Most 10 minute triangle occurrences have no common 20 second contact slot in six sensor datasets

We tested whether triangles in coarse contact networks share an original contact slot. We used all six contact networks in Zenodo version 2540795. A triangle occurrence is a set of three people with all three pair contacts in one window. It has simultaneous support if all three pairs share at least one original 20 second slot. In 10 minute windows at zero offset, 54.7% to 80.9% of occurrences lacked this support. The share stayed above 50% in every dataset at all 30 possible offsets on the 20 second grid. Requiring at least 60 recorded seconds on each edge reduced the unsupported share to 12.4% to 35.1%. This rule retained only 32.2% to 47.7% of baseline supported occurrences. Observed supported counts were 2.55 to 4.81 times the expected counts from independent slot placement with fixed edge durations. This limited reference does not preserve node attendance or contact persistence. The study measures recorded support, not verified group conversations. The six records do not establish a population rate or a causal effect. A separate count method passed checks on 24 real windows. A full repeat matched all six numerical files. The package includes source code, aggregate results, figures, source hashes, retrieval code, and reproduction instructions. Raw data remain at the source under CC BY-NC 4.0. The general risk from time aggregation is known; this contribution measures its size and the cost of a duration filter in these fixed records.

temporal networks · human contact · triangles · time aggregation · measurement · higher order interactions · SocioPatterns · reproducible benchmark