Five bands from voter history before 2008 left a daily agreement excess of 1.12 percentage points in selected 2008–2013 Wikipedia votes
We tested adjacent support and oppose votes in Wikipedia administrator requests. The sample has 55,273 adjacent pairs from 59,668 votes on 2,848 candidate-days in 2008–2013. We set five voter support-rate bands from records before 2008, with a separate band for insufficient history. Signs shuffled within each candidate-day and band. Observed agreement was 78.471 percent. Its exact expectation was 77.346 percent. The excess was 1.124 percentage points, with a descriptive 95 percent candidate bootstrap interval of 0.872 to 1.386 points. The daily model without history bands gave 1.129 points. Only 24.8 percent of test votes met the history rule of at least 20 earlier votes on other candidates. In an added check of 4,363 pairs where both voters met this rule, the bands reduced the excess from 1.376 to 0.674 points; the latter interval included zero. An hourly model with history bands left 0.097 points, with an interval that also included zero. We report fixed signs and the remaining shuffle freedom. Independent parsing, exact small-case checks, and a numerical replay agreed. An added matched audit found daily excesses of 1.142 and 1.124 points with history ending before 2007 and 2008 on the same pairs. Only 410 pairs met the 20-vote rule in both windows. The contribution is a robustness check of prior history-control work, with code and a record of limited coverage. It does not identify social influence. Raw source records remain external; code, aggregate results, figures, provenance and repeat instructions are preserved.
Wikipedia · sequential voting · voter history · conditional shuffles · social influence · reproducibility
References
7- usesAdjacent binary votes agreed above the expectation from fixed daily counts for 2,290 Wikipedia candidates in 2004–2013 sci:sha256:e75c75325e03664e26662114d75cb1c0e882a039293de65ebd279ce25bec9f3b
We reuse its source selection, adapt its MIT parser and daily test, reproduce its full daily counts, and add fixed bands from earlier voter history.
- usesdoi:10.1162/tacl_a_00184
This is the source citation for the Wikipedia vote data.
- doi:10.1609/icwsm.v4i1.14013
Prior work studies public Wikipedia votes, voter and candidate characteristics, and vote order.
- doi:10.1177/0049124111404820
This work explains why observational social dependence need not identify influence.
- doi:10.48550/arXiv.2609.09321
This preprint studies vote text and sequential evidence in Wikipedia. Our sign test does not test its text measure.
- supportsDaily shuffles with prior support bins leave a 1.00 percentage point agreement excess in 2004–2013 Wikipedia votes sci:sha256:4fb312a5f5dc103778851d9afb6d51aebe804776179fba68015b240128b7443b
Our fixed-history audit is consistent with its limited-coverage warning. We test a 2008 cutoff and add a matched cutoff comparison. Both studies use the same source, so this is not independent data confirmation.
- Adjacent Wikipedia votes had a 1.10 percentage point daily agreement excess under a voter history model in 2004–2013 sci:sha256:3cc6cdf0e4d1cda0077d4834f4d55772278d4b660b6acde6766b88a9453f3e5c
Related work uses conditional voter odds and finds a small pooled change. Our fixed-band control is a different robustness check on the same source.
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.