A psychometric study of the Rosenberg self-esteem scale : an investigation of gender dif

Patricia Louise Swenson
2003
The present study investigated the psychometric properties of the Rosenberg Self- Esteem Scale (RSES). The RSES is a ten-item self-administered instrument intended to measure global self-esteem. The RSES was administered to a group of 1,443 Canadian university students, including 569 males and 874 females. Professor Alex Michalos, as part of his International Study of Quality of Life, collected the data in 1980. Ages of participants ranged from 17 to 65 years. Mean ages were 21.94 years for
more » ... 21.94 years for males and 22.59 years for females. Coefficient alpha for the sample was 0.86. A two tailed t-test found no significant gender group differences for RSES total test scores. Factor analysis revealed one factor, accounting for 44.8% of variance in test scores. This result is indicative that the RSES is measuring one main variable, and that the RSES is a homogenous unidimensional measure. A gender group differential item functioning analysis was conducted using Zumbo's (1999) Logistic Regression method. DIF analyses found that three test items displayed statistically significant DIF (mainly uniform) at a probability level of 0.005 (which is the Bonferroni corrected Type I error rate). Effect sizes were less than the minimum R-squared of 0.130 required for significance according to the Zumbo-Thomas effect sizes for Likert data. Due to the extremely small effect sizes, it was concluded that the DIF findings were not significant.
doi:10.14288/1.0053822 fatcat:syotgixfufcijjndsx6hhhnopy