Content area
Full Text
Abstract: The study deals with the structure of the Rosenberg Self-Esteem Scale. In previous research several samples of Czech adolescents were obtained, yielding a similar three-factor structure as exploratory results. Confirmatory factor analysis was conducted on the joint sample (N = 708, 293 boys and 415 girls) and separately on a high school students group (N = 428, 184 boys and 244 girls) and a university students group as well (N = 280, 109 boys and 171 girls). Analyses corroborated the hypothesis about three-factor structure invariant in both groups. These factors are assertion of positive statements about self, denial of negative statements about self and factor of social comparison.
Key words: Rosenberg Self-Esteem Scale, confirmatory factor analysis, adolescents, culture
INTRODUCTION
According to Blascovich and Tomaka (1991), Rosenberg's Self-Esteem Scale (RSES, Rosenberg, 1965) belongs to the most widespread instruments for the measurement of global self-esteem. It was originally intended for the adolescents, but it was also applied to other samples of population. The RSES involves ten items evaluated by the respondents on a four- point (sometimes five-point) scale ranging from "I agree" to "I don't agree at all".
The RSES was originally intended as unidimensional and some current studies confirm this assumption (e.g., Gray-Little, Williams, Hancock, 1997). However, when subjected to factor analysis, two factors usually emerged: the first factor is formed by positively worded items, the second by negatively worded items (Goldsmith, 1986; Hagborg, 1993, 1996; Hensley, Roberts, 1976; Kaplan, Pokorny, 1969; Tafarodi, Swan, 1995).
The interpretation of these factors differs with various authors. Some of these assume that it is about the positive and negative component of self-esteem: e.g. Stolin (1987) assumes that it is about positive self-esteem (or self-confidence) and negative self-esteem (or self-depreciation). Also according to Tafarodi and Swan (1995) the global self-esteem contains two dimensions: a sense of social worth, or self-liking, and a sense of personal efficacy, or self-competence.
Other authors look upon the two-factor structure as the artifact of measurement and consider the scale as unidimensional. They consider the emergence of two ortogonal factors as a result of the response set. Currently, this interpretation prevails on the grounds of results of a confirmatory factor analysis (e.g., Marsh, 1996; Shevlin, Bunting, Lewis, 1995), as well as on the...