95% CI [−0.223, −0.031]). This direct effect represented 72.99% of the total impact.

How Sleep Quality Alters the Psychological Dynamic

The indirect effect transmitted through self-efficacy was also statistically significant (beta = −0.047, standard error = 0.015, 95% CI [−0.078, −0.019]). It accounted for 27.01% of the total effect. These findings confirm partial mediation. Self-efficacy clearly explains part of the academic benefits derived from physical activity.

Researchers applied PROCESS Model 8 to determine whether sleep quality alters these dynamics. This model estimated moderation between exercise and self-efficacy, as well as the direct path to procrastination. Continuous predictors were mean-centered before creating interaction terms.

With self-efficacy as the dependent variable, the overall model explained a significant variance proportion (R2 = 0.108, F = 5.626, p < 0.001). Physical activity showed a significant positive link with self-efficacy (beta = 0.179, t = 3.718, p < 0.001, 95% CI [0.085, 0.273]). Conversely, PSQI scores had a significant negative association (beta = −0.190, t = −4.049, p < 0.001, 95% CI [−0.282, −0.097]).

The interaction between physical activity and PSQI was significantly negative (beta = −0.147, t = −3.169, p < 0.01, 95% CI [−0.236, −0.056]). This reveals that exercise boosts self-efficacy much more effectively in students with low PSQI scores (good sleepers) than in those with high scores (poor sleepers).