An objective method for measuring blood pressure would enhance the usefulness of comparative survey research in hypertension. Previous studies have demonstrated that inexpensive, commercially available machines can achieve satisfactory levels of precision. We evaluated one such device in three field settings. Among 64 untreated hypertensive patients, electronic readings both in the clinic and at home were well correlated with 24 h ambulatory readings (correlation ranged from 0.6 to 0.8), consistently better than measurements obtained by human observers. Evaluation as part of a quality control exercise in an ongoing random population surveys from four countries demonstrated average correlations of 0.92 for systolic and 0.85 for diastolic between human and electronic readings. Subsequent use in survey research involving 4000 individuals in three African communities demonstrated consistency of the electronic devices. These cross-sectional surveys, in which measurements were made by non-professionals, provided qualitative estimates of rural-urban variation in blood pressure, demonstrating the practical utility of the technique. If verified by other evaluative studies, the new generation of semi-automatic electronci blood pressure devices could markedly enhance the comparability of hypertension prevalence studies.