The World Health Organization suggested that the prevalence of Schistosoma mansoni among 7- to 14-year-olds be used to guide treatment strategies in endemic areas. This study explores how well the prevalence in that age group predicted the overall prevalence in the community in data from stool examinations (Kato-Katz method) from 180,000 people in 3 municipalities in Brazil in 1984 and 1985. The median prevalence was higher in 1984, before community treatment was introduced. There was a strong relationship between the prevalence among 7- to 14-year-olds and the overall prevalence in the community. We present sensitivities and positive predictive values for the use of prevalence in the indicator group to select communities for mass treatment as recommended by WHO. For a range of assumptions sensitivity and positive predictive value were never both above 80 %. We suggest that the estimates of validity presented in this paper inform future evaluations of strategies for S. mansoni control.