Modification of the Rosenberg Scale to Assess Self-Esteem in Children

Rosenberg's scale (RSES) is widely used to assess global self-esteem (SE) in adults and adolescents but is not validated for children <12 years. This study assessed the internal consistency, convergent validity, and factor structure of a modified RSES for schoolchildren (CRSES) aged 7-12 years. A total of 711 children aged 9.0 +/- 1.5 years completed the CRSES; a subset (n = 417) also completed a life satisfaction (LS) scale. Data were submitted for confirmatory factor analysis (CFA) and tests of factorial invariance by sex. Two-way ANOVA compared scores by age-group and sex; whilst Pearson's correlations examined the relationship between LS and SE. Following the use of modification indices the fit for the global SE model met the goodness of fit statistic criteria: chi(27, n = 711) = 77.22; chi(normed) = 2.860 CFI = 0.961; RMSEA = 0.051 with 90% CI = 0.038-0.065; SRMR = 0.037; and displayed respectable internal consistency (alpha = 0.79). The model was also factorially invariant by sex. SE scores did not vary sex (p > 0.05); but were significantly reduced in children aged 9-10 and 11-12 years compared to children aged 7-8 years. The global SE score was significantly correlated (r = 0.51; P < 0.001) with LS. The current version of the CRSES can reliably examine global SE in children aged 7-12 years; extending the use of the RSES to allow tracking across the life course.