Observed evidence study

Uncovering Changes in Parkrun Participation and Demographics (2023–2025)

Observed result records

Across the complete years 20232025, observed parkrun result records grew from 7,222,532 to 10,733,029 (+48.6%) across 36,44746,699 event instances. Average field size rose from 198.2 to 229.8, while median finish time moved from 28:41 to 29:57 and the interquartile range widened from 8:34 to 10:23. Composition also shifted: recorded female share moved from 39.94% to 41.58% and the 60-plus age band from 14.49% to 16.98%.

Every Saturday morning, a field gathers at the start line. This study asks a simple question of the records that field leaves behind: has the visible parkrun field changed?

We did not survey anyone, and we did not model anything forward. We counted what the result records actually show — how many people finished, how long they took, and what the source fields recorded about them — and then we set the complete years side by side.

What follows is not a story about parkrun getting better or worse. It is a description of a field that appears to have grown and broadened. The numbers move in one consistent direction, and the honest thing to do is simply to show them.

Evidence snapshot · complete years 2023–2025
Observed result records · 2023 → 2025
+48.6%

Result records grew from 7,222,532 to 10,733,029 across the three complete years — a rise of +48.6%.

Average field size · 2025
229.8

observed finishers per event instance, up from 198.2 in 2023

Median finish time · 2025
29:57

in observed finish-time records, from 28:41 in 2023

60-plus age band · 2025
16.98%

of records, up from 14.49% in 2023 · observed source field

Recorded female share · 2025
41.58%

up from 39.94% in 2023 · observed source field, not verified identity

The story is not that parkrun has become better or worse. It is that the observed field appears to have grown and broadened.
01

The field grew

More records, larger fields

The clearest shift is the simplest one. Across the complete years 2023–2025, observed result records rose by almost half, and the number of event instances grew alongside them. The average field standing at the line got bigger, year on year.

Observed · not predicted
+48.6%
growth in observed result records across the complete years 2023–2025

From 7,222,532 to 10,733,029 result records.

Over the same three complete years, event instances grew from 36,447 to 46,699. These are observed result records — a syncing, partial-coverage sample — not a count of verified unique runners.

Average field size by yearObserved finishers per event instance · complete yearsfield size
Average · 2023
198.2
Average · 2025
229.8
Median · 2023
161
Median · 2025
176
Average field size reached 229.8 in 2025, from 198.2 in 2023. Median field size was 176 in 2025.
02

The field got slower, and more spread out

In observed finish-time records

As the field grew, its observed finish times moved too. The median finish time rose, and the spread between quicker and slower finishers widened. The interquartile range — the gap between the 25th and 75th percentiles — grew by nearly two minutes.

Median finish time by yearObserved finish-time records · complete yearsmm:ss
Median · 2023
28:41
Median · 2025
29:57
IQR · 2023
8:34
IQR · 2025
10:23
Median finish time was 29:57 in 2025, with an IQR of 10:23.

The widening spread shows up at both ends. The slower end of the field grew as a share of finishers, while the quickest end thinned.

Where the field sitsShare of valid finish times · all observed records% of valid times
26:00–29:5924.9% · 8,214,191 records
Under 25:00 · 2023
25.76%
Under 25:00 · 2025
21.29%
60:00 or over · 2023
0.59%
60:00 or over · 2025
1.87%
The 26:00–29:59 band is largest with 8,214,191 records, 24.9% of valid times.
03

The recorded composition shifted

From observed source fields

Alongside size and pace, the source fields recorded a modest change in who the field was made up of. Recorded female share edged up, and the 60-plus age band grew. These are the values the source recorded — observed source fields, not verified person identity.

Recorded sex share · 2025Observed source field, not verified identity% of records
Male 52.86%Female 41.58%Unknown 5.56%
Female · 2023
39.94%
Female · 2025
41.58%
Recorded female share rose to 41.58% in 2025. Recorded male share was 52.86%, unknown 5.56%.
60-plus age band share by yearObserved source field · complete years% of records
60-plus · 2023
14.49%
60-plus · 2025
16.98%
39.36%the 40–59 age band in 2025
The 60-plus age band reached 16.98% in 2025. The 40–59 age band was 39.36%.
Observed source fields, not verified identity. Recorded sex, recorded female share and age band reflect what the source recorded. They are not verified against any register, and this study does not claim to have resolved unique runner identity.

What the data shows

Plain English

Read together, the complete years 2023–2025 point the same way. The honest summary is short.

What the data cannot show

Where we stop

These are observed shifts, not causal effects. The same care that makes the findings trustworthy also marks their edges clearly.

The evidence base

The evidence base · all observed records

34,814,233

observed parkrun result records

160,630

parkrun event instances

1,047

distinct event series

How we counted this

Methodology · in full view
  • Observed records only. We count observed parkrun result records — no surveys, no self-reported intent, nothing modelled forward.
  • Observed sex field. 18,777,064 recorded male, 14,224,949 recorded female, 1,812,220 unknown or blank — an observed source field, not verified identity.
  • Complete-year basis. Year-on-year comparison uses complete years 2023, 2024 and 2025. 2022 H2 and 2026 H1 are partial context only.
  • 94.95% of records carry a valid observed finish time (33,057,601); 5.05% (1,756,632) have a missing or zero time
  • 94.79% of records have age group present; club is present on 27.45% of records
  • 24.9% of valid times fall in the most common band, 26:00–29:59 (8,214,191 records)
All figures are descriptive counts of observed records. Changes are observed shifts, not causal effects — we publish what the records show so you can make your own sense of them. SEL holds a syncing, partial-coverage sample of parkrun results; counts are lower bounds, not the full parkrun record.

Published by Sports Evidence Lab on 9 June 2026 · every figure traces to a recorded query over identity-linked results · descriptive and unadjusted unless stated · observed, not predicted.