Daily excess adjacent vote agreement stayed near one point with prior voter groups in 2008–2013 Wikipedia records

We tested whether prior voter differences change a reference for adjacent vote agreement. We used Wikipedia requests for administrator status. The test has 66,548 adjacent binary vote pairs from 1,057 candidates in 2008–2013. Prior groups use support rates from 2003–2007 only. Observed agreement was 78.55%. Excess above fixed candidate-day sign counts was 0.95 percentage points. Fixing sign counts by candidate-day and prior voter group gave 0.98 points. The paired change was +0.03 points, with a descriptive 95% candidate resampling interval from -0.01 to +0.07 points. These groups cover only 25.0% of test votes. In a secondary test on the same pairs, hourly sign counts gave 0.26 points of excess. Adding prior groups to hourly counts gave 0.19 points. All planned daily sensitivity settings gave similar results. Independent parsing and counting matched the primary results. Exact small examples checked the reference formula. The result applies to these fixed historical groups and selected records. It does not rule out voter selection or identify social influence. The package has code, provenance, aggregate results, random draws, figures and repeat instructions. Raw source records are external and retain their original terms.

Wikipedia · vote order · collective decisions · historical voter groups · conditional permutation · reproducible research