a. relative contribution of the factors scaled to 100% (nb = number, TP = true positive); b. marginal effects of the age of the AFP; c. marginal effects of the number of true positive AFP cases in the last month; d. marginal effects of the delay in days between AFP onset and AFP stools sampling for several values of the delay between the collection of stools from the AFP and those of the contacts. The marginal effects of the other factors are presented in S6 Fig.