Included Studies Assessment of bias.

1Emphasizing binary or continuous assessment of safety.23 information conditions: Informed about common source and household stored water quality at times 1 and 2 (not at time 0), Informed about common source at times 1 and 2 and household stored water quality at time 2 (not at time 0) or informed about common source at time 2 (not at time 1 or 0) (i.e. a waiting list control group only). Villages were randomly assigned to information treatment, with 133, 123 and 130 households in each arm.