Morris Rosenberg’s (1965) Self-Esteem Scale (RSES) is perhaps the most widely used instrument for the assessment of trait self-esteem, defined as relatively stable feelings of overall self-worth. The importance of self-esteem in the prediction of other self-attitudes and behavior in conjunction with the ease of administration and scoring of the RSES make the instrument useful for social scientists in many different settings.
Rosenberg originally designed the RSES as a face-valid, unidimensional measure of self-esteem designed to assess adolescent self-worth. Example items from the RSES include "I feel that I am a person of worth, at least on an equal basis with others" and "At times I think I am no good at all." Negatively worded items are reverse-scored to create an overall score. The underlying view of self-esteem on which the scale was founded defines self-esteem as a positive or negative attitude about the self. People with high self-esteem respect and consider themselves worthy. Rosenberg noted that high-self-esteem people do not necessarily feel that they are better than others, but they do not consider themselves to be worse than others. People with low self-esteem express "self-rejection, self-dissatisfaction," and "self-contempt" (Rosenberg 1965, p. 31). According to Rosenberg, low-self-esteem people do not respect themselves and wish that they could evaluate themselves more favorably.
Though Rosenberg fashioned the RSES as a ten-item Guttman scale, researchers most commonly adopt five- or seven-point Likert-style response formats anchored by, for example, 1 = not at all like me to 7 = very much like me. (Blascovich and Tomaka 1991). Internal consistency estimates for the scale typically fall in a range from .77 to .88, indicating acceptable internal reliability. Additionally, test-retest estimates for the RSES range from .85 to .82, revealing that the RSES demonstrates excellent test-retest reliability.
In addition to being reliable, the RSES surfaces as a valid measure of self-worth as well. RSES scores are positively related to an abundance of conceptually related constructs such as confidence and social acceptance, as well as to other self-esteem scales (Blascovich and Tomaka 1991). Moreover, RSES scores are negatively associated with constructs that are typically symptomatic of low self-worth, including anxiety, depression, and feelings of rejection.
Researchers have adapted the RSES for use with different populations (i.e., younger children; Rosenberg and Simmons 1972) and for different purposes. By modifying instructions for the RSES, asking that respondents answer according to how they feel about themselves right now, at the present moment, researchers may assess state self-esteem, an evaluation of self-worth that is more variable than trait self-esteem using the RSES (Kernis, Grannemann, and Barclay 1992).
Although the RSES is a conceptually and psycho-metrically sound measure of unidimensional self-esteem, it does not assess all facets of self-worth. Researchers who are interested in assessing domain-specific elements of self-esteem (e.g., appearance self-esteem, school competence, moral self-worth) should include domain-specific measures in addition to or instead of the RSES, depending on their research questions. Additionally, the RSES is transparent, or face valid, making it easy for the respondent to surmise that they are completing a measure of self-esteem. Such obvious face validity might result in artificially inflated self-esteem scores. Researchers who are more interested in people’s automatic or uncontrolled ratings of self-worth should make use of available implicit measures of self-worth.