Body mass index (BMI) is widely used in epidemiological studies to categorize individuals to their level of obesity. Yet, several researchers have questioned its validity in terms of being used as a cut-off score, particularly when a single cut-off standard is used for different genders and ethnicities. The purpose of this study was to validate BMI cut-point values associated with obesity in different gender and ethnic groups. Archival data from 596 undergraduate students enrolled in a required fitness and health physical education class were used for this study, which included both males and females, as well as four different ethnic groups (i.e., Asian Americans, African Americans, Caucasians, and Hispanics). Variables utilized for this study included: age, gender, ethnicity, BMI, height, weight, and sum of skinfolds. Validity was examined by means of decision validity, sensitivity, and specificity. The criterion-referenced validity coefficients were high for each ethnic group as well as for males and females. The decision validity coefficients ranged from .85 for Hispanic males to .99 for Asian females. Sensitivity, the percent of true positives, was 57.98% for men and 59.26% for women when BMI was used as a diagnostic index for obesity in the total sample. However, when gender was examined among the ethnic groups, Hispanic males and females had a substantially higher sensitivity (i.e., 80% and 100% respectively). Specificity, the percent of true negatives, was 89.22% for males and 94.65% for females, and was equally high when the ethnic groups were separated. The sensitivity of BMI ranged from 47.83% in Caucasians to 80% in Hispanics. Inversely, specificity was high starting at 85.26% in Hispanics and ranged up to 97.25% in Asians. The negative predictive values were high in all the ethnic groups, which suggested that a high percentage of those who tested negative for obesity using BMI actually were not obese. However, the positive predictive value ranged from moderate (42.86% in African Americans) to low (28.57% in Asians) for each ethnic group indicating that only a small percentage of those who are diagnosed as obese using the skinfold measure were actually found obese using BMI. Therefore, multiple BMI cut-off standards to identify obesity need to be applied to the different gender and ethnic groups.Keyword(s): adult fitness, diversity, measurement/evaluation