It has been suggested that residential exposure to contact currents may be more directly associated with the potential for an increased risk of leukemia in childhood than magnetic fields. Contact current exposure occurs when a child contacts a bathtub's water fixtures, which are usually contiguous with a residence's electrical ground, and when the drainpipe is conductive. The Northern California Childhood Leukemia Study (NCCLS) is the only epidemiological study known to address whether contact current may confound the reported association between residential magnetic fields and childhood leukemia. The study contributed contact voltage and magnetic-field data for over 500 residences of leukemia cases and control children. We combined these data with the results of previous measurement studies of contact voltage in other communities to conduct an analysis of the relationship of magnetic fields with contact voltage for a total sample of 702 residences. The Spearman correlation of magnetic field with contact voltage was 0.29 (Spearman, P < 0.0001). Magnetic-field and contact voltage data were both divided into tertiles, with an upper magnetic-field cutpoint of 0.3 μT suggested by values used in epidemiological results and an upper contact voltage cutpoint of 60 mV based on dosimetric considerations. Expressed as an exposure odds ratios (EOR), we report an association of contact voltage with magnetic fields of 15.1 (95% CI 3.6-61) as well as a statistically significant positive trend across magnetic-field strata (EOR of 4.2 per stratum with 95% CI 2.4-7.4). The associations appear to be large enough to support the possibility that contact current could be responsible for the association of childhood leukemia with magnetic fields.