Author(s): DIEWALD, S., GUZMAN, N., WAGNER, K., RASSKE, K., SUMNER, J., Institution: BROOKS RUNNING, Country: UNITED STATES, Abstract-ID: 2269

INTRODUCTION:
Widespread adoption of wearable technology has enabled the longitudinal tracking of physical activity with unprecedented ecological validity [1]. Although sex and age-related declines in physiological capacity are well-documented in specific cohorts, it remains unclear how these demographic factors explain variability in everyday behaviours of recreational runners [2]. Specifically, the extent to which demographics account for differences in fundamental components of training (volume and intensity) is unknown. The aim of this study was to describe sex- and age-related patterns in outdoor running behaviour and to quantify how much of the variability in key training metrics is explained by these factors.
METHODS:
Data were captured from individuals (aged 19-79 years) with > 25 runs recorded via wearable wrist devices on the DashLX platform between 6 Sept 2024 and 6 Sept 2025. A total of 2,515,687 valid runs (duration > 2 minutes & distance > 100 meters) were recorded by 17,863 (male=7770, female=10,093) individual runners. For each individual, median, capacity (min/max), and variation (IQR) were calculated for distance (km), average pace (min/km), and duration (min). To describe sex- and age-specific patterns in outdoor running behaviour, multiple linear regression models were used, with Bonferroni p-value corrections to account for multiple comparisons.
RESULTS:
Sex, age, and their interaction explained the most variance in maximal (Adj R2=19.3%) and typical median pace (Adj R2=14.4%), but notably less variance in all distance and duration metrics (R2<4%). Men generally exhibited greater typical volume and intensity, with significantly higher typical distances (+1.41 km) and faster typical paces (-1.13 min/km) compared to women. Men also demonstrated significantly higher upper-end capacity, with max distances averaging 6.79 km longer and max durations 27.07 minutes greater than women. Regarding variability (IQR), men showed significantly higher variation in distance (+1.31 km) and duration (+2.57 min), though they exhibited slightly less variation in pace (-0.22 min/km) compared to women. Age acted as a moderator for several metrics; notably, the sex gap in maximum distance and duration narrowed as runners aged (β=-0.11 km/year; β=-0.77 min/year, respectively). Whilst typical pace slowed with age (β=+0.04 min/km per year), this effect was consistent across both sexes.
CONCLUSION:
Sex and age were significant predictors of recreational running behaviour but explained only a modest proportion of variance in real-world training patterns, highlighting substantial individual heterogeneity. These findings suggest that population-level demographic segmentation provide an incomplete framework for understanding training behaviour. As wearable technologies and footwear systems aim to optimise for lifelong performance and experiences, moving beyond broad demographic categories toward individualised (n=1) modelling may provide a more holistic and effective path for personalisation.