Pathways to Adolescent Political Participation across Race and Ethnicity

Abstract

This study assesses impacts of racial/ethnic identification on adolescent civic development to inform interventions to counter civic disconnect on the part of ethnic minority youths. Analyses of 4 years of national Monitoring the Future data find that Black and Latino adolescents hold negative political attitudes and low rates of political behavior. Structural equation models find dissimilarities in paths between political attitudes and behaviors for White, Black, Latino, and Asian adolescents. Findings suggest that political attitudes may operate as precursors to political behavior in different ways across races/ethnicities. Implications for civic interventions and future minority youth civic development research are discussed.

KEYWORDS:

• Adolescence
• political participation
• political attitudes
• ethnic identity

Notes

The author acknowledges the Monitoring the Future Project and the National Institute on Drug Abuse for granting access to a limited private-use dataset. I express gratitude to Amanda McBride, John Bricout, Nancy Morrow-Howell, Michael Sherraden, Ed Spitznagel, and William Tate for their helpful comments and feedback. This project was supported by the Center for Social Development at the George Warren Brown School of Social Work, Washington University of St. Louis.

^aThe residual variance of the electoral behavior latent construct is fixed to 0, unlike the other subgroup models.

^bThe covariance term between two variables composing the external political attitudes construct (trust in government and government knows what it is doing) is fixed to 0 to achieve acceptable model fit.

***p < .001

**p < .01

*p < .05.

***p < .001

**p < .01

*p < .05

^{^} p < .08.

***p < .001.

1. Data analysis was limited to 2002–2005, due to the private-use nature of the data. Although major political events impacting youth participation, such as immigration rallies in certain states and the 2008 presidential election, have taken place since these data were collected, this analysis is designed to capture the aggregate nature of attitude-behavior relationships, rather than the individual impact of specific political phenomena.

2. Analyses use sampling weights (M = 1.39, SD = 0.84) to account for variations in sample size across schools and in selection probabilities due to the stratified clustered sampling procedure (Johnston et al., 2005).

3. The R ² value for the electoral behavior construct is high in all structural models presented here. R ² values may over-account for the variance explained by the three other dependent variables in the model (Joreskog, 2000). The high correlation between the political voice and electoral factors (in this model, r = .89) suggest some contribution to the high R ²; however, the R ² of .15 when regressing the latent political voice factor on the latent electoral factor suggests that this relationship is not the primary explanation.