SUMMARY A multiple regression analysis was performed on statistically independent factors derived from blood pressure measurements and possible predictive variables in 618 Utah adults. Nine blood pressure factors obtained in a previous study composed the dependent variables; 35 anthropometric, questionnaire, and biochemical variables were reduced by factor analysis to 10 factors and used as independent variables. Body size and obesity had significant independent effects on different types of blood pressure: body size correlated most highly with systolic blood pressure, while obesity correlated most highly with sitting diastolic blood pressure measurements. Smoking did not correlate with sitting blood pressure but did show a significant positive correlation (after controlling for obesity) with tilt and supine diastolic pressure. Alcohol consumption correlated positively with sitting diastolic pressure when the effects of body size and obesity were controlled. No correlations were found between urinary potassium or sodium excretion and any blood pressure factors, but a significant positive correlation was seen between plasma sodium concentration and several different types of diastolic blood pressure measurements. Psychological stress showed a significant independent positive correlation with systolic blood pressure measurements that was strongest in adults over 35 years of age. The multiple correlation values for the multiple regression equations ranged from 0.19 to 0.52. (Hypertension 8: 243-251, 1986) KEY WORDS • blood pressure determinants • factor analysis • causal factors • multiple regression V IRTUALLY all previous analyses of multiple predictive factors for blood pressure have used systolic and diastolic blood pressure measured in subjects in a sitting position only.'" 5 Because blood pressure can vary considerably depending on posture, activity levels, and external stimuli, a previous study examined 57 blood pressure measurements in six different clinic situations (sitting, standing, lying, tilting, gripping, and just before blood drawing). 6 Factor analysis of these variables showed that all systolic blood pressure measurements composed one factor, while pulse and diastolic blood pressure measurements formed eight separate additional factors. The present study used these nine factors as dependent variables and evaluated the effects of 35 independent variables by using correlation and multiple regression techniques.