Hourly Wikipedia views in 26 European editions moved by 78–85% of the daylight saving clock change on weekdays and by 49% on autumn weekend mornings, 2015–2026

Question: when clocks change for daylight saving time (DST), does daily human activity move by the full hour? Approach: we used hourly reader counts (Wikimedia API, user agent) of 26 Wikipedia editions whose readers live mostly in EU-rule DST countries, at 21 clock changes each (2015–2026; 546 edition-changes). For each day, we measured the time position of the morning rise (05:00–12:00) and the evening fall (19:00–03:00) by log-profile matching with a free scale. Planned simulations showed that this method, unlike mean-crossing times, does not change when only the activity level changes. A step model with trend and day-of-week effects gave the shift fraction S (1 = follows the clock, 0 = no change in UTC). Results (mean over editions, 95% bootstrap interval): on weekdays, S was 0.85 (0.82–0.89) spring and 0.85 (0.78–0.92) autumn for the morning, and 0.78 (0.72–0.85) and 0.82 (0.74–0.92) for the evening. The morning moved more than the evening in spring only (+0.07, 20/26 editions). On autumn weekends the morning moved only 0.49 (0.41–0.56; 25/26 editions below weekdays) and then followed sunrise (exploratory sun-tracking coefficient 0.34). Desktop mornings moved 0.95 in spring. There was no clear adaptation from week 1 to weeks 2–4. Twelve editions without DST gave S from −0.03 to 0.06 on the same dates. Turkish Wikipedia gave 0.41–0.83 at its last real changes and 0.08–0.15 on the same dates after Turkey stopped DST in 2016. Readers outside DST regions did not explain the partial shift (fit at full EU-rule view share: 0.76 morning, 0.77 evening). Limitations: aggregate reading data, not sleep; fake-date placebos near school breaks gave S up to 0.2, so calendar events can bias estimates; two plan amendments were made after inspection of German data; the editions are not independent. Code, plan, amendments, all fits and the retained raw data are provided.

daylight saving time · circadian rhythm · social jet lag · Wikipedia pageviews · collective behavior · natural experiment · clock time · sun time · online activity · time zones