Growing Social Inequalities in Youth Civic Engagement? Evidence from the National Election Study

Social class differences in civic engagement persist for both youth and adults. Althoughempiricalevidence ismixed, several recentsocial changespertainingtoyouthsuggest that social inequalities in civic engagement may be growing over time for young people. Using data from theNational Election Study, we compared trends for youth and older adults of varying education levels and tested the hypothesis of an increasing educational disparity in youth political participation. Results for voting supported our expectations: declines over timewere found for less-educated youth only. Unexpectedly, participation in otherpoliticalactivities formore-educatedyouthdeclinedmoreover timecomparedtoother groups.Ourfindingshighlight theneedtocreateequalopportunities foryouthcivicengagement across social groups. American democracy is founded on the premise of citizens’ participation, and likewise, citizenship entails full expression of the rights and responsibilities inherent in societal membership (Walzer 1989). Unfortunately, social class differences in political participation and community engagement in the United States persist (Verba, Burns, andLehman 2003), and these inequalities seriously undermine the representativeness of democracy. From a developmental perspective, historical shifts in the nature of the transition from adolescence to adulthood, along with other social changes affecting youth, indicate that social inequalities in civic engagement among young people have increased in recent years (Finlay,Wray-Lake, and Flanagan 2010). Building on recent empirical findings among youth (Center for Information & Research on Civic Learning and Engagement (CIRCLE) 2011; Syvertsen,Wray-Lake, Flanagan, Osgood, and Briddell 2011), this study used data from the National Election Study collected during the past 50 years to determine whether social class differences in civic engagement are increasing. SOCIAL INEQUALITY IN CIVIC ENGAGEMENT Social inequality is an obstacle for civic engagement because educational and economic resources confer advantages in the civic domain. Associations between socioeconomic disparities and civic engagement have been widely documented among adults as well as youth, at the neighborhood level, and for a range of behaviors including voting, other political activities, and volunteering (Hart and Atkins 2002; Kinder and Sears 1985; McFarland and Thomas 2006;Verba, Schlozman, andBrady 1995). Social inequalities in civic engagement are also evident among youth around the world (Amadeo,Torney-Purta, Lehmann, Husfeldt, andNikolova 2002). The social class divide in civic engagement can be partly attributed todisadvantages thataccumulateduringchildhoodand adolescenceaswellas the lackof institutionalopportunities forcivic participation available to individuals who do not attend college (FlanaganandLevine2010).Beyondparticipatorybehaviors, socioeconomicdisparitiesalsoextendtocivicknowledge, skills, andfeelingsofempowerment (Levinson2010).Verbaandcolleagues (2003) argued that class differences in political participation are largely the result of educationaldisparities,witheducationbeing the “singlemost substantial andmostmulti-faceted influence on political activity” (47). Thus, investigations of socioeconomic inequalities in civic engagement typically operationalize socioeconomic status in terms of education (CIRCLE 2011; Syvertsen et al. 2011). Here we ask an important, unanswered, question: have social inequalities in civic engagement grown in recent years? Evidence from a national study of high school seniors from 1976 to 2005 suggested that social class disparities (measured by adolescents’ college aspirations) in voting intentions and volunteering have expanded since the early 1990s (Syvertsen et al. 2011). Persistent, yet stable, class differenceswere found for political behaviors, trust in government, and public hope. However, other evidence has not LauraWray-Lake is an assistant professor in the School of Behavioral and Organizational Sciences at Claremont Graduate University. She can be reached at [email protected]. Daniel Hart is professor of childhood studies and psychology, and director of the Center for Children and Childhood Studies at Rutgers University. He can be reached at [email protected] Fea tu res ............................................................................................................................................................................................................................................................. 456 PS • July 2012 doi:10.1017/S1049096512000339 found increasing social inequalities in civic trends. For example, data from the Census Current Population Survey showed a large but stable gap in the voting rates of 18 to 29 year olds from 1972 to 2008 when comparing youth based on college experience versus no college experience (CIRCLE 2011). These conflicting results may be due to differences in civic engagement measures (i.e., voting intentions versus reported voting behavior) or in population (high school seniors versus 18–29 year olds). To better understandwhether disparities are growing and forwhom further exploration is needed. UNIQUETRENDS FOR YOUTH? These previous analyses of educational disparities in civic engagement over time have exclusively examined youth (i.e., 18 to 29 year olds). Thus, an implicit assumption is that social inequalities may be growing for young people in particular; indeed, theory and some research suggest that this assumption is plausible for at least three reasons. First, changes to the education system, such as cuts to funding and institution of the No Child Left Behind Act, have forced secondary education to focus on basic subjects and deemphasize civic education (Kahne andMiddaugh 2008). Therefore, youth of varying backgrounds are less likely to gain equal exposure in school to civic education and preparation for civic life than in previous decades. Second, the transition to adulthood has become more protracted in recent decades, with youth taking longer to settle into adult roles and establish residential stability (Arnett 2000; Settersten, Furstenberg, and Rumbaut 2005). Delays in establishing adult roles and community ties parallel documented delays in civic participation such as voting (Flanagan and Levine 2010). This delay suggests that historical trends for youth as compared to older adults exhibit more variability in civic engagement over time. The uncertainty of this developmental transition may compromise the optimal development of more vulnerable youth (Osgood, Foster, Flanagan, and Ruth 2006). Third, fewer institutional opportunities now exist for youth to become involved in politics and community life after high school (Finlay,Wray-Lake, and Flanagan 2010; Flanagan and Levine 2010; Jennings and Stoker, 2004). For example, members of theWorldWar II generation were offered a range of institutionalized civic opportunities from the New Deal, Civilian Conservation Corps, and G.I. Bill; yet, since the 1970s, these types of government support programs have eroded (Finlay et al. 2010). For youthwho are exposed to less civic education in high school and who do not plan to attend college, institutional opportunities for civic engagement may be increasingly important. Our study more rigorously tested the idea that social inequalities in civic engagement are growing among youth by using data fromtheNationalElectionStudy(NES), inwhichparticipantsspan the entire range of adulthood, and compared trends for youthwith trends forolderadults. In linewithpreviousstudies (CIRCLE2011; Syvertsen et al. 2011; Verba et al. 2003), we operationalized social class in terms of education. Based on the measures available, our analyses focused on voting and other political activities. Specifically, given that previous studies examined evidence for the social classdivideexclusivelyamongyoungpeople (CIRCLE2011;Syvertsen et al. 2011), we considered whether a growing educational disparity in voting and political behaviors would bemore evident, or only evident, among 18 to 29 year olds. In otherwords, we hypothesized an education× survey year × age interaction such that rates of political participation for less-educated young people would declinemore steeply over time compared tomore-educated youth and older adults of any education level. METHOD Data came from the National Election Study, a cross-sectional time series study designed to understand Americans’ voting and public opinion. Face-to-face interviews primarily have been used to survey a nationally representative sample of US adults ages 17 to 99 (M 45.76, SD 17.15) biennially since 1952. A multistage probability sampling design was used to sample individuals of voting age from US Census regions, Census blocks, and housing units that are stratified by geography, size, and median per capita income. Our study used postelection interview data from the 1952 to 2008 presidential election years (i.e., 15 data points across 56 years) and from individuals at least 18 years old. Sample size ranged from 1,060 to 2,366 across years, with approximately 26,000 individuals participating overall (56% female). See table 1 for ethnicity of sample across survey years. Data weights were used in all analyses to ensure representativeness of the sample. Measures Voting was measured by asking individuals whether they voted in the most recent presidential election ( yes 1, no 0). Participantswere also askedwhether or not they attended politicalmeetings or rallies, worked for a party or candidate, displayed a candidate button or sticker, and donatedmoney to a party or candidate during the campaign ( yes 1, no 0). Political activities were measured as the sum of these four behaviors. Frequencies of voting and political activities across years are shown in table 1 and means for study variables are displayed in table 2. Respondents reported highest level of education completed on a four-point scale: 8th grade or less (1), high school (2), some college (3), and college or advanced degree (4). Age was calculated from respondents’ reports of their birthdates and measured in years (i.e., 18 to 99). Year of survey was entered as a linear continuous variable. Sex ( female 2,male 1), ethnicity (dummy coded into Black and Other with White as the reference group), and employment (employed 1, unemployed 0)were included inmodels as control variables. RESULTS Two regression models were examined to test the theory-driven hypothesis of an education × year × age interaction. A logistic For youth who are exposed to less civic education in high school and who do not plan to attend college, institutional opportunities for civic engagement may be increasingly important. .............................................................................................................................................................................................................................................................