English Wikipedia top-list attention became slower, not faster, from 2016 to 2025: longer top-50 stays, lower turnover and slower burst decay in daily user pageview data

Question: did collective attention on English Wikipedia accelerate from 2016 to 2025, as reported for other platforms by Lorenz-Spreen et al. (2019)? Data: the Wikimedia daily list of the 1,000 most-viewed English Wikipedia articles for human (user) traffic, all-access and mobile-web, 2015-07-01 to 2026-10-21 (8.26 million rows), and per-article daily views around 2,750 sampled bursts. Method: we cleaned the lists, removed probable bot spikes, and calculated the mean days in the top N, stay lengths, daily turnover and 30-day novelty for N = 25, 50 and 100. We tested trends on 120 monthly values with month dummies and HAC errors. We measured decay after the peak for 30,240 isolated top-25 bursts. Results: attention became slower, not faster. Mean days in the top 50 increased from 3.29 to 3.71 (+2.1% per year, 95% CI +1.2 to +3.1). Daily top-50 turnover decreased from 46.5% to 39.6% (-2.3% per year, CI -3.2 to -1.5). All 60 trend estimates (5 measures, 3 list sizes, 4 list variants) point to slower attention, and all intervals exclude zero. The median next-day share of peak views increased from 0.30 to 0.42 (Spearman rho with year 0.99). The share of bursts with a half-life of 3 days or more increased from 9% to 17%. The peak-to-baseline size of bursts did not change. Limitations: the result applies only to the most-viewed English Wikipedia articles and to Wikimedia's user traffic class, which changed in 2020. The trend is already present in 2016-2019, and it remains in mobile-web data with few bots. The list measures were slowest in 2022 and partly reversed after 2022. We did not test causes.

collective attention · Wikipedia · pageviews · social acceleration · attention decay · popularity dynamics · turnover · burst dynamics · computational social science