Time-effect relationship between spare-time exercise and sleep quality in middle school student by parallel latent growth and cross-lagged panel model

Wei Zhong, Mo Sha, Zimei Bu, Yujie Liu, Yue Gao, Fan Dong, Shan Jiang, Da HuangFrontiers in Psychology 2025 · DOI

Abstract

This longitudinal study investigates the developmental trajectories and causal relationship between spare-time exercise (STE) and sleep quality (SQ) in secondary school students. Key findings reveal that: (1) Initial levels of STE negatively predict the rate of SQ degradation, suggesting that early physical activity serves as a protective factor. (2) The influence of STE on SQ is unidirectional and stable; STE consistently predicts future SQ, while SQ does not significantly impact future STE. (3) Both exercise levels and sleep quality tend to decline over time, highlighting the critical need for sustained exercise to mitigate sleep debt during high-pressure academic periods.

Introduction

Adolescence is a period of intense physiological change and academic pressure, frequently resulting in a "sleep debt" crisis. While the link between physical activity and sleep is well-documented in cross-sectional research, the long-term causal direction—whether exercise drives sleep or vice versa—remains unclear. This study aims to untangle these developmental trends and causal directions over a one-year period using advanced structural equation modeling.

Methods and Data

The researchers tracked 894 students from six secondary schools in Beijing across three time points in 2024 (March, July, and November). Exercise was measured via the International Physical Activity Questionnaire (IPAQ), and sleep quality via the Pittsburgh Sleep Quality Index (PSQI). Notably, in the PSQI, higher scores indicate poorer sleep quality.

Two sophisticated statistical frameworks were employed:

  1. Parallel Latent Growth Modeling (PLGM): Used to map the starting points (intercepts) and rates of change (slopes) for both behaviors simultaneously.
  2. Cross-Lagged Panel Model (CLPM): Used to determine directional causality by testing if exercise at one time point predicts sleep quality at the next, while controlling for previous levels.
Figure 1: Latent growth model showing the changing trajectories of spare-time exercise (STE) across three time points.

Key Insights

The Unidirectional Driver

The relationship was found to be a one-way street. Spare-time exercise consistently predicted future sleep quality (T1 STE predicted T2 SQ with $\beta = -0.28, P < 0.01$; T2 STE predicted T3 SQ with $\beta = -0.19, P < 0.005$). However, better sleep quality did not lead to increased exercise, suggesting that exercise is a proactive intervention tool rather than a byproduct of being well-rested.

The "Protective Shield" Effect

Students who entered the study with higher levels of exercise saw a significantly slower rate of increase in their PSQI scores ($\beta = -0.35, P < 0.001$). This indicates that establishing exercise habits early on buffers students against the sleep-disrupting stress of the academic year.

The Natural Decline

Both variables showed worrying trends as the year progressed. Exercise levels followed a linear decline, while sleep quality worsened (indicated by a linear increase in PSQI scores). This suggests a common "trap" where students sacrifice physical activity for study time, inadvertently degrading the sleep quality necessary for cognitive function.

Limitations and Future Work

The study's reliance on self-reported surveys (IPAQ and PSQI) introduces potential subjectivity; future research should utilize objective tools like wearable actigraphy. Additionally, the geographic scope was limited to Beijing, and the one-year duration represents the minimum required for latent growth modeling. Longer-term tracking is recommended to see if these trends persist as students transition into higher education environments.

Summary updated

Abstract

(1) Initial levels of STE in secondary school students negatively predicted the rate of development of SQ over the course of the study, and higher levels of STE may have been a protective factor in the development of SQ levels. (2) The effect of STE on SQ was more stable, negatively predicting SQ at the next time point for all three measurements. SQ, on the other hand, had no significant effect on changes in STE. (3) STE decreases over time, so continued exercise and intensity are key to improving SQ.
BibTeX
@inproceedings{zhong2025time,
  author = {Wei Zhong and Mo Sha and Zimei Bu and Yujie Liu and Yue Gao and Fan Dong and Shan Jiang and Da Huang},
  title = {{Time-effect relationship between spare-time exercise and sleep quality in middle school student by parallel latent growth and cross-lagged panel model}},
  booktitle = {Frontiers in Psychology},
  year = {2025},
  doi = {10.3389/fpsyg.2025.1590573},
}
Figure