Consumption of nitrate and nitrite is associated with a variety of health outcomes. Commercially available test strips that allow semi-quantitative estimation of these contaminants in drinking water are inexpensive relative to laboratory testing, and are simple to use. To examine the accuracy of a nitrate/nitrite test strip, we recruited Washington State residents to estimate levels of these contaminants in their tap water using these strips, and simultaneously provide a tap water sample for laboratory analysis. Paired results were available from 102 homes. On the basis of laboratory assay, nitrate levels as nitrogen ranged from no nitrate (27%) to 40.5 mg/l (median 0.4 mg/l). Spearman's correlation coefficient between test strip- and laboratory-measured nitrate indicated moderate precision overall (r=0.72). Correlation was similar for homes inside and outside city/town limits, but differed by primary source of water for the purveyor indicated by residents (r=0.72 for groundwater and r=0.34 for surface water). Seven (7%) participants reported difficulty in distinguishing colors (contaminant levels) when using the test strip; and among the samples with nitrate, the laboratory assay indicated higher nitrate levels than the test strip for 81%. Nitrite was not detected by laboratory assay; in comparison, five (5%) subjects reported any nitrite according to the test strip. Nitrate/nitrite test strips may be useful in some epidemiologic studies, but should be used with caution, preferably as a screening tool or when laboratory assays are not feasible.