Identify Gender Specific Risk Areas for Female and Male Adolescents: An Exploratory Factor Analysis with Youth Violence

Background and Purpose: Youth violence is a growing concern for public health in the United States. Physical violence-related behaviors are important risk factors threat public health and influence children's development. Multiple studies have indicated that gender is one of the risk factors contributing to violent behavior among adolescents. Therefore, it is crucial to explore the underlying dimensions between female and male adolescents in the United States. The purpose of the present Exploratory Factor Analysis, conducted separately for females and males, was applied to examine the underlying dimensions of YRBS 11 risk items.

Method: Analysis was conducted on the 2015 Youth Risk Behavior Survey (YRBS) Public Use File, applying the Exploratory Factor Analysis. The construction sample consisted of 11,914 adolescents in the United States (female: n=5,986; male: n=5,859). In the present EFA, to assess the optimal number of factors to retain multiple methods were employed. The eigenvalues-greater-than-one rule retained the factors with eigenvalues greater than one, and the scree-plot presented the number of factors by a drop in the eigenvalues in the scree-plot items with factor loadings smaller than .40 removed. Then, the Raw Parallel Analysis was applied to determine the number of factors ultimately. Due to the non-normal distribution, An EFA was conducted applying Unweighted Least Square factoring, Direct Oblimin rotation method with Kaiser normalization. Reliability was measure utilizing Cronbach’s alpha value.

Results: The suitability of data for factor analysis was assessed before performing the EFA. Inspection of the correlation matrix revealed the presence of coefficients of 0.4 and above. The Kaiser overall Kaiser-Meyer-Oklin (KMO) values was 0.759, exceeding the recommended value of 0.6 (Kaiser, 1974), and Bartlett's test of sphericity was statistically significant (Bartlett, 1954), then supporting the suitability of factor analysis. The same EFA procedure was repeated separately on female and male gender, resulting in a 3-factor solution for female gender that explained 41.65% of the overall variance, and 4-factor solution for male gender that explained 44.57% of the total variance. Raw Parallel Analysis suggested that 3-factor solution and 4-factor solution provided the best fits.

 Conclusion and Implications: The present study explored the underlying dimensions of the YRBS youth violence items in female and male adolescents. EFA was identified that the underlying dimensions of youth violence on female and male were slightly different. The three factors were defined as follows for both female and male gender: Factor 1: physical fighting; Factor 2: accessibility of weapons; Factor 3: dating violence victimization. Also, the Factor 4 which was identified by EFA for male gender was threatened at school. The competence of youth violence risk assessment to inform about the gender differences in risks could be improved by providing social work practice insights into the underlying structure of empirically validated risk measurement instruments like the YRBS (Hilterman, Bongers, Nicholls, & Van Nieuwenhuizen, 2016). The separate risk factor structures of youth violence for female and male adolescents empirically based information on social work practice in the identification of gender-specific treatment needs and violence prevention programs.