The Culture-Free Self-Esteem Inventory (CFSEI-2) was administered to 7 groups of children: 84 White Catholic school students from a New Orleans suburb, 78 White rural public school students from Virginia, 62 Hispanic Migrant student from Florida, 90 Aboriginal and White students from an isolated Canadian community, 199 African American students attending an inner city school, 60 Hispanic and White international students from Venezuela, and 61 Innuit students from isolated community in Labrador. The four elder groups also wrote three words to describe themselves (the Adjective Generation Technique [AGT]). Significant differences in responding between groups were found on all CFSEI-2 scales and for AGT favorability means. Although several possible reasons for these results are discussed, we conclude that the CFSEI-2 is not culture-free. Recommendations are: change the title of the test to avoid misrepresentation, limit test usage to elementary school children, develop an adolescent version with age appropriate language, and construct local norms before using the CFSEI-2 to make decisions about a child's self-esteem. To determine relevance of scores, a team of professionals and lay persons should review items from this or any test given to children who may be different from the normative or standardization group.