Hostname: page-component-6898cfcb55-th4rj Total loading time: 0 Render date: 2026-11-11T12:28:15.550Z Has data issue: false hasContentIssue false
  • English
  • Français

Childhood Skill Development and Adult Political Participation

Published online by Cambridge University Press:  19 June 2017

JOHN B. HOLBEIN
JOHN B. HOLBEIN*
Affiliation:
Brigham Young University
*
John B. Holbein is Assistant Professor, Brigham Young University, 745 Kimball Tower, Brigham Young University, Provo, UT 84602 (john.holbein@byu.edu).
Rights & Permissions [Opens in a new window]

Abstract

Recent child development research shows that the psychosocial or noncognitive skills that children develop—including the ability to self-regulate and integrate in social settings—are important for success in school and beyond. Are these skills learned in childhood also important for adult political behaviors like voting? In this article, I use a unique school-based 20-year field experiment to explore whether children who develop psychosocial skills early on are more likely to vote in adulthood than those who do not. Matching subjects to voter files, I show that this intervention had a noticeable long-run impact on political participation. These results highlight the need to better understand how childhood experiences shape civic behaviors later in life. During this critical period, children can be taught the not explicitly political, but still vital, skills that set them on a path toward political participation in adulthood.

Information

Type
Research Article
Information
American Political Science Review , Volume 111 , Issue 3 , August 2017, pp. 572 - 583
Copyright
Copyright © American Political Science Association 2017 

In seeking to understand why some people vote whereas others do not, political scientists tend to focus on adulthood—when citizens are already eligible to vote—rather than on childhood or early adolescence. Political socialization research once focused on childhood in hopes of discovering the roots of participation (e.g., Dawson and Prewitt Reference Dawson and Prewitt1968; Searing et al. Reference Searing, Schwartz and Lind1973), with early researchers arguing that “the more important a political orientation is in the behavior of adults, the earlier it will be found in the learning of the child” (Greenstein Reference Greenstein1965, 12). However, socialization studies struggled to measure political inputs in childhood, find childhood attributes that consistently predicted adult behaviors, and establish causality. Given these difficulties, studies of political behavior in recent years have “eschew[ed] . . .young children” and have instead “focus[ed] on the political learning years [of early adulthood]” (Niemi and Hepburn Reference Niemi and Hepburn1995, 7), justifying this focus by arguing that “the degree of activity or involvement in politics . . . seem[s] to be best explained in terms of [adult] experiences” (Verba and Almond Reference Verba and Almond1989, 267–68). As a result, there remains “little current empirical research shedding light on [the] early childhood antecedents to . . .civic engagement” (Astuto and Ruck Reference Astuto and Ruck2010, 249).Footnote 1

This approach stands in sharp contrast to work from other disciplines, which suggests that childhood is a critical period in determining adult outcomes (e.g., Becker and Tomes Reference Becker and Tomes1986; Chetty et al. Reference Chetty, Friedman, Hilger, Saez, Schanzenbach and Yagan2011; Currie and Thomas Reference Currie and Thomas1995; CPPRG 1999a ). In particular, psychologists, neuroscientists, and economists have noted that a set of so-called psychosocial or noncognitive skills developed in childhood appear to be especially important for success later in life. These skills capture the abilities, attitudes, and motivations that help individuals integrate in society, but that are not measured by achievement tests of cognitive proficiency (Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014). These include the general ability to interact in social settings and to control one’s thoughts, emotions, motivations, and actions. Researchers have shown that individuals who develop psychosocial skills tend to stay in school longer, perform better while in school, have higher earnings, and report higher levels of life satisfaction (Dee and West Reference Dee and West2011; Duckworth et al. Reference Duckworth, Peterson, Matthews and Kelly2007; Heckman Reference Heckman2000; Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014). Still, no work has considered whether psychosocial skills promote a broader set of well-being metrics, including active participation in democracy.

Are psychosocial or noncognitive skills developed in childhood important for adult civic behaviors like voting? Some have speculated that such attributes might matter (e.g., Astuto and Ruck Reference Astuto and Ruck2010). Moreover, there is a smattering of separate strands of empirical work showing that psychosocial attributes like grit or perseverance (Hillygus et al. Reference Hillygus, Holbein and Snell2016), general self-efficacy (e.g., Condon and Holleque Reference Condon and Holleque2013), emotion regulation (Denny and Doyle Reference Denny and Doyle2008), locus of control (Dawes et al. Reference Dawes, Cesarini, Fowler, Johannesson, Magnusson and Oskarsson2014), patience or delayed gratification (Fowler and Kam Reference Fowler and Kam2006; Hill Reference Hill2016; Schafer Reference Schafer2016), overall mental well-being (Ojeda Reference Ojeda2015), impulse control (albeit indirectly measured, e.g., Pacheco and Plutzer Reference Pacheco and Plutzer2007), and sociability (e.g., Fowler and Kam Reference Fowler and Kam2007; Uhlaner Reference Uhlaner1989)—to name a few—appear to predict voter participation. However, these studies have examined psychosocial attributes in isolation and have mostly relied on observational, cross-sectional data. As such, critical questions remain unanswered about causality and the potential for policies to promote such attributes in a way that also increases civic engagement. In other words, can early life psychosocial interventions have a long-run effect on civic engagement later in life? Or are psychosocial attributes innate and immovable attributes?

To explore these questions, I consider the broader effects of a program targeting, and moving, childhood psychosocial attributes. By using a unique dataset with random assignment, I can, for the first time, establish a causal connection between early life investments in psychosocial abilities and voter participation in adulthood. To do so, I leverage a unique 20-year, multisite field experiment—the Fast Track intervention. Previous research shows that this childhood intervention targeted—and successfully moved—a bundle of students’ psychosocial attributes like self-control, general self-efficacy, and empathy while leaving their cognitive abilities and other attributes relevant to political participation virtually unchanged (CPPRG 1999a , 2002, 2004, 2007, 2010).Footnote 2 Matching Fast Track participants to state voter files, I show that despite a low propensity for participation at baseline, children provided with the Fast Track program vote at a noticeably higher rate in adulthood than those in the control group. Assignment to the Fast Track program increased turnout by 7 to 9 percentage points (intent-to-treat [ITT]), with full exposure increasing participants’ turnout by 11 to 14 percentage points (treatment-on-the-treated [TOT])—constituting a 30 to 40% increase in baseline participation rates. Using mediation models, I provide some descriptive evidence that this effect was not the consequence of well-known correlates of political participation—such as cognitive ability, income, or social status—but instead appears to be directly the result of improvements in childhood psychosocial skills, which helped set children on a path toward participation in adulthood.

This article makes several important contributions to the study of civic participation. In exploring psychosocial attributes together, this work overcomes some of the challenges that have plagued the previously scattered research in this area.Footnote 3 By couching the study in the context of a randomized-control intervention, it provides the first causal evidence that psychosocial attributes targeted, and developed, early on can have a substantial impact on adult political behavior. By incorporating psychosocial attributes formally into studies of political behavior, this work connects to a broader framework and burgeoning literature from other disciplines (e.g., Dee and West Reference Dee and West2011; Duckworth et al. Reference Duckworth, Peterson, Matthews and Kelly2007; Farrington et al. Reference Farrington, Roderick, Allensworth, Nagaoka, Keyes, Johnson and Beechum2012; Heckman Reference Heckman2000; Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014; Heckman and Rubinstein Reference Heckman and Rubinstein2001; Heckman et al. Reference Heckman, Stixrud and Urzua2006), therefore expanding collective understanding of the types of life outcomes influenced by psychosocial attributes. Doing so in a causal manner provides clear support for new avenues of research on the family of nonpolitical attributes, abilities, and motivations that drive people to participate in politics.

Second, the evidence presented here provides important theoretical insights into the points of the life course that matter in determining who votes. Although early life political socialization studies have dwindled in recent years, the results presented here provide clear support for a refocusing of attention back on early childhood. This article takes a meaningful step in that direction—providing concrete evidence that what happens in childhood can matter a great deal for political participation in adulthood. Although children may not be developing political attitudes and values when they are very young (Niemi and Hepburn Reference Niemi and Hepburn1995), they are developing the not explicitly political, but still vital, psychosocial skills that persist as resources that they can call upon in adulthood. These skills go beyond the political values, political knowledge, or cognitive abilities that children are thought to learn when they are in school (Campbell Reference Campbell2006; Nie et al. Reference Nie, Junn and Stehlik-Barry1996; Verba et al. Reference Verba, Schlozman and Brady1995). Instead, they involve the ability to regulate thoughts, motivations, emotions, and behaviors, as well as the ability to work with others. These skills play a vital, yet until now unexplored, role in encouraging political participation.

Finally, this article’s contributions cut across disciplinary boundaries and speak to current policy debates. Whereas most of the standard predictors of civic participation (e.g., age, race, income, and—according to some—psychological characteristics) are resistant to change, this work shows very clearly that psychosocial attributes important for voting are malleable. As such, this article offers insights into the design of new interventions to help remediate dismally low and unequal rates of civic participation. These findings suggest that to encourage active participation in democracy, schools should do more than just nurture students’ cognitive abilities—the predominant focus of schools captivated by students’ performance on standardized tests of academic achievement—and teach them knowledge and facts about government and politics. Schools should also promote the general psychosocial skills that play a vital role in encouraging active civic participation.

BACKGROUND AND CONCEPTUAL FRAMEWORK

Recent research from developmental psychology, neuroscience, economics, and other related fields shows that children who develop psychosocial skills are more likely to achieve various markers of life success in school and beyond (for recent overviews of this literature, see Farrington et al. [Reference Farrington, Roderick, Allensworth, Nagaoka, Keyes, Johnson and Beechum2012] and Heckman and Kautz [Reference Heckman, Kautz, Heckman, Humphries and Kautz2014]). Psychosocial skills are often called noncognitive skills, thus striking a contrast with the cognitive skills of logic, reasoning, learning, memory, and problem solving that have long been used in models of political behavior (e.g., Denny and Doyle Reference Denny and Doyle2008; Nie et al. Reference Nie, Junn and Stehlik-Barry1996; Verba et al. Reference Verba, Schlozman and Brady1995).Footnote 4 In general, scholars sort psychosocial skills into two subcomponents: those involving self-regulation and those involving social skills (Sorensen and Dodge Reference Sorensen and Dodge2016). Self-regulation skills help individuals “develop, implement, and . . .maintain planned behavior in order to achieve one’s goals” (Kanfer Reference Kanfer1970). Components of self-regulation include, but are not limited to, general self-efficacy, belief that one can do what he or she sets out to do; delayed gratification, the ability to resist temptation and patiently wait for later rewards (Fowler and Kam Reference Fowler and Kam2007); grit, the ability to persevere despite obstacles (Duckworth et al. Reference Duckworth, Peterson, Matthews and Kelly2007); emotion recognition and regulation, the ability to understand and control individual affect (Wellman et al. Reference Wellman, Cross and Watson2001); and behavioral control, the ability to avoid negative behavior and exhibit positive behavior (Astuto and Ruck Reference Astuto and Ruck2010). Social skills involve the ability to work with others productively—to communicate, build friendships, and solve group-based problems (CPPRG 1999a ).

Although there may be multiple subcomponents of psychosocial ability, scholars have shown conceptually and empirically that these attributes belong to a shared family of latent ability separate from cognitive aptitude (e.g., Heckman Reference Heckman2000; Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014; Heckman and Rubinstein Reference Heckman and Rubinstein2001; Heckman et al. Reference Heckman, Stixrud and Urzua2006; Jackson 2012; Park et al. Reference Park, Tsukayama, Goodwin, Patrick and Duckworth2017). Conceptually, whereas cognitive ability involves logic, reasoning, and memory, noncognitive or psychosocial ability involves self-regulation, executive functioning, top-down processing, motivation, and sociability.Footnote 5 Empirical work supports this conceptual distinction. For example, measures of cognitive ability and measures of psychosocial ability load onto two distinct factors (Heckman et al. Reference Heckman, Stixrud and Urzua2006; Jackson Reference Jackson2012; Sorensen and Dodge Reference Sorensen and Dodge2016; Park et al. Reference Park, Tsukayama, Goodwin, Patrick and Duckworth2017; see also Section 1 of the Online Appendix). Further, fMRI studies show that psychosocial and cognitive abilities require different areas and networks within the human brain (e.g., Yang and Raine Reference Yang and Raine2009). Measures of individuals’ psychosocial ability also together predict several adult outcomes (education, income, health, etc.) above and beyond measures of cognitive ability (Farrington et al. Reference Farrington, Roderick, Allensworth, Nagaoka, Keyes, Johnson and Beechum2012). Finally, psychosocial abilities move together in shared interventions, as I show in the following analysis. (For further discussion and evidence on psychosocial attributes’ nature as a separate bundle of attributes, see Section 1 of the Online Appendix.) Thus, although a handful of previous empirical studies have shown a relationship between individual psychosocial attributes and political engagement, the current study is the first to evaluate the causal implications of this distinct bundle of individual abilities.

Why Psychosocial Attributes and Participation?

There are theoretically compelling reasons to suspect that psychosocial abilities could play a role in encouraging political participation. Scholars have long noted that participation comes with costs and barriers (e.g., Verba et al. Reference Verba, Schlozman and Brady1995). These costs include registering (Corvalan and Cox Forthcoming; Wolfinger and Rosenstone Reference Wolfinger and Rosenstone1980), locating and traveling to polling locations (Brady and McNulty Reference Brady and McNulty2011), waiting in line at the ballot box (Pettigrew Reference Pettigrew2016), finding a time to vote that works with one’s schedule (Garmann Reference Garmann2016), navigating potential inclement weather on Election Day (Gomez et al. Reference Gomez, Hansford and Krause2007), learning about the candidates and issues, and navigating a complex web of institutional rules (Leighley and Nagler Reference Leighley and Nagler2013). Political science research suggests that the resources people have—such as their levels of knowledge about politics, income, or how many years they spend in school—help them overcome these costs. In a complicated political environment, participation is thought to be a “skill-based act”; simply put, individuals’ skills act as resources that help make political participation “less daunting and costly” (Verba et al. Reference Verba, Schlozman and Brady1995, 304).

Chief among the skills thought to matter for voting are the verbal cognitive skills captured by standardized tests of individual ability (Denny and Doyle Reference Denny and Doyle2008; Luskin Reference Luskin1990; Nie et al. Reference Nie, Junn and Stehlik-Barry1996; Verba et al. Reference Verba, Schlozman and Brady1995; Wolfinger and Rosenstone Reference Wolfinger and Rosenstone1980). However, cognitive ability may not capture all of the individual abilities that matter for political participation. Simply put, a broader set of psychosocial or noncognitive skills may also act as resources that help citizens overcome voting barriers. Consistent with this view, some have argued—but not tested—that the psychosocial skills children learn in schools may “foreshadow the skills . . . of a civically engaged [citizen]” (Astuto and Ruck Reference Astuto and Ruck2010, 258). Perhaps independent of individuals’ cognitive ability, their ability to drum up the motivation, energy, and focus required to overcome the many obstacles and distractions that get in the way of voting may be important. Individuals who develop these skills early in life may be more oriented toward participating, able to follow through on their desire to participate, and less likely to experience negative life events that make them ineligible or unable to vote.

First, psychosocial skills may increase citizens’ general motivation, which may, in turn, increase their capacity to participate in politics. For example, the psychosocial ability to empathize with others—to be able to “put oneself into another’s shoes,” or to recognize that they have thoughts, emotions, beliefs, and desires that are independent from one’s own—may promote altruistic behavior, a known predictor of civic participation (Fowler and Kam Reference Fowler and Kam2007).Footnote 6 Thus, empathy may help citizens recognize when social problems are present and increase their desire to mobilize on behalf of others. Indeed, for these reasons, political theorists have speculated that those who develop an “empathic understanding” may be more likely to develop “civic virtue,” or an inclination toward participating (Dahl Reference Dahl1992, 53, 56–57). Similarly, a sense of general self-efficacy—the belief that actions across various life domains will yield desired results—may help individuals believe that they have the specific capacity to participate in politics (Condon and Holleque Reference Condon and Holleque2013). Together, the family of psychosocial abilities may give individuals the domain-specific motivation needed to participate in politics.

Second, psychosocial skills may empower individuals to follow through once they have a desire to participate (Hill Reference Hill2016). Given that voting is costly, citizens’ ability to self-regulate—or to voluntarily control behaviors, emotions, motivations, and thoughts in the presence of temptations or diversions—may be particularly important in helping them avoid the distractions and obstacles that get in the way of voting. For example, grit or perseverance may help individuals overcome the specific obstacles that impede voting, such as long lines on Election Day (Hillygus et al. Reference Hillygus, Holbein and Snell2016). Indeed, some have speculated that “perseverance could conceivably account for why some . . .overcome the start-up costs of registration [to vote]” (Plutzer and Wiefek Reference Plutzer and Wiefek2006, 674). Similarly, the ability to delay gratification may help individuals overcome the immediate costs and delayed benefits of voting (Fowler and Kam Reference Fowler and Kam2006; Hill Reference Hill2016; Schafer Reference Schafer2016). Other psychosocial attributes such as emotion regulation, locus of control, and impulse control may work in a similar way (Denny and Doyle Reference Denny and Doyle2008; Dawes et al. Reference Dawes, Cesarini, Fowler, Johannesson, Magnusson and Oskarsson2014; Pacheco and Plutzer Reference Pacheco and Plutzer2007). Together, psychosocial abilities may work to reinforce voters’ ability to follow through on their vote intentions.

Finally, psychosocial skills may influence participation by making citizens less likely to experience negative life events that make them ineligible or unable to vote. Previous work shows that individuals’ psychosocial skills may reduce the likelihood of negative health (e.g., teenage pregnancy), education (e.g., dropping out of school), and criminal behaviors (Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014; Sorensen and Dodge Reference Sorensen and Dodge2016). Moreover, previous work in political science shows that individuals with such negative life experiences are, on average, much less likely to vote (e.g., Gerber et al. Reference Gerber, Huber, Meredith, Biggers and Hendry2015; Pacheco and Plutzer Reference Pacheco and Plutzer2007). Putting these parallel strands together, psychosocial skills may together help children avoid demobilizing life events that make them ineligible, unwilling, or unable to participate in adulthood.

Undoubtedly, the mechanisms are many and complex, but based on these theoretical reasons and the strong relationship found for nonpolitical outcomes, psychosocial skills are ripe for examination as to their role in influencing political participation. As already mentioned, a few studies have found an observational relationship between psychosocial skills and political participation (e.g., Condon and Holleque Reference Condon and Holleque2013; Dawes et al. Reference Dawes, Cesarini, Fowler, Johannesson, Magnusson and Oskarsson2014; Denny and Doyle Reference Denny and Doyle2008; Fowler and Kam Reference Fowler and Kam2006, Reference Fowler and Kam2007; Hill Reference Hill2016; Hillygus et al. Reference Hillygus, Holbein and Snell2016; Ojeda Reference Ojeda2015; Pacheco and Plutzer Reference Pacheco and Plutzer2007; Schafer Reference Schafer2016; Uhlaner Reference Uhlaner1989). However, it remains unclear whether psychosocial attributes actually causally affect participation or instead reflect other observable factors—such as income or cognitive ability—or unobservable factors that also influence political participation. Given this, understanding the role which psychosocial skills developed in childhood play for adult participation requires a causal identification strategy.

EMPIRICAL CASE: FAST TRACK INTERVENTION

The Fast Track intervention provides an opportunity to explore the causal effect of psychosocial skills. This intervention was one of the earliest, and largest, randomized-control trials to target children’s psychosocial skills as a means of attempting to improve their life outcomes. Started in 1992, this multisite program began with a sample of children who were just finishing kindergarten. The intervention was administered in four communities—Durham, North Carolina, Nashville, Tennessee, rural central Pennsylvania, and Seattle, Washington—and three cohorts. Within these sites, the principal investigators targeted children “for whom the prospect of [life] success . . .[was] most daunting” (CPPRG 2011, 332). Following a clustered design, the Conduct Problems Prevention Research Group (CPPRG) selected 55 high-risk schools to be randomized to either treatment or control groups. This resulted in 445 children in the treatment group and 446 in the control group. More details about subject recruitment can be found in Section 2.1 of the Online Appendix.

The Fast Track intervention drew heavily from the broad psychology literature on child development. It was explicitly “devised to increase [children’s] emotion regulation and social-cognitive skills” (CPPRG 1999a , 633). It was hoped that targeting students’ psychosocial skills would encourage them to “becom[e] a nondisruptive and nonaggressive member of a classroom” (CPPRG 1999a , 633). To achieve its goals, the principal investigators designed a multicomponent, multiyear treatment program that was primarily implemented while children were in elementary school. This six-part intervention included a formal curricula, home visits, parent training groups, tutoring, friendship groups, and peer pairing. One of the key components was an instructor-led curriculum called PATHS. This curriculum taught children, through hands-on applications, “skills for emotional understanding and communication (i.e., recognizing and labeling emotions), friendship skills (i.e., participation, cooperation, fair play, and negotiation), self-control skills (i.e., behavioral inhibition and arousal modulation), and social problem-solving skills (i.e., problem identification, response generation, response evaluation, and anticipatory planning)” (CPPRG 1999a , 637). The other five components of the program were designed to reinforce these skills by promoting positive behavior management skills, enhancing students’ social skills, and fostering problem-solving and life management skills. The intervention featured discussions, stories, films, games, crafts, joint reading activities, and role-plays (CPPRG 1999a ). Sessions built on each other and focused on reviewing and practicing the skills taught in previous weeks. Group meetings were held often in the initial stages but later declined in their frequency. In short, although the program had several components, all were unified in their targeting of improvements in psychosocial skills. Given that Fast Track was not designed with the explicit goal of making active citizens, this intervention contained no expressly political content. More detail about the specifics of the Fast Track program can be found in Section 2 of the Online Appendix.

From follow-up data collection efforts, we know that Fast Track helped children develop psychosocial skills (CPPRG 1999a , 1999b , 2002, 2004, 2007, 2010; Sorensen and Dodge Reference Sorensen and Dodge2016). Relative to the control group, those in the treatment arm manifested “significantly lower rates of aggressive behaviors at school” (CPPRG 2011, 333). Intervention children could more readily identify emotions, scoring 0.28 standard deviations above the control group. Four years into the program, the treatment group scored 0.20 standard deviations above the control group in ability to work through difficult group problems and scored 0.29 standard deviations higher on self-control. Eight years into the program, the treatment group manifested higher levels of self-efficacy, being 0.16 standard deviations above the control group. In adolescence, treated children manifested reduced conduct problems in the home, school, and community, with decreases in involvement with deviant peers, hyperactivity, delinquent behavior, and conduct disorders (CPPRG 2004; Sorensen and Dodge Reference Sorensen and Dodge2016). Treated subjects also showed signs of improved social competence and social cognition (CPPRG 2010). In adulthood, those in the treatment group had 30 to 40% lower conviction rates for violent and drug-related crimes than the control group, suggesting that the program had set children on a long-run path of controlling their emotions, motivations, thoughts, and behavior and being able to interact productively in social settings (Sorensen and Dodge Reference Sorensen and Dodge2016).

In contrast, Fast Track had only a small and limited effect on cognitive ability. From the time participants were in the second grade, the treatment and the control group were statistically and substantively indistinguishable on measures of verbal and math proficiency (see Table A5 in the Online Appendix). Although not an ideal result from a policy perspective, this posttreatment balance has the virtue from a design perspective of allowing us to extend our understanding of the skills that matter for political participation beyond the cognitive skills that so many have previously studied (e.g., Denny and Doyle Reference Denny and Doyle2008; Luskin Reference Luskin1990; Nie et al. Reference Nie, Junn and Stehlik-Barry1996; Verba et al. Reference Verba, Schlozman and Brady1995; Wolfinger and Rosenstone Reference Wolfinger and Rosenstone1980).

In short, Fast Track’s design and long-run effects make it ideal for evaluating the impact of psychosocial skills, and programs targeting these, on voter turnout. It should be noted, however, that despite this valuable virtue, Fast Track’s multicomponent design makes it admittedly difficult to distinguish which specific aspects of the program drive any effect on voter turnout (a point I return to later).

MATCH TO VOTER FILES

To evaluate the impact of Fast Track’s psychosocial skill treatment on political participation, in July 2014 (when the participants in the three cohorts were 26 to 29 years old) I matched Fast Track participants to state voter files. The process of matching Fast Track participants to voter files consisted of uniquely matching individuals using their first names, last names, and birthdays.Footnote 7 Fortunately, all subjects had these matching characteristics collected at baseline. The search process matched individuals based on their most recent address from the year 19 follow-up (2011) that was maintained through when the match was conducted. Using these current addresses, I searched for all individuals in their current and original intervention state.Footnote 8 More details about the match can be found in Section 3 of the Online Appendix.

This match was possible because the Fast Track organization implemented a robust system of regular follow-up surveys of both the treatment and control groups. These started in the first year of the intervention (1991–1992) and continued almost yearly until the most recent follow-up in year 19 of the study (2010–2011, when subjects in the three cohorts were 22–25 years old). This rigorous follow-up procedure has assured that, even more than two decades later, very few students have dropped out of either the treatment or control groups—in year 19 of the study (2011), 93% remained in the sample with a current address in the Fast Track data files. Importantly, this metric of attrition is balanced across the treatment and control conditions, with 91.7% of individuals in the treatment group and 93.3% in the control group having a current address (p = .38). Moreover, the types of people who have attrited appear to be similar—based on their baseline characteristics—across the treatment and control groups. Out of about 34 pretreatment characteristics that are available in the Fast Track files, only one of them (child’s emotional appropriateness score) predicts attrition when it is interacted with the treatment variable—about what we would expect simply by random chance. Simply put, there is little evidence of different types of people dropping out of the Fast Track pool across the treatment and control groups. This is important because it helps ensure similar match quality across the treatment and control groups.

One might be worried that other posttreatment behaviors—such as moving, getting married (and changing one’s name), or dying—could make matching subjects to voter files difficult and potentially bias the results. However, these behaviors remain balanced across the treatment conditions. The comprehensive and high-quality nature of the Fast Track data is of critical importance. Given balanced attrition, movement rates, and quality of matching inputs, the match itself is unlikely to bias the results outlined in the following. If individuals were missed or mistakenly matched in the procedure used, this should only introduce additional uncertainty into the model estimates. Even still, to guard against the small possibility of matching biasing the results, I also use a self-reported measure of voting as a robustness check in the Online Appendix (see Figure A6).

Following previous practice (e.g., Ansolabehere and Hersh Reference Ansolabehere and Hersh2012; Sondheimer and Green Reference Sondheimer and Green2010), those who could not be located were marked as having never registered nor voted. Overall, 44% of individuals in the Fast Track intervention were matched to voter files. This registration rate is comparable to those of this age range (27–29 years at the time of the match) and socioeconomic status (SES) in the general population.Footnote 9 Moreover, registration within the sample follows well-known patterns, with subjects with higher SES and educational attainment being more likely to register.

Based on the data available from the voter file match, it is apparent that without the Fast Track program the subjects were unlikely to participate in politics. Being largely low-income, at-risk children of low-participation parents, these children had the standard characteristics of low-propensity voters. As a reference, only about 26% of subjects in the control group voted in at least one election from 2004 to 2012. In short, these children had the deck stacked against their participation in politics in adulthood.

METHODS

Given program randomization, showing the ITT effect of the Fast Track program is relatively straightforward. In the following, I provide both raw estimates of differences between the treatment and the control groups and those that benefit from the added precision of baseline covariates. Estimating the effect of exposure to the Fast Track psychosocial skill intervention (i.e., the TOT) requires the standard instrumental variables approach provided by Angrist et al (Reference Angrist, Imbens and Rubin1996).Footnote 10 This is required, as over time there was some noncompliance by program participants. Despite the fact that there was little to no noncompliance in the control group (due to the clustered initial design and individual tracking thereafter), some individuals in the treatment arm did stop attending program sessions. Overall, those in the treatment group attended about 66% of documented program sessions over the 6 year period.Footnote 11

Under this standard approach, in the first stage, Fast Track program receipt (T i ) is instrumented by random assignment to Fast Track (A s ). The coefficient for this estimate—γ1—shows the rate of compliance. In the second simultaneously estimated stage, the effect of program receipt ( $\hat{T_{i}}$ ) on voter turnout (V i ) is estimated. This approach is displayed in Equation (1).

(1) $$\begin{equation} T_{i}=\gamma _{0}+\gamma _{1}A_{s} + \gamma _{2}\: X_{i} + \epsilon _{is} \end{equation}$$
$$\begin{equation*} V_{i}=\beta _{0}+\beta _{1}\: \hat{T_{i}} + \beta _{2}\: X_{i} + \sigma _{is} \nonumber \end{equation*}$$

The key outcome of interest is whether an individual voted in adulthood (V i ). I specify this outcome in two primary ways. Following the approach used by Sondheimer and Green (Reference Sondheimer and Green2010), the first uses an indicator for whether an individual ever voted in the elections for which they were eligible. The second uses an adjusted count of the number of times individuals voted, with the numerator being the number of times an individual voted and the denominator being the number of times they were eligible to vote. In the Online Appendix, I also present results with a self-reported turnout item that produces similar results.

The models include a set of pretreatment controls (X i ), which consist of race, gender, implementation site, cohort, age at the start of the intervention, SES, and information about whether the subject had siblings and married parents at baseline. (Descriptive statistics of these baseline covariates can be found in Table A1 in the Online Appendix.) These have been included in previous examinations of Fast Track’s impacts and are simply included to increase precision (CPPRG 1999a , 2002, 2004, 2007, 2010).Footnote 12

The coefficient β1 is the causal effect of exposure to the full Fast Track childhood psychosocial program on adult voter turnout (i.e., the TOT).Footnote 13 For ease in interpretation, I use a linear probability model with the ever-voted outcome. However, I also display average marginal effects from a probit model in the results (the results do not change when I use this approach). When I use the proportion of elections voting as the outcome, I use a tobit model because this outcome is pseudocontinuous (it takes on nine values) and is truncated at 0 and 1—but, again, the results are statistically indistinguishable when I employ ordered logit or two-part hurdle models.

RESULTS

Raw turnout differences between the treatment and control groups indicate that individuals assigned to the Fast Track program in childhood turned out at a rate approximately 6.6 percentage points higher than the control group (treatment = 26.2, control = 32.8; p ≈ .04). Model estimates provided in Table 1 support this basic comparison. These indicate that assignment to the Fast Track treatment (ITT) increased turnout by 7.3 percentage points above the control group (95% confidence interval [CI]: [0.02, 0.13]). The TOT indicates that individuals exposed to Fast Track turned out to vote in at least one of the federal elections held during 2004–2012 at a rate 11.1 percentage points higher than the control group (95% CI: [0.03, 0.20]).

TABLE 1.

Fast Track’s Impact on Voter Turnout

DV = Dependent Variable; Prop.= Proportion; ITT, intent to treat; TOT, treatment on the treated; *, p ≤ .05; +, p ≤ .10. Notes: The 95% confidence intervals are displayed below the coefficients. These are based on cluster robust standard errors that adjust for the treatment being at the school level and the potential for heteroskedasticity (results unchanged with multilevel models). Control coefficients are reported in the Online Appendix. Controls include race, gender, age, socioeconomic status, whether the child has a sibling, whether the child has a male present in the home, whether the child’s parents are married, and site and cohort fixed effects. For those in Washington, vote history was only available after 2006. F is the test statistic for instrument strength. (The analogous χ2 statistic for the joint hypothesis of the instruments being equal to 0 is shown for tobit models.) Columns 3 and 4 report a tobit specification for the proportion of elections voting (nine categories, truncated at 0 and 1). Probit specifications report average marginal effects.

When I use the proportion of elections voting as the outcome, the results are larger still and slightly less precise, yet still statistically significant. Individuals assigned to the Fast Track program vote at a predicted rate of 9.4 percentage points higher than those in the control group (95% CI: [0.0004, 0.19]; p = .049), with those fully exposed voting at a predicted rate 14 percentage points higher (95% CI: [0.00, 0.28]; p = .052). If I use an ordered logit or a two-part hurdle model, the results are similar, showing that the program was particularly effective at moving individuals from never voting to voting in at least one election. These and other robustness checks can be found in Section 8 of the Online Appendix.

Figure 1 provides two visualizations of these estimates, with the first panel (a) showing the ITT and the second (b) the TOT. As can be seen, the estimates of Fast Track’s impact are statistically distinct from zero at the 95% significance level. Moreover, the point estimates are substantively meaningful. When benchmarked with turnout levels in the control group, the estimates represent about a 28% (ITT) to 42% (TOT) increase in baseline participation rates. Three things should be noted about the size of this effect. First, although the point estimate is sizable and is statistically distinct from zero, the 95% confidence intervals do not allow us to totally rule out smaller effects. Second, it should be kept in mind that this subject pool was particularly disadvantaged—having a low propensity to vote at baseline. The reader should keep in mind that it is possible that there could be diminishing returns to programs like Fast Track—that in more advantaged populations the turnout effects could be smaller.Footnote 14 Even if this were the case, however, the effects would be important, helping to perhaps narrow stubborn resource-based gaps in voter turnout. Finally, Fast Track’s effect on turnout is similar—albeit slightly larger—to other childhood school programs that target disadvantaged populations. Indeed, the confidence intervals for the raw mean difference in turnout for Fast Track (6.6 percentage points) overlap with the similar mean differences from the Perry Preschool program (5.6 percentage points) provided by Sondheimer and Green (Reference Sondheimer and Green2010, see their Table 2).Footnote 15

FIGURE 1.

Fast Track’s Impact on Voter Turnout

Panel (a) in Figure 1 shows predicted probabilities (distributions) and mean levels (dashed lines) of ever voting based a model with treatment status and baseline characteristics (i.e., the ITT). Panel (b) shows mean differences in predicted turnout for those fully exposed to the Fast Track program (i.e., the TOT). N = 812.

All of that said, the Fast Track effect remains important. Taken together, these results solidify the point that the Fast Track program that targeted psychosocial skills in childhood was effective at increasing voter turnout in adulthood. This result provides strong evidence that early life interventions targeting psychosocial skills can substantially move participation downstream. Such a result finds meaning given that previous research has focused on what happens to citizens in adolescence and adulthood. Such a finding is also meaningful because of the population in which it was observed (children of low-socioeconomic parents who were identified as being high risk). Given the normal exclusion of voters with the attributes of the Fast Track participants from the electorate (e.g., Wolfinger and Rosenstone Reference Wolfinger and Rosenstone1980; Leighley and Nagler Reference Leighley and Nagler2013), this result has important implications for how we think about ameliorating low (i.e., bringing in more voters who under the counterfactual would not have voted) and unequal (i.e., drawing these gains from low-propensity citizens) rates of voter turnout.

POTENTIAL MECHANISMS

Although the design employed here is not perfectly suited to explore individual mechanisms, I am able to offer some insights about the channels through which Fast Track potentially mobilized its participants—be it through psychosocial skills or some alternate path. To do so, I estimate mediation models that allow for interdependence among multiple mediators, following the approach suggested by Kohler et al. (Reference Kohler, Karlson and Holm2011).Footnote 16 These models allow me to explore what proportion of the treatment effect is explained by improvements in observed psychosocial skills or other alternate factors. Although these mediation models have value, they should be viewed as descriptive due to the inherent difficulties with unobserved mediators (Green et al. Reference Green, Ha and Bullock2010).

The results from these mediation models are shown in Figure 2. These provide clarity as to what was likely not behind Fast Track’s impact on turnout. First, Fast Track’s effect does not appear to be due to enhanced social status (as measured by marital status and social memberships). Although social status is strongly related to voter turnout, the Fast Track intervention did not alter this outcome. Further, the mediation models suggest that more traditional individual resources were also not behind Fast Track’s effect. The mediation estimates for income represent only 2.7% of the average direct effect (ADE), a comparatively small amount. Fast Track may have increased educational attainment some, and as such, this resource may have played a role in the program’s effect on turnout (about 5.6% of the ADE). Still, this by no means explains Fast Track’s entire effect. Likewise, the mediation models suggest that cognitive skills were unlikely behind Fast Track’s effect—with these accounting for only about 4.5% of the treatment effect. Finally, it may be tempting to argue that Fast Track’s effect is explained by the fact that children in the program simply received a lot of attention from adults. The mediation analyses show that this too is unlikely. Survey measures of subjects’ perceptions about the attention they receive from adults are included in the mediation models, and they explain no more than 4% of the treatment effect.Footnote 17 In total, these results suggest that Fast Track’s effect was about more than improving social status, individual resources, and cognitive ability, or simply providing increased attention from adults.

FIGURE 2.

Fast Track’s Potential Mechanisms

Figure 2 provides the mediation estimates that allow for interdependence across multiple mediators using KHB decomposition (Kohler et al. 2011). The mediators include registration status from validated voting records (N = 807); educational attainment in year 19 (N = 665); income in year 19 (N = 578); religious, nonreligious, social, and club memberships and marital status from year 15 (N = 350); the number of delinquent acts in years 8 through 13 based on public crime records (N = 746); scores on a verbal achievement test in years 5 and 6 (N = 665); emotion recognition (y2–y3), emotion regulation (y5–y9), grit (y3–y6), attentiveness (y2–y6), social problem solving (y2–y4), withdrawn (y2–y6), internalizing (y2–y6), antisocial (y2–y6), social problems (y2–y6), prosocial (y2–y9) and behavioral (y8–y13), hostility (y2–y4), aggression (y2–y4), and self-control (y1–y3) (N = 335); the child’s perception of how much their parents care about them (i.e., positive representation of parents, y6, N = 693); and the child’s perception of how much their teachers care about them from years 7 through 10 (N = 726).

Instead, the results suggest that Fast Track’s effect likely largely came from the bundle of psychosocial skills the program improved. The mediation models show that improvement in students’ measured psychosocial skills explains approximately 31.2% of the treatment effect. (Depending on the combination of measured psychosocial skills used, this number rises to approximately 52% of the total effect explained.Footnote 18 ) The size of this mediation estimate is consistent with previous work exploring the mediators of Fast Track’s impact on crime (Sorensen and Dodge Reference Sorensen and Dodge2016). Moreover, these mediation estimates are much larger than any of the other possible channels. If we think of voting as a two-step process—with the first step being registering to vote and the second being actually voting—Fast Track’s effects on registration explains about 44.5% of the effect on voting (see the first mediator shown in the left panel of Figure 2). In addition, observed psychosocial skills explain about 42.5% of the increase in registration (see the right panel of Figure 2). Among individual skills, the strongest mediators are general self-efficacy, empathy, and ability to controls one’s emotions and behavior (see Table A6 in the Online Appendix).

Still, a significant amount of the treatment effect remains unexplained, suggesting that there may be other unmeasured psychosocial skills of importance or downstream consequences from targeting psychosocial skills in childhood that drove this effect. For instance, it appears that some of the increase in turnout can be attributed to the fact that participants spent less time in prison. The mediation estimates represent about 8 to 12% of the average treatment effect—one of the larger individual mediation estimates. It is important to note, however, that Fast Track’s reduction of criminal behavior has been shown to be attributable, in large part, to upstream improvements in participants’ psychosocial skills (Sorensen and Dodge Reference Sorensen and Dodge2016).

These results suggest that the development of psychosocial skills played a large role in driving the Fast Track program’s effect on turnout. It appears that those who develop the general ability to control their emotions, motivations, thoughts, and behaviors, and to work in group settings, are more likely to target those skills in productive, mobilizing ways throughout their lives. In short, children who develop psychosocial skills early on are more likely to be set on a path toward active civic participation in adulthood.

CONCLUSION

In this article, I provided evidence that childhood programs targeting psychosocial or noncognitive abilities—the individual skills involving self-regulation and sociability not captured by cognitive ability—have a substantively meaningful effect on adult voter turnout. This result has important implications for our understanding of political participation. It suggests that what happens to children early in their lives can have a long-lasting impact on their rates of participation in adulthood. This finding suggests a reorientation of political socialization studies toward early childhood—a previously neglected critical period in the development of participatory predispositions.

Such reorientation should go beyond simple measures of cognitive ability. In the past, political scientists from many fields have focused narrowly on cognitive ability—from topics as diverse as civic participation (e.g., Verba et al. Reference Verba, Schlozman and Brady1995) to the formulation of political attitudes (e.g., White Reference White1968)Footnote 19 to even how people respond to political surveys (Gooch and Vavreck Reference Gooch and VavreckForthcoming). Although some observational work has started to explore the implications of psychosocial attributes (e.g., Condon and Holleque Reference Condon and Holleque2013; Fowler and Kam Reference Fowler and Kam2006, Reference Garmann2007; Hillygus et al. Reference Hillygus, Holbein and Snell2016; Uhlaner Reference Uhlaner1989), this work has been conducted in isolation. The results in this article suggest that psychosocial attributes should be viewed as part of the same taxonomy of attributes. Future work would do well to expand our understanding of how this bundle of attributes influences various political processes.

The results outlined here also have meaningful policy implications. They suggest that the specific childhood programs that schools implement—including those targeting individual psychosocial skills—appear to impact civic participation many years later. Further, they show that psychosocial attributes important for voting are not innate or immovable, but are instead malleable or teachable. At present, there is a major controversy in education circles surrounding how much attention should be given to helping children develop psychosocial skills (Farrington et al. Reference Farrington, Roderick, Allensworth, Nagaoka, Keyes, Johnson and Beechum2012). The results presented here add support to reforms that target these skills—providing evidence that psychosocial skills have benefits beyond those identified previously. This result is important, given that a common criticism of policies that encourage these skills is that they are prohibitively expensive. As as engagement in democracy is considered to be a critical mission of the public education system (Ravitch and Viteritti Reference Ravitch and Viteritti2001), the civic externalities documented here may help justify these costs. More generally, in a landscape of stubbornly low and unequal rates of voter participation and small estimates for many mobilization efforts, this finding should give policymakers and advocates for higher levels of civic participation from a more diverse electorate renewed hope and direction.

Still, much work remains to be done. In many ways, this work represents the first step in understanding the importance of psychosocial skills for political participation. For example, this intervention has focused its attention primarily on a group of relatively disadvantaged individuals. Given this group’s perpetually low rates of political participation (Leighley and Nagler Reference Leighley and Nagler2013), this focus is warranted. Future work would do well to also consider how well psychosocial interventions affect the broader population. Similarly, it would be beneficial for future research to formally explore what period in the life course psychosocial programs are most effective. More work also remains to be done to more narrowly evaluate the costs and benefits of the specific components of psychosocial programs and to explore empirical relationships between individual psychosocial skills and how they interact with political behavior.

That said, this article has documented that pursuing this agenda may provide quite fruitful. Expanding our view to include psychosocial attributes should help increase our understanding of why some people vote whereas others do not.

SUPPLEMENTARY MATERIAL

To view supplementary material for this article, please visit https://doi.org/10.1017/S0003055417000119

Footnotes

I wish to thank the National Science Foundation (SES-1416816) for its generous financial support. I would also like to thank Kenneth Dodge for providing invaluable feedback and assistance throughout this project; the other members of the Conduct Problems Prevention Research Group (CPPRG)—Karen L. Bierman, John D. Coie, Mark T. Greenberg, John E. Lochman, Robert J. McMahon, and Ellen E. Pinderhughes—for their very generous access to the Fast Track data; Jennifer Goodwin and Jeff Quinn for their help with the Fast Track data; and Marisa Abrajano, Nicholas Carnes, Anna Gassman-Pines, Christina Gibson-Davis, Donald Green, D. Sunshine Hillygus, Bradley Jones, Helen Ladd, Jan Leighley, Quin Monson, Marcos Rangel, Lucy Sorensen, Steven Sexton, Steven Snell, Jacob Vigdor, five anonymous reviewers, and participants at the Midwest Political Science Association (MPSA), the American Political Science Association (APSA), the Association for Public Policy Analysis and Management (APPAM), Boston University, Brigham Young University, Columbia University, Duke University, University of Virginia, Princeton University, University of Chicago, University of Tennessee, and Tufts University for their thoughtful feedback.

1 Even with some lamenting the “abandonment” of childhood (Sapiro Reference Sapiro2004, 1) and arguing for a “rebirth” of political socialization (Niemi and Hepburn Reference Niemi and Hepburn1995, 7), the fact remains that most political studies only start in late adolescence (e.g., Langton Reference Langton1969; Plutzer Reference Plutzer2002).

2 The Fast Track data used in this article is proprietary and confidentiality agreements prohibit disclosure, as the information contained therein includes sensitive information. However, interested scholars can apply for access through the Fast Track organization. For eligibility rules, restrictions, data security provisions, and how to apply to access the data, please visit http://fasttrackproject.org/request-use-data.php.

3 For example, previous research has not acknowledged that individual psychosocial attributes are conceptually and empirically related to one another (see Section 1 of the Online Appendix).

4 The appropriate term for this set of attributes remains controversial. These abilities have also been referred to as socioemotional skills, soft skills, character, emotional intelligence, social cognitive skills, and metacognitive learning skills (Farrington et al. Reference Farrington, Roderick, Allensworth, Nagaoka, Keyes, Johnson and Beechum2012; Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014).

5 There is some debate over how malleable psychosocial skills are and when in the life course they are most movable. If these skills are traits comparable to the Big Five, then psychosocial skills may have the properties that some scholars see inherent in the Big Five, marking them as “stable psychological characteristics” (Gerber et al. Reference Gerber, Huber, Doherty and Dowling2011, 265) that are “biologically influenced and enduring” (Mondak Reference Mondak2010, 6; see also Mondak et al. Reference Mondak, Hibbing, Canache, Seligson and Anderson2010). Conversely, psychosocial skills may be attributes that “are not set in stone at birth” but are instead teachable (Heckman and Kautz Reference Heckman, Kautz, Heckman, Humphries and Kautz2014, 413). At present, how best to characterize psychosocial skills remains unresolved because empirical evidence in this area is lacking—with interventions showing malleability, like the one I use in the following, being relatively rare.

6 Empathy is also referred to as emotion recognition or a theory of mind.

7 These are the standard inputs used in matching intervention (Sondheimer and Green Reference Sondheimer and Green2010) and survey data (Ansolabehere and Hersh Reference Ansolabehere and Hersh2012) to voter files.

8 At the time of matching to voting records, about 90% of individuals in the sample remained in their original intervention state. Overall, movement to an outside state was balanced across the treatment and control groups (p = .50). Among individuals located in other states, almost all were searchable through the individual states’ voter registration portals. Those few individuals who lived in unmatchable outside states were also balanced across the treatment and control groups (p = .99).

9 In 2012, approximately 42% of those low SES individuals aged 25 to 29 years were registered (Current Population Survey).

10 The standard assumptions of instrumental variables are likely satisfied due to the cluster-randomized assignment of the Fast Track program and the high degree of program compliance.

11 As expected, attendance declined over time (attendance in years 1–6 was 77%, 72%, 71%, 64%, 59%, and 53%, respectively).

12 See Table A2 in the Online Appendix for the covariate balance check.

13 To estimate the ITT, one need only substitute A s for $\hat{T_{i}}$ in the second stage of Equation (1).

14 In Figure A7 in the Online Appendix, I provide estimates by individuals’ baseline SES (based on Fast Track’s composite SES measure). I find some evidence that the effect estimate was larger for those with lowest SES in the sample and smaller for those of higher SES at baseline. Although these heterogeneous effects are not statistically significant (p = .30) and both point estimates are sizable on their own, there is a substantively meaningful difference between the two. This heterogeneity provides some suggestive evidence that if psychosocial skill interventions were scaled up, they might see their largest effects among disadvantaged, low-propensity citizens.

15 Although Sondheimer and Green (Reference Sondheimer and Green2010) use the Perry Preschool program to explore the role of educational attainment, the unexplored possibility remains that this program channeled its turnout gains through increases in psychosocial abilities. Indeed, Heckman et al. (Reference Heckman, Pinto and Savelyev2013) show that the Perry Preschool program increased participants’ psychosocial abilities, reducing externalizing behaviors and increased academic motivation. Increasing psychosocial abilities was not the primary purpose of the Perry Preschool program (unlike Fast Track); however, unintentional gains in these attributes may be behind the comparable, albeit slightly smaller, effect of this early life program on voter participation.

16 In the Online Appendix, I also include mediation models that examine potential mediators individually using the process outlined by Imai et al. (Reference Imai, Keele, Tingley and Yamamoto2011) (see Table A6).

17 The specific items asked whether parents/teachers “pay attention to me,” “listen to what I have to say,” “accept me as I am,” “care about me,” “can tell when I am upset about something,” “ask me [when they see something is wrong],” and “are proud of the things I do.”

18 This model includes emotion recognition and regulation, as well as behavioral, hostility, and aggression control.

19 As White (Reference White1968) bluntly put it, “the bright . . .child is likely to feel more politically efficacious than the slow . . .child” (731).

References

REFERENCES

Angrist, Joshua D., Imbens, Guido W., and Rubin, Donald B.. 1996. “Identification of Causal Effects Using Instrumental Variables.” Journal of the American Statistical Association 91 (434): 444–55.Google Scholar
Ansolabehere, Stephen, and Hersh, Eitan. 2012. “Validation: What Big Data Reveal About Survey Misreporting and the Real Electorate.” Political Analysis 20 (4): 437–59.Google Scholar
Astuto, Jennifer, and Ruck, Martin D.. 2010. “Early Childhood as a Foundation for Civic Engagement.” In Handbook of Research on Civic Engagement in Youth. Hoboken, NJ: John Wiley & Sons, 249–76.Google Scholar
Becker, Gary S., and Tomes, Nigel. 1986. “Human Capital and the Rise and Fall of Families.” Journal of Labor Economics 4 (3): S1–39.CrossRefGoogle ScholarPubMed
Brady, Henry E., and McNulty, John E.. 2011. “Turning Out to Vote: The Costs of Finding and Getting to the Polling Place.” American Political Science Review 105 (01): 115–34.CrossRefGoogle Scholar
Campbell, David E. 2006. Why We Vote: How Schools and Communities Shape Our Civic Life. Princeton, NJ: Princeton University Press.Google Scholar
Chetty, Raj, Friedman, John N., Hilger, Nathaniel, Saez, Emmanuel, Schanzenbach, Diane Whitmore, and Yagan, Danny. 2011. “How Does Your Kindergarten Classroom Affect Your Earnings? Evidence from Project Star.” Quarterly Journal of Economics 126 (4): 1593–660.Google Scholar
Condon, Meghan, and Holleque, Matthew. 2013. “Entering Politics: General Self-Efficacy and Voting Behavior Among Young People.” Political Psychology 34 (2): 167–81.Google Scholar
Corvalan, Alejandro, and Cox, Paulo. Forthcoming. “The Impact of Procedural Information Costs on Voting: Evidence from a Natural Experiment in Chile.” Political Behavior.Google Scholar
CPPRG. 1999 a. “Initial Impact of the Fast Track Prevention Trial for Conduct Problems: I. The High-Risk Sample.” Journal of Consulting and Clinical Psychology 67 (5): 631–47.Google Scholar
CPPRG. 1999 b. “Initial Impact of the Fast Track Prevention Trial for Conduct Problems: II. Classroom Effects.” Journal of Consulting and Clinical Psychology 67 (5): 648–57.Google Scholar
CPPRG. 2002. “Evaluation of the First 3 Years of the Fast Track Prevention Trial with Children at High Risk for Adolescent Conduct Problems.” Journal of Abnormal Child Psychology 30 (1): 19–35.Google Scholar
CPPRG. 2004. “The Effects of the Fast Track Program on Serious Problem Outcomes at the End of Elementary School.” Journal of Clinical Child and Adolescent Psychology 33 (4): 650–61.Google Scholar
CPPRG. 2007. “Fast Track Randomized Controlled Trial to Prevent Externalizing Psychiatric Disorders: Findings from Grades 3 to 9.” Journal of the American Academy of Child and Adolescent Psychiatry 46 (10): 1250–62.CrossRefGoogle Scholar
CPPRG. 2010. “Fast Track Intervention Effects on Youth Arrests and Delinquency.” Journal of Experimental Criminology 6 (2): 131–57.Google Scholar
CPPRG. 2011. “The Effects of the Fast Track Preventive Intervention on the Development of Conduct Disorder Across Childhood.” Child Development 82 (1): 331–45.Google Scholar
Currie, Janet, and Thomas, Duncan. 1995. “Does Head Start Make a Difference?” American Economic Review 85 (3): 341–64.Google Scholar
Dahl, Robert A. 1992. “The Problem of Civic Competence.” Journal of Democracy 3 (4): 45–59.Google Scholar
Dawes, Christopher, Cesarini, David, Fowler, James H., Johannesson, Magnus, Magnusson, Patrik K. E., and Oskarsson, Sven. 2014. “The Relationship Between Genes, Psychological Traits, and Political Participation.” American Journal of Political Science 58 (4): 888–903.Google Scholar
Dawson, Richard E., and Prewitt, Kenneth. 1968. Political Socialization: An Analytic Study. Boston, MA: Brown.Google Scholar
Dee, Thomas S., and West, Martin R.. 2011. “The Non-Cognitive Returns to Class Size.” Educational Evaluation and Policy Analysis 33 (1): 23–46.Google Scholar
Denny, Kevin, and Doyle, Orla. 2008. “Political Interest, Cognitive Ability and Personality: Determinants of Voter Turnout in Britain.” British Journal of Political Science 38 (02): 291–310.CrossRefGoogle Scholar
Duckworth, Angela L., Peterson, Christopher, Matthews, Michael D., and Kelly, Dennis R.. 2007. “Grit: Perseverance and Passion for Long-Term Goals.” Journal of Personality and Social Psychology 92 (6): 1087–101.Google Scholar
Farrington, Camille A., Roderick, Melissa, Allensworth, Elaine, Nagaoka, Jenny, Keyes, Tasha Seneca, Johnson, David W., and Beechum, Nicole O.. 2012. “Teaching Adolescents to Become Learners: The Role of Noncognitive Factors in Shaping School Performance—a Critical Literature Review.” Technical Report. University of Chicago CCSR.Google Scholar
Fowler, James H., and Kam, Cindy D.. 2006. “Patience as a Political Virtue: Delayed Gratification and Turnout.” Political Behavior 28 (2): 113–28.CrossRefGoogle Scholar
Fowler, James H., and Kam, Cindy D.. 2007. “Beyond the Self: Social Identity, Altruism, and Political Participation.” Journal of Politics 69 (3): 813–27.CrossRefGoogle Scholar
Garmann, Sebastian. 2016. “The Effect of a Reduction in the Opening Hours of Polling Stations on Turnout.” Public Choice 171: 99–117.Google Scholar
Gerber, Alan S., Huber, Gregory A., Doherty, David, and Dowling, Conor M.. 2011. “The Big Five Personality Traits in the Political Arena.” Annual Review of Political Science 14: 265–87.Google Scholar
Gerber, Alan S., Huber, Gregory A., Meredith, Marc, Biggers, Daniel R., and Hendry, David J.. 2015. “Can Incarcerated Felons Be (Re) Integrated into the Political System? Results from a Field Experiment.” American Journal of Political Science 59 (4): 912–26.Google Scholar
Gomez, Brad T., Hansford, Thomas G., and Krause, George A.. 2007. “The Republicans Should Pray for Rain: Weather, Turnout, and Voting in US Presidential Elections.” Journal of Politics 69 (3): 649–63.Google Scholar
Gooch, Andrew, and Vavreck, Lynn. Forthcoming. “How Face-to-Face Interviews and Cognitive Skill Affect Item Non-Response: A Randomized Experiment Assigning Mode of Interview.” Political Science Research and Methods.Google Scholar
Green, Donald P., Ha, Shang E., and Bullock, John G.. 2010. “Enough Already About Black Box Experiments: Studying Mediation Is More Difficult Than Most Scholars Suppose.” Annals of the American Academy of Political and Social Science 628 (1): 200–08.Google Scholar
Greenstein, Fred. 1965. “Children and Politics”. New Haven, CT: Yale University Press Google Scholar
Heckman, James J. 2000. “Policies to Foster Human Capital.” Research in Economics 54 (1): 3–56.Google Scholar
Heckman, James J., and Kautz, Tim. 2014. “Fostering and Measuring Skills: Interventions That Improve Character and Cognition.” In The Myth of Achievement Tests: The GED and the Role of Character in American Life, eds. Heckman, James J., Humphries, John E., and Kautz, Tim. Chicago, IL: University of Chicago Press, 341–430.Google Scholar
Heckman, James J., and Rubinstein, Yona. 2001. “The Importance of Noncognitive Skills: Lessons from the GED Testing Program.” American Economic Review 91 (2): 145–49.CrossRefGoogle Scholar
Heckman, James, Pinto, Rodrigo, and Savelyev, Peter. 2013. “Understanding the Mechanisms Through Which an Influential Early Childhood Program Boosted Adult Outcomes.” American Economic Review 103 (6): 2052–86.Google Scholar
Heckman, James J., Stixrud, Jora, and Urzua, Sergio. 2006. “The Effects of Cognitive and Noncognitive Abilities on Labor Market Outcomes and Social Behavior.” Journal of Labor Economics 24 (3): 411–82.Google Scholar
Hill, Seth J. 2016. “Following Through on an Intention to Vote: Present Bias, Norms, and Turnout.” Working Paper. http://www.sethjhill.com/SJH_PresentBias.pdf.Google Scholar
Hillygus, D. Sunshine, Holbein, John B., and Snell, Steven A.. 2016. “The Nitty Gritty: The Unexplored Role of Grit and Perseverance in Voter Turnout.” SSRN Working Paper (2675326). http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2675326.Google Scholar
Imai, Kosuke, Keele, Luke, Tingley, Dustin, and Yamamoto, Teppei. 2011. “Unpacking the Black Box of Causality: Learning About Causal Mechanisms from Experimental and Observational Studies.” American Political Science Review 105 (04): 765–89.CrossRefGoogle Scholar
Jackson, Kirabo C. 2012. “Non-Cognitive Ability, Test Scores, and Teacher Quality: Evidence from 9th Grade Teachers in North Carolina.” NBER Working Paper (w18624). http://www.nber.org/papers/w18624.Google Scholar
Kanfer, Frederick H. 1970. “Self Monitoring: Methodological Limitations and Clinical Applications.” Journal of Consulting and Clinical Psychology 35 (2): 148–58.CrossRefGoogle Scholar
Kohler, Ulrich, Karlson, Kristian Bernt, and Holm, Anders. 2011. “Comparing Coefficients of Nested Nonlinear Probability Models.” Stata Journal 11 (3): 420–38.CrossRefGoogle Scholar
Langton, Kenneth P. 1969. Political Socialization. Princeton, NJ: Princeton University Press.Google Scholar
Leighley, Jan E., and Nagler, Jonathan. 2013. Who Votes Now? Demographics, Issues, Inequality, and Turnout in the United States. Princeton, NJ: Princeton University Press.Google Scholar
Luskin, Robert C. 1990. “Explaining Political Sophistication.” Political Behavior 12 (4): 331–61.Google Scholar
Mondak, Jeffery J. 2010. Personality and the Foundations of Political Behavior. New York, NY: Cambridge University Press.Google Scholar
Mondak, Jeffery J., Hibbing, Matthew V., Canache, Damarys, Seligson, Mitchell A., and Anderson, Mary R.. 2010. “Personality and Civic Engagement: An Integrative Framework for the Study of Trait Effects on Political Behavior.” American Political Science Review 104 (01): 85–110.Google Scholar
Nie, Norman H., Junn, Jane, and Stehlik-Barry, Kenneth. 1996. Education and Democratic Citizenship in America. Chicago, IL: University of Chicago Press.Google Scholar
Niemi, Richard G., and Hepburn, Mary A.. 1995. “The Rebirth of Political Socialization.” Perspectives on Political Science 24 (1): 7–16.Google Scholar
Ojeda, Christopher. 2015. “Depression and Political Participation.” Social Science Quarterly 96 (5): 1226–43.CrossRefGoogle ScholarPubMed
Pacheco, Julianna Sandell, and Plutzer, Eric. 2007. “Stay in School, Don't Become a Parent: Teen Life Transitions and Cumulative Disadvantages for Voter Turnout.” American Politics Research 35 (1): 32–56.Google Scholar
Park, Daeun, Tsukayama, Eli, Goodwin, Geoffrey, Patrick, Sarah, and Duckworth, Angela. 2017. “A Tripartite Taxonomy of Character: Evidence for Intrapersonal, Interpersonal, & Intellectual Competencies in Children.” Contemporary Educational Psychology 48: 16–27.Google Scholar
Pettigrew, Stephen. 2016. “The Downstream Effects of Long Lines: How Long Waits at the Precinct Depress Future Turnout.” Working Paper. Presented at the 2016 Annual Meeting of the Midwest Political Science Association, Chicago, IL .Google Scholar
Plutzer, Eric. 2002. “Becoming a Habitual Voter: Inertia, Resources, and Growth in Young Adulthood.” American Political Science Review 96 (01): 41–56.Google Scholar
Plutzer, Eric, and Wiefek, Nancy. 2006. “Family Transitions, Economic Status, and Voter Turnout Among African-American Inner-City Women.” Social Science Quarterly 87 (3): 658–78.Google Scholar
Ravitch, Diane, and Viteritti, Joseph P.. 2001. Making Good Citizens: Education and Civil Society. New Haven, CT: Yale University Press.Google Scholar
Sapiro, Virginia. 2004. “Not Your Parents’ Political Socialization: Introduction for a New Generation.” Annual Review of Political Science 7: 1–23.Google Scholar
Schafer, Jerome Pablo. 2016. ‘‘Delayed Gratification in Political Behavior: Time Preferences Predict Turnout and Donations.” Working Paper (2756566). http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2756566.Google Scholar
Searing, Donald D., Schwartz, Joel J., and Lind, Alden E.. 1973. “The Structuring Principle: Political Socialization and Belief Systems.” American Political Science Review 67 (02): 415–32.Google Scholar
Sondheimer, Rachel M., and Green, Donald P.. 2010. “Using Experiments to Estimate the Effects of Education on Voter Turnout.” American Journal of Political Science 54 (1): 174–89.Google Scholar
Sorensen, Lucy C., and Dodge, Ken A.. 2016. “How Do Childhood Interventions Prevent Crime?” Child Development 87 (2): 429–45.Google Scholar
Uhlaner, Carole Jean. 1989. “Relational Goods and Participation: Incorporating Sociability into a Theory of Rational Action.” Public Choice 62 (3): 253–85.Google Scholar
Verba, Sidney, and Almond, Gabriel A.. 1989. The Civic Culture: Political Attitudes and Democracy in Five Nations. Thousand Oaks, CA: Sage.Google Scholar
Verba, Sidney, Schlozman, Kay Lehman, and Brady, Henry E.. 1995. Voice and Equality: Civic Voluntarism in American Politics. New York, NY: Cambridge University Press.Google Scholar
Wellman, Henry M., Cross, David, and Watson, Julanne. 2001. “Meta-Analysis of Theory-of-Mind Development: The Truth About False Belief.” Child Development 72 (3): 655–84.CrossRefGoogle ScholarPubMed
White, Elliott S. 1968. “Intelligence and Sense of Political Efficacy in Children.” Journal of Politics 30 (03): 710–31.Google Scholar
Wolfinger, Raymond E., and Rosenstone, Steven J.. 1980. Who Votes? New Haven, CT: Yale University Press.Google Scholar
Yang, Yaling, and Raine, Adrian. 2009. “Prefrontal Structural and Functional Brain Imaging Findings in Antisocial, Violent, and Psychopathic Individuals: A Meta-Analysis.” Psychiatry Research: Neuroimaging 174 (2): 81–88.CrossRefGoogle ScholarPubMed
Figure 0

TABLE 1.Fast Track’s Impact on Voter Turnout

Figure 1

FIGURE 1.Fast Track’s Impact on Voter Turnout

Panel (a) in Figure 1 shows predicted probabilities (distributions) and mean levels (dashed lines) of ever voting based a model with treatment status and baseline characteristics (i.e., the ITT). Panel (b) shows mean differences in predicted turnout for those fully exposed to the Fast Track program (i.e., the TOT). N = 812.
Figure 2

FIGURE 2.Fast Track’s Potential Mechanisms

Figure 2 provides the mediation estimates that allow for interdependence across multiple mediators using KHB decomposition (Kohler et al. 2011). The mediators include registration status from validated voting records (N = 807); educational attainment in year 19 (N = 665); income in year 19 (N = 578); religious, nonreligious, social, and club memberships and marital status from year 15 (N = 350); the number of delinquent acts in years 8 through 13 based on public crime records (N = 746); scores on a verbal achievement test in years 5 and 6 (N = 665); emotion recognition (y2–y3), emotion regulation (y5–y9), grit (y3–y6), attentiveness (y2–y6), social problem solving (y2–y4), withdrawn (y2–y6), internalizing (y2–y6), antisocial (y2–y6), social problems (y2–y6), prosocial (y2–y9) and behavioral (y8–y13), hostility (y2–y4), aggression (y2–y4), and self-control (y1–y3) (N = 335); the child’s perception of how much their parents care about them (i.e., positive representation of parents, y6, N = 693); and the child’s perception of how much their teachers care about them from years 7 through 10 (N = 726).
Supplementary material: PDF

Holbein supplementary material

Online Appendix

Download Holbein supplementary material(PDF)
PDF 669.7 KB
Submit a response

Comments

No Comments have been published for this article.