is among the most prevalent psychopathologies in American Indian communities. Yet little is known about the performance characteristics of common assessment tools in this population. This article describes the factor structure of the Center for Epidemiologic Studies Depression Scale (CES-D) on the basis of data from a sample of 605 American Indian college students. A 3-factor model fit the data well and was judged the most parsimonious. Comparisons of the factor structure across gender demonstrated that the differences were due largely to 1 item: cry. The implications of this work in the context of studies with other populations are discussed.Depression is among the most frequently acknowledged psychopathologies in American Indian communities, second only to suicide and alcoholism (Shore & Manson, 1983). Its occurrence has been linked to a variety of circumstances, including frequent interpersonal conflict (Ross & Davis, 1986), prolonged, unresolved grief (Jilek-All, Jilek, & Flynn, 1978), chronic familial instability (Resnik & Dizmang, 1971), unemployment (Travis, 1984, and rapid culture change (Kraus & Huffier, 1979). Yet little is known about the performance characteristics of even the most common tools for assessing the presence of such a debilitating condition in this special population (Manson, Shore, & Bloom, 1985). Rare exceptions include seminal, but now dated, work with the Cornell Medical Index, Health Opinion Survey, and Minnesota Multiphasic Personality Inventory (see Manson, Walker, & Kivlahan, 1987, for a review). The present report returns to this earlier tradition and describes one in a series of recent attempts to understand the psychometric properties of a widely used measure of depressive symptomatology, the Center for Epidemiologic Studies Depression Scale (CES-D;Radloff, 1977), when used with American Indians.