BACKGROUND: In most resource-poor settings, malaria is usually diagnosed based on clinical signs and symptoms and not by detection of parasites in the blood using microscopy or rapid diagnostic tests (RDT). In population-based malaria surveys, accurate diagnosis is important: microscopy provides the gold standard, whilst RDTs allow immediate findings and treatment. The concordance between RDTs and microscopy in low or unstable transmission areas has not been evaluated. OBJECTIVES: This study aimed to estimate the prevalence of malaria parasites in randomly selected malarious areas of Amhara, Oromia, and Southern Nations, Nationalities and Peoples' (SNNP) regions of Ethiopia, using microscopy and RDT, and to investigate the agreement between microscopy and RDT under field conditions. METHODS: A population-based survey was conducted in 224 randomly selected clusters of 25 households each in Amhara, Oromia and SNNP regions, between December 2006 and February 2007. Fingerpick blood samples from all persons living in even-numbered households were tested using two methods: light microscopy of Giemsa-stained blood slides; and RDT (ParaScreen device for Pan/Pf). RESULTS: A total of 13,960 people were eligible for malaria parasite testing of whom 11,504 (82%) were included in the analysis. Overall slide positivity rate was 4.1% (95% confidence interval [CI] 3.4-5.0%) while ParaScreen RDT was positive in 3.3% (95% CI 2.6-4.1%) of those tested. Considering microscopy as the gold standard, ParaScreen RDT exhibited high specificity (98.5%; 95% CI 98.3-98.7) and moderate sensitivity (47.5%; 95% CI 42.8-52.2) with a positive predictive value of 56.8% (95% CI 51.7-61.9) and negative predictive value of 97.6% (95% CI 97.6-98.1%) under field conditions. CONCLUSION: Blood slide microscopy remains the preferred option for population-based prevalence surveys of malaria parasitaemia. The level of agreement between microscopy and RDT warrants further investigation in different transmission settings and in the clinical situation.