Sample sizes were calculated using WIN EPISCOPE 2.0 (Thrusfield et al., 2001) using 95% confidence (two tailed)
and 80% power to detect a difference between 3 and 15% prevalence (Dung et al., 2010) between groups.
This resulted in a sample size of 87 snails per group.
As 20% of snails would not be alive (Dudgeon,1986), at least 110 snails per group should be sampled.
Based on this we have decided to sample 120 M. tuberculata snails per sampling area.
Each pond was “divided” in 9 sampling areas, being the 4 corners, 4 sides between corners and the middle of the pond in
rectangular ponds.
For each of these areas, 1 or 2 types of land use bordering a sampling area were recorded (Table 1).
Road, dike, rice cultured in pond, none (middle of the pond), or agriculture (crops and fruit trees) were considered as low-risk land use for infection of snails.
Presence of livestock sty (pigs and/or poultry), human access point to pond or water connection to canal were considered
as higher risk.
Based on type of land use bordering a sampling area, they were categorized in 5 risk-categories: low-risk, human access
point to pond, livestock sty, both human access point and livestock sty, and water connection to canal.
In case two different types of land use (e.g. livestock sty and agriculture) bordered a sampling area, it was assigned to the risk-category with potentially the highest risk (in the example: livestock sty).
It was assumed that both human access point and livestock sty was at highest risk, and was therefore categorized as a separate risk-category.