The validity of several published investigations of the possibility that residential exposures to 50 Hz or 60 Hz electromagnetic fields might cause adverse psychological effects, such as suicide and depression, may have been limited by inadequate controlling for confounders or inadequate measurement of exposures. We investigated the relationships between magnetic field exposure and psychological and mental health variables while controlling for potential confounders and careful characterising individual magnetic field exposures. Five-hundred-and-forty adults living near transmission lines completed neuropsychological tests in major domains of memory and attentional functioning, mental health rating scales and other questionnaires. Magnetic field measurements were taken in each room occupied for at least one hour per day to provide an estimate of total-time-integrated exposure. The data were subjected to joint multivariate multiple regression analysis to test for a linear relation between field exposure and dependent variables, while controlling for effects of possible confounders. Performance on most memory and attention measures was unrelated to exposure, but significant linear dose-response relationships were found between exposure and some psychological and mental health variables. In particular, higher time-integrated exposure was associated with poorer coding-test performance and more adverse psychiatric symptomatology. These associations were found to be independent of participants' beliefs about effects of electromagnetic fields.