n paired items 400 (kvfull 400, kvctx 400, newline 400, kvswap 400, paddot 400, padrand 400, mean 400, swap2 400)
kvfull: hop-3 laundered 20.0%, empty hop-3 0, margin 4.562
kvctx: hop-3 laundered 55.5%, empty hop-3 0, margin 6.598
newline: hop-3 laundered 27.8%; - kvctx -27.8 [-32.2, -23.2]; - kvfull +7.8 [+4.2, +11.2]; share 78% [69, 87] -> gives part of the reduction; empty hop-3 7 (share if counted as adopters 73% [64, 83]); margin 4.113, margin share 122% [113, 133]; hop-1 open adoption 7/400
kvswap: hop-3 laundered 21.0%; - kvctx -34.5 [-39.2, -29.8]; - kvfull +1.0 [-2.2, +4.5]; share 97% [88, 107] -> reproduces the latent-step reduction; empty hop-3 0 (share if counted as adopters 97% [88, 107]); margin 4.565, margin share 100% [93, 108]; hop-1 open adoption 16/400
  kvswap: hop-3 = donor's planted config 0/400; same-parity donors 0/400
paddot: hop-3 laundered 54.0%; - kvctx -1.5 [-5.2, +2.2]; - kvfull +34.0 [+29.5, +38.8]; share 4% [-7, 14] -> inconclusive; empty hop-3 0 (share if counted as adopters 4% [-7, 14]); margin 6.126, margin share 23% [15, 31]; hop-1 open adoption 4/400
padrand: hop-3 laundered 9.8%; - kvctx -45.8 [-50.7, -41.0]; - kvfull -10.2 [-14.0, -6.5]; share 129% [117, 143] -> inconclusive; empty hop-3 0 (share if counted as adopters 129% [117, 143]); margin 4.070, margin share 124% [115, 134]; hop-1 open adoption 7/400
mean: hop-3 laundered 17.8%; - kvctx -37.8 [-42.5, -33.0]; - kvfull -2.2 [-5.5, +1.0]; share 106% [97, 117] -> inconclusive; empty hop-3 0 (share if counted as adopters 106% [97, 117]); margin 4.464, margin share 105% [97, 113]; hop-1 open adoption 19/400
swap2: hop-3 laundered 21.0%; - kvctx -34.5 [-39.2, -29.8]; - kvfull +1.0 [-2.0, +4.2]; share 97% [89, 106] -> reproduces the latent-step reduction; empty hop-3 0 (share if counted as adopters 97% [89, 106]); margin 4.578, margin share 99% [92, 106]; hop-1 open adoption 16/400
  swap2: hop-3 = donor's planted config 0/400; same-parity donors 400/400
COMBINED (fillers): report the filler shares as a range; the positions share depends on the filler
SWAP2: the A16 swap result does not depend on the opposite-holder link
