Two abstention rules gave the same US versus China vote agreement order in 1,856 of 1,869 country and year records from 2013–2022

We tested how two abstention rules change vote agreement with the United States and China. We used 902 roll calls dated 2013–2022 in Harvard Dataverse release 30.0. Each country and year first had one common set of votes with both reference countries. One rule placed abstention halfway between yes and no. The other removed abstentions for each partner. Of 1,869 records that met fixed count limits, 1,856 kept the same order, 12 had opposite orders, and one changed from a tie. All 12 reversals favored China after removal. None had a margin of five percentage points under both rules. The median US-minus-China margin fell by 9.36 points. An exact identity shows that an order reversal requires a US or China abstention. A separate calculation checked 1,908 nonempty score rows and 51,570 vote-pattern cells. A second run reproduced all 15 prepared and result files exactly. The source has gaps and conflicting records. The findings apply to its selected votes. They do not show a change in political allegiance. The package contains original CC0 data, source hashes, code, audit tables, figures, and reproduction steps.

United Nations · voting agreement · abstention · measurement sensitivity · data audit · collective decisions