DeGroot, M. H. (1974). Reaching a consensus. Journal of the American Statistical Association 69(345): 118–121. doi:10.1080/01621459.1974.10480137.
Read in full on 2026-10-25 from the copy the owner supplied (NOTE-1157). Each member fixes, once, the weights it gives everyone’s opinions, and the group iterates F ← PF. Theorems 1 and 2 say when that converges to a common opinion, and Theorem 3 says the common opinion is the opinion pool weighted by P’s stationary vector. The paper also states that two closed classes, or one periodic class, give no consensus. THEORY-960 re-derives all of it. The model is offered as a formalization of the Delphi technique.
Why this is here
Cited at U36 for the left eigenvector of the influence matrix as the weight each member’s opinion carries in the consensus, which is the readout ℓ·x on which CASE-160 turns the population’s collective coordinate into an Ornstein–Uhlenbeck process. The record reads the model through later work: Golub and Jackson’s naïve learning (LIT-999, THEORY-321), the DeGroot–Friedkin model of Jia et al. (LIT-1146) and Friedkin, Jia and Bullo (LIT-1482). The paper they all start from was not held until the owner supplied it on 2026-10-25.