Blood pressure is a cardiovascular measurement with dynamic characteristics that can be influenced by a number of internal and external factors. The preferred blood pressure determination method would be one that reduces variability between measurements and that reflects the true blood pressure level. In this article, we present the variability of, and agreement between, the blood pressures collected by two indirect methods on the same patients during a hypertensive research project. Data obtained on patients in a typical clinical setting are also provided. Twenty-four-hour diastolic pressures obtained by the automated method demonstrated no regression to a lower mean, while blood pressures obtained casually in the office exhibited such regression. The 95% confidence interval of repeated measures for casual office blood pressure on a patient in a research setting (35/17 mm Hg) or in typical clinic practice (26/19 mm Hg) were similar, while the range of the mean 24-hour automated blood pressure monitoring (21/11 mm Hg) was smaller and demonstrated less variability. The magnitudes of the differences in blood pressures obtained on separate occasions in the same subjects were significantly lower with automated vs casual blood pressure determination methods (7.9/4.6 vs 13.7/7.4 mm Hg for both systolic and diastolic pressures). The agreement (95% confidence interval) between blood pressures obtained by the two methods (19/12 mm Hg) was found to be similar to the repeatability of automated blood pressure monitoring alone, and superior to that for data recorded casually in the office (35/17 mm Hg). Thus, the variability in mean 24-hour automated blood pressures is less than that for casual office blood pressures. The clinician should understand that the variability of blood pressures measured on an individual may be much greater than that reported for populations of hypertensive patients, and must be considered when applying epidemiologic group data to a specific patient. Moreover, any methodology of indirect blood pressure measurement that may reduce the variability and improve repeatability of casual office blood pressures deserves further consideration.