When Clubs Turn Up Together, Race Results Do Not Simply Get Faster

When a Club Appears in Numbers

When a running club appears in numbers at a race, it is natural to wonder whether those records land faster than clubs represented only once. The observed record is more measured.

This is a descriptive, record-level view. It reads the club field that some race results carry, counts how many finishers at an event share the same observed club string, and compares the resulting groups within the same event. It does not track people over time, and it does not test cause. What follows is what the records show, non-causal and unadjusted.

Coverage and Cohort Readiness

SEL has observed 527,602 non-parkrun race result records across 1,146 event instances (771 series) between 2021-10-24 and 2026-05-12; 527,599 (100%) carry a valid finish time. A club string is present on 96,779 records (18.34%), of which 96,776 are analysable, a valid finish time and a club string, for the comparison.

The club field is optional, and its availability has shifted as the observed record has grown. Earlier, smaller years carry a club string on most records. The large recent years carry one on a smaller share.

The Primary Within-Event Comparison

At the primary threshold, where the same observed club appears on at least five records at an event instance, 523 event instances hold both a mass-entry cohort and a solo-representative cohort.

Mass-entry records had the faster within-event median in 161 of 523 instances (30.78%). Across the full set of qualifying instances, the median within-event difference was +3:22, with the solo-representative cohort faster on that measure.

What We Found

Sensitivity by Threshold

Changing the threshold changes which events qualify, but not the overall shape. At N ≥ 3, 767 instances qualify and mass-entry records had the faster within-event median in 29.6% of them. At N ≥ 5, 523 instances qualify and the share is 30.78%. At N ≥ 10, 215 instances qualify and the share is 35.35%. The mass-entry-faster share rises as the threshold rises, yet stays a minority at every threshold tested.

The median within-event difference holds steady: +3:22 (solo faster) at N ≥ 3 and N ≥ 5, and +3:00 (solo faster) at N ≥ 10.

Instance-level results varied across races, with some qualifying instances favouring larger same-club groups and others favouring solo-representative records.

Exclusions Applied

• 39,328,327 parkrun records excluded (approved design); parkrun appears here only as an exclusion count.
• 3 records excluded for a missing or zero finish time.
• 430,823 valid-time records carry no club string and are excluded from the cohort comparison only.

What This Study Can and Cannot Say

In these observed non-parkrun race results, at the primary threshold (≥5 from the same observed club at an event instance) mass-entry cohorts had the faster within-event median in 161 of 523 instances (30.78%). The median within-event difference is +3:22, indicating the solo-representative cohort's median was typically faster. Sensitivity checks show mass-entry-faster shares of 29.6% (N ≥ 3), 30.78% (N ≥ 5) and 35.35% (N ≥ 10).

These are descriptive, within-event comparisons of observed record groups. They do not establish causal mass-entry effects. Results are unadjusted for age, sex or ability mix, depend on an optional club field present on 18.34% of non-parkrun records, and club strings are not verified membership. The study counts result records, not unique runners, and does not track individuals over time.