Intensive longitudinal study of 99 participants finds conditional activity–affect associations that weaken after day-level adjustment

An intensive longitudinal study of 99 participants found that momentary self-rated activity/exertion was associated with higher positive affect and lower negative affect at the next assessment, but the estimates weakened substantially after day-level means were removed. Positive affect also showed a smaller association with the subsequent activity/exertion rating, and that pathway was not distinguishable from zero after person-day demeaning. These results describe conditional next-assessment relationships rather than causal effects.

Participants who did not withdraw contributed 5,850 valid assessments over 14 days, with six prompts per day. At each prompt, they provided single-item ratings for momentary activity/exertion, positive affect, negative affect and fatigue. The activity/exertion rating was anchored by sedentary behavior and maximum exertion; the researchers did not treat it as a verified measure of physical activity behavior. The primary analysis examined 3,472 consecutive same-day assessment pairs recorded 90–240 minutes apart.

Four within-person standardized lagged regressions included a random intercept and independent random slopes for all four predictors. In the primary analysis, momentary activity/exertion was associated with subsequent positive affect (β = 0.162, 95% confidence interval 0.128 to 0.197) and subsequent negative affect (β = −0.157, 95% confidence interval −0.190 to −0.128). Positive affect was associated with the next activity/exertion rating (β = 0.118). After person-day demeaning, the corresponding coefficients were 0.081, −0.074 and 0.007, respectively; the confidence interval for the last estimate included zero.

Confidence intervals were based on 1,000 participant bootstrap resamples, and 16 coefficient tests were adjusted for the false discovery rate. Sensitivity analyses covered exclusion of six participants, interval restrictions, trend adjustment, person-day demeaning and conditional response weighting. Moderation measured with the International Physical Activity Questionnaire–Short Form was exploratory, and no questionnaire-based moderation remained significant after false-discovery-rate correction. Because the lagged coefficients summarize associations across the observed distribution of assessment intervals, they do not represent effects at a fixed elapsed time. The authors concluded that day-level covariation may contribute to the primary pattern and that the strength and within-day temporal interpretation of some pathways remain uncertain; the findings do not establish effects of physical activity behavior, causal mechanisms or intervention targets.

Subject of Research: Temporal associations among momentary self-rated activity/exertion, positive and negative affect, and fatigue in intensive longitudinal data

Article Title: Dynamic interplay between momentary self-rated activity/exertion and affective states: a temporal network analysis using intensive longitudinal data

Article References: Tan, L., Tan, Y., Hu, X., Zhang, W., & Hua, H. (2026). Dynamic interplay between momentary self-rated activity/exertion and affective states: a temporal network analysis using intensive longitudinal data. BMC Psychology. https://doi.org/10.1186/s40359-026-05814-6

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05814-6

Keywords: intensive longitudinal data, momentary self-assessment, affective states, positive affect, negative affect, activity and exertion, lagged regression, temporal network analysis, person-day adjustment, experience sampling, Dynamic, interplay