In a longitudinal seroepidemiological study in the Schistosoma japonicum endemic area of Leyte, the ELISA technique to determine prevalence and incidence rates in elementary school children was compared with similar determinations made by a modified quantitative stool examination (MIFC). In the area of this study, Barrio Salvador, Tanauan, Leyte, the ongoing schistosomiasis japonica control program in the Philippines is dependent on stool examination by MIFC and/or the quantitative thick smear (Kato-Katz) for measurement of prevalence and incidence. Over a 3-year period with multiple periodic examinations, infection rates were measured and the serologic technique was compared to stool examination in 598 untreated children (mostly 7-10 years of age) of Salvador Elementary School. A group of 150 school children from a non-endemic area, Milagros, Masbate, provided sera as a reference negative control. ELISA results are expressed as ELISA activity (EAc) in reference to a positive control serum pooled from parasitologically confirmed cases, dilutions of which were always included in each assay. A convenient positive-negative discrimination level was chosen based on the EAc values obtained from 170 stool-positive Salvador pupils and the 150 pupils of the non-endemic area. Using the chosen discrimination level, ELISA in this study had a sensitivity of 98% and a specificity of 96%. ELISA was significantly more sensitive than stool examination in detecting infections; only 28% of the children were stool positive on a single examination in contrast with 56% positive by ELISA. A single stool examination underestimated serologic positives by 50% while two stool examinations 4 months apart reduced the underestimate to 29%. The underestimation varied by age and sex, and showed no consistent pattern in this regard. Stool-positive children had a wide variation of egg counts with a geometric mean of 6.4 eggs/g of stool, with 52% of the stool positives excreting only 1-5 eggs/g. A high percentage of infected children have a misdiagnosis of infection by stool examination. This has, in the past, resulted in many being misclassified as noninfected. This erroneous classification has serious consequences on the measurement of prevalence and incidence, on studies of clinical manifestations of the disease, and on the evaluation of serologic techniques for diagnosis. Stool examination does not give an accurate measurement of prevalence, and therefore it cannot be relied upon for the evaluation of the current control program. It is recommended that the capability to undertake serodiagnostic tests for schistosomiasis japonica be encouraged and adopted in the Philippines for field