A Psychometric Study of the Prosocial Behavior Scale: Differential Item Functioning by Gender

Some scales co-exist in the literature to measure prosocial behavior in adolescents. Gender differences in prosocial behavior have been a controversial topic of research. To strengthen future research in the area, the psychometric properties of the most used instruments must be guaranteed, especially its gender non-bias. Our study provides psychometric evidence for the Prosocial Behavior Scale in a sample of adolescents, exploring: (a) its factor structure; (b) reliability; (c) gender-related differential item functioning (DIF); (d) nomological validity. A sample of 512 high school students (mean age = 13.62 (SD = 1.34), 51.6% females) participated in the research. Confirmatory Factor Analysis (CFA) was used to test the factor structure of the scale, which adequately fitted the data (χ² (35) = 152.224, p < 0.001, Comparative Fit Index (CFI) = 0.905, Root Mean Squared Error of Approximation (RMSEA) = 0.085 90%CI [0.072-0.099], Standardized Root Mean Squared Residual (SRMR) = 0.079). Reliability results were good (α = 0.74, ω = 0.74). Regarding the DIF, five items presented some gender-related bias, generally benefiting females. However, the DIF impact could be considered negligible. Correlations with the subdimensions of the psychological capital offered evidence of the nomological validity of the scale. In conclusion, the scale presented adequate psychometric properties that support its ability to effectively assess prosocial behavior and gender differences in the prosocial behavior samples of adolescents. Additionally, the results obtained imply that gender differences in the manifestations of prosocial behavior require measurements that can fairly sample behaviors characteristic of each gender.