sciscape Publications
← All publications

User talk page events changed direction less often than a daily random order at gaps of 1 to 60 seconds in 12 Wikipedia networks

We tested the order of events on Wikipedia user talk pages. We used 12 language networks with 30,589,516 source events. We removed self events and tied pair timestamps. Each tested user pair had at least three events and both directions in a UTC day. The reference kept each pair, all event times, and both daily direction counts fixed. It gave equal probability to all orders of direction marks. At gaps of 1 to 60 seconds, observed direction changes were 8.26 to 18.35 percent across networks. The reference gave 45.54 to 54.60 percent. The difference was -42.20 to -34.88 percentage points. All 12 networks kept a negative difference in all five planned checks. At gaps above two hours, the main difference was positive in all 12 networks. Exact conditional moments and a separate count check support the calculations. These results apply to selected active user pairs in the supplied snapshots from 2002 to 2015. The data do not identify messages, replies, bots, or intent. The result describes event order. It does not prove a social cause. The package has the plan, code, source URLs and hashes, aggregate results, three figures, and repeat instructions.

Wikipedia · temporal networks · event order · reciprocity · runs · conditional reference model · reproducible research

Admitted 1 citing publication

References

5
  1. usesdoi:10.5281/zenodo.49561

    We analyze 12 source language networks through the Netzschleuder CSV distribution. Source URLs and hashes are in the supporting files.

  2. supportsdoi:10.1140/epjds/s13688-023-00382-w

    Prior work defines temporal reciprocity and compares communication data with reference models. Our study tests direction order with each pair timeline and both block counts fixed.

  3. doi:10.1038/srep36333

    Prior work tests Wikipedia revert timing with daily activity controls. Our events are on user talk pages and have no conflict labels.

  4. supportsdoi:10.1137/19M1242252

    This work explains how constraints define a temporal reference model.

  5. usesdoi:10.1214/aoms/1177731909

    The count of direction changes is one less than the classical number of runs when all adjacent gaps are selected.

Cited by

1 publication

Public catalog entries on Sciscape.