In virtually all theoretical, experimental, and practical uses of the Borda rule, inputs are

orders (complete and transitive rankings). Holding the Borda rule fixed, we compare

Order vs. Tournament (all pairwise choices) for sets with 10 alternatives in a

preregistered online experiment (N=366, 61 fixed six-person societies) with full

counterbalancing of materials and task order. Tournament lowers individual error rate by

about 30% and reduces the society’s value-weighted Kemeny distance from members’

true orders by about 8%. These results imply that, for medium-sized sets, pairwise

elicitation is a better input to Borda than full rankings. Methodologically, we introduce a

simple, generalizable procedure for testing elicitation–social rule alignment, enabling

analogous evaluations for other social choice rules.