Physical activity and mental health: Comparing between-person and within-person associations in longitudinal analysis

Background

Extensive research has consistently highlighted the strong association between physical activity and mental health. But few have examined how this association may differ at the between-person and the within-person level. This longitudinal study compares how within-person and between-person changes in physical activity are associated mental health during the COVID-19 pandemic among the American population.

Methods

We analyzed data from the Understanding Coronavirus in America (UCA) survey, a probability-based online panel of American adults. An average of more than 6000 panel members repeatedly participated across 33 waves from the start of the pandemic in March 2020 to January 2023. Measures of physical activity include 1) number of days exercised in the past week (0–7), and 2) whether gone outside to walk, hike, and exercise in the past week (0 = no, 1 = yes), capturing both the frequency and the context (e.g., outdoor) of physical activity. Mental health is indicated using the four-item Patient Health Questionnaire (PHQ-4). We used multilevel Within-Between Random Effects models (REWB) that provide coefficients for both between-individual as well as within-individual associations.

Results

We reported three main findings. First, both within-person and between-person increases in physical activity are significantly associated with lower levels of anxiety and depression regardless of the measure. Second, between-person changes in physical activity are more strongly associated with mental health than within-person changes. Third, a large share of the between-person association (about 32%) can be attributed to structural inequalities captured by status markers such as gender, race, and SES.

Conclusions

Our findings indicate the value of physical activity as a mental health intervention and highlight that both between-individual inequality in physical activity, as well as within-individual changes in physical activity, should be considered. The finding of a stronger between-person association that is largely related to status markers such as gender, race, and SES can be used to develop targeted intervention and prevention strategies to improve mental health through physical activity among the most vulnerable populations.