This study prospectively evaluated psychosocial as well as biologic, behavioral, and demographic factors in the development of duodenal ulcer. Baseline data were collected in 1963 on 8,458 Israeli men over age 40 years who had no history of duodenal ulcer. The subjects were followed for 5 years, and 254 reported the development of radiographically proven duodenal ulcer. The average annual incidence was six per 1,000 subjects. Multivariable logistic regression analysis confirmed several previously reported risk factors: smoking (odds ratio (OR) = 1.64, 95% confidence interval (Cl) 1.25-2.16), greater age (60 years and over) (OR = 1.85, 95% Cl 1.25-2.74), lower salary (OR = 1.50, 95% Cl 1.14-1.96), and lower systolic blood pressure (less than 140 mmHg) (OR = 1.58, 95% Cl 1.16-2.15). In addition, logistic regression analysis revealed a significant association of duodenal ulcer incidence with preceding measures of family stress, emotional support, and coping style. These were: increased family problems (OR = 1.60, 95% Cl 1.19-2.16), low level of perceived love and support from subject's wife (OR = 2.06, 95% Cl 1.05-4.05), and restraining retaliation when hurt by coworkers (OR = 1.89, 95% Cl 1.19-3.00). This study confirms several previously reported risk factors and underlines the importance of stress, lack of social support, and coping style in the development of duodenal ulcer.