The effectiveness of a family-based cardiovascular disease risk reduction intervention was evaluated in two ethnic groups. Participants were 206 healthy, volunteer low-to-middle-income Mexican-American and non-Hispanic white (Anglo-American) families (623 individuals), each with a fifth or a sixth-grade child. Families were recruited through elementary schools. Half of the families were randomized to a year-long educational intervention designed to decrease the whole family's intake of high salt, high fat foods, and to increase their regular physical activity. Eighty-nine percent of the enrolled families were measured at the 24-month follow-up. Both Mexican- and Anglo-American families in the experimental groups gained significantly more knowledge of the skills required to change dietary and exercise habits than did those in the control groups. Experimental families in both ethnic groups reported improved eating habits on a food frequency index. Anglo families reported lower total fat and sodium intake. There were no significant group differences in reported physical activity or in tested cardiovascular fitness levels. Significant differences for Anglo-American experimental vs. control adult subjects were found for LDL cholesterol. Significant intervention-control differences ranging from 2.2 to 3.4 mmHg systolic and/or diastolic blood pressure were found in all subgroups. Direct observation of diet and physical activity behaviors in a structured environment suggested generalization of behavior changes. There was evidence that behavior change persisted one year beyond the completion of the intervention program. It is concluded that involvement of families utilizing school based resources is feasible and effective. Future studies should focus on the most cost-effective methods of family involvement, and the potential for additive effects when family strategies are combined with other school health education programs.