Objectives: We sought to determine what differences, if any, existed between white and Mexican American women with polycystic ovary syndrome (PCOS) and whether the same values for fasting insulin, fasting glucose/insulin ratio, and homeostasis model assessment (HOMA) might be applied when screening both ethnic groups for insulin resistance. Study Design: Eighty-three consecutive women suspected to have PCOS but who demonstrated absence of other endocrine disorders comprised the study population. Nineteen healthy ovulatory women volunteered as controls. Fasting serum samples were obtained for determination of thyroid-stimulating hormone (TSH), prolactin, glucose, insulin, free and total testosterone, 17-hydroxyprogesterone, and dehydroepiandrosterone sulfate in the early proliferative phase. An oral glucose load was administered, and blood samples for glucose and insulin were drawn at 1, 2, and 3 hours. Those with impaired glucose tolerance or diabetes mellitus were excluded from our final study population. Four different groups were defined: (1) women with PCOS and insulin resistance, (2) women with PCOS without insulin resistance, (3) women with irregular cycles but without PCOS or another identifiable endocrinopathy, and (4) regular, cycling control subjects. Each group was subdivided by ethnicity (white or Mexican American). A total of 65 white and 37 Mexican American women were studied, including control subjects. Results: Among all study participants, Mexican American women with PCOS had significantly higher mean values for body mass index, fasting insulin, and HOMA but lower mean fasting glucose/insulin levels than white women. When group 1 patients (PCOS with insulin resistance) were compared between ethnic groups, mean fasting insulin and HOMA levels were significantly lower and glucose/insulin ratios higher in white than in Mexican American women. A single cutoff value for insulin resistance in PCOS was insensitive when applied to both ethnic groups. A fasting insulin value >20 μU/mL, HOMA value > 3.8, and glucose/insulin value <7.2 were reasonable screening values in our population of white women, whereas a fasting insulin value >23 μU/mL, HOMA value >4.5, and glucose/insulin ratio <4.0 were feasible screening values in Mexican American women. Conclusions: We conclude that (1) Mexican American women with PCOS are more insulin resistant than white women, (2) the incidence of insulin resistance is higher in Mexican American women with PCOS than in white women, (3) a single “screening” value for PCOS-related insulin resistance screening cannot be applied to both white and Mexican American women, and (4) normative values for insulin resistance screening in the PCOS population should be individualized for different racial or ethnic populations. (Am J Obstet Gynecol 2002;187:1362-9.)