Two pupils per class chosen by first-day contact time between classes covered 35.2% of next-day contact time between classes in one school

We tested contact selection with two days from one primary school in Lyon, France. The main sample had 222 pupils in ten classes. Each rule chose two pupils per class from first-day data. Selection by time with other classes covered 35.2% of next-day contact time between classes. Selection by total time covered 27.4%. A contact was covered when at least one endpoint was selected. A class-matched random choice had an exact mean of 17.2%. The observed difference was 7.87 percentage points. In 9,999 pupil-identity permutations within class, the two-sided p value was 0.075. The gain over total time remains uncertain under this reference model. A count of all first-day partners gave 35.8% coverage in a secondary check. Requiring both endpoints to be selected reduced the first two results to 6.50% and 3.01%. The study has one school and one forward day transition. It does not establish a gain for disease control or for a sensor system. The package preserves code, figures, the local analysis plan, source provenance, summary results, all null draws, and reproduction instructions. A source-based check and a full numerical replay passed. The restricted source graph remains external under its source licence.

contact networks · school contacts · chronological validation · class mixing · network sampling · identity permutation · contact coverage