| Most of the studies addressing adolescent sexual behavior in America are asymmetrically focused on either one of boys and girls or one race, primarily African Americans. Studies also are small scale and ethnographic by nature. There is lack of cross-sectional studies addressing both adolescent boys and girls, and African Americans and whites making use of a wide range of variables in terms of their sexual behavior. The present study explores sexual behavior making use of a nationally representative cross-sectional (race and gender) dataset addressing both social structural and demographic variables of interest. The data analyzed by this study were collected by Add Health as the second wave of a longitudinal study of health and well being of American adolescents. This study analyzes data collected from a school based stratified sample of African American and white adolescents only (N = 2834).;Eight major determinants of adolescent sexual behavior are explored. These are religiosity, self-esteem, family support, peer group involvement, school attachment, gender, race and age. Five of these variables are developed as additive scale. These scales had Cronbach's coefficient-alpha that ranged from school attachment (.994), religiosity (.993), family support (.796), and self-esteem (.714) to peer group (.309). The other three variables are race, gender, and age which are dichotomized. Dependent variable is sexual activity which is dichotomized as yes and no. Statistically significant relationships are found between school attachment and adolescent sexual behavior, peer group and sexual behavior, age and sexual behavior, family support and sexual behavior, and race and sexual behavior. But self-esteem, religiosity, and gender are found not to have significant relationship with sexual behavior. Results of logistic regression analysis reveals that only school attachment, peer group, race and age are significant predictors of sexual behavior among these adolescents. However, religiosity, self-esteem, family support, and gender are not significant predictors.;Further research is recommended to take into account other contextual variables pertinent to social class, neighborhood factor, urbanicity, adolescents' attitudes towards sex, and their perception of parental attitude towards sex because the dataset did not contain them. |
