Citizenship education in American schools and its role in developing civic engagement: a review of the research

Few studies provide an overview of citizenship education from the primary to secondary grade levels in American schools. Citizenship education consists of specific teaching practices designed to encourage students to become more involved in their communities. This review critically evaluates three kinds of programmes related to citizenship education: (1) character education programmes, (2) political simulations and, (3) service-learning programmes. Students in the primary grades are mainly exposed to character education programmes, which emphasise the importance of developing ethical values. Political simulations are more common in high school civic courses, where students learn the importance of community-level civic engagement (e.g. volunteering). Service-learning programmes can help students in the secondary grades develop a broader range of civic engagement outcomes that pertain to the school and community-level context. This study reaffirms the importance of increasing students’ exposure to citizenship education, while emphasising that certain instructional practices can be more effective in helping students develop civic engagement.


Introduction
In the United States and many western European nations, the primary government model is based on representational democracy and characterised by regular elections (Haste 2010). Moreover, Haste (2010, 171) argues that "the ordinary citizen's primary access to power is via lobbying their representative, or using the media and pressure groups to influence public opinion and government policy". Citizens can be involved in the political process by being active participants in their local communities (Adler and Goggin 2005). For instance, a person can be civically engaged in the community by helping a neighbour or more broadly in the political arena by joining public demonstrations. It is important to examine the development of civic engagement amongst young people because these specific kinds of dispositions are likely to translate into his or her adulthood (Youniss et al. 2002). Community action has been problematic for various models of citizenship because these activities may not be directly related to party or voting activities. However, active participation in the school and neighbourhood context can help adolescents develop a critical understanding of political and social forces (Morgan and Streb 2001;Westheimer and Kahne 2004).
In the United States, various scholars and researchers have argued that there is no single conception of what constitutes good citizenship (Haste 2010). This study focuses on two citizenship goals drawing the most attention in the K-12 school system, which involves helping students develop personally responsible and participatory citizenship (Westheimer and Kahne 2004). The personally responsible citizen is mainly concerned with following laws, while the participatory citizen is inclined to think critically about various social issues. As a result of civic education standards, students from elementary to high school have opportunities to take coursework related to learning about democracy (Cogan 1999). Although high schools generally require students to take government courses (e.g. Civics), teachers in elementary classes can inform students about government and democracy during social studies lessons (Carnegie and CIRCLE 2003;Palonsky 1987). Althof and Berkowitz (2006) and others (Sears and Hyslop-Margison 2006) argue that teachers often rely on teacher-centred approaches to civic education, in which students focus on learning factual knowledge about government processes. However, citizenship education programmes can supplement mainstream social studies and civics courses to give students a participatory experience of learning about democracy.
Based on school or classroom-level reform, citizenship education primarily consists of three kinds of programmes: (1) character education programmes, (2) political simulations and, (3) service-learning programmes. The primary goal of citizenship education is to help students develop civic engagement at the school and community level (Carnegie and CIRCLE 2003). Thus, the goals of this paper are to understand the various programmes related to citizenship education in K-12 schools, and evaluate the extent to which these opportunities can develop students' civic engagement at the school and community level.

Conceptual framework
Three forms of citizenship education According to Hoge (2002), citizenship education refers to instructional strategies that promote democratic ways of thinking that foster informed and active citizenship. Campbell (2006) believes that citizenship education should be student-centred with a strong emphasis on classroom discussion and cooperative activities. Citizenship education is broad in nature and characterised by engaging students with different types of learning activities that relate to discussion of public issues and participation in simulated government activities (Althof and Berkowitz 2006). The next section examines citizenship education in its three forms: (1) character education programmes, (2) political simulations, and (3) service-learning programmes. These programmes consist of specific instructional practices that may be appropriate for students at certain grade levels.
Character education programmes are typically school-based reforms that focus on promoting ethical values (e.g. caring and honesty) to students (Altholf and Berkowitz 2006). These programmes are more common in the elementary grades 1 (Pearson and Nicholson 2000). According to Westheimer and Kahne (2004), character education programmes are closely tied with early stages of citizenship education. In these early stages of citizenship education, a young person learns about core values (e.g. honesty and trustworthiness) that are foundational to becoming a personally responsible citizen. The authors explain that a personally responsible citizen is someone who acts responsibly in their community by obeying laws. In contrast to older youth, young children may benefit more from character education because they are taught that communities foster a particular type of core values (Berkowitz 2000).
Another form of citizenship education consists of role play and political simulations. Students can learn about the political process by engaging in debates about political issues and mock elections that may require speech making and voting (Westheimer and Kahne 2004). Additionally, political simulations are focused on helping students to become participatory citizens. Participatory citizenship is mainly concerned with certain attitudes and dispositions related to thinking critically about social issues and finding solutions to various problems. Compared to character education programmes, political simulations may be more suitable for older students, in particular to those at the secondary level. 2 Hoge (2002) argues that older students are developmentally capable of understanding abstract concepts of democracy, which involves deliberating on complex social issues characterised by conflicting goals and interests.
The last form of citizenship education consists of service-learning programmes. Brandell and Hinck (1997) define service-learning programmes as academic curricula that assign students to various service projects addressing community needs. Similar to political simulations, this form of citizenship education is largely concerned with helping students to become participatory citizens. In terms of educational timing, these programmes may be more suitable for older students because they are more likely to have autonomy from their parents to participate in community service (Eccles and Barber 1999;Michelson, Zaff, and Hair 2002).
Civic engagement at the community and school level Leading experts in the field of citizenship education agree that American schools hold a civic mission to help students to become "competent and responsible citizens" (Carnegie and CIRCLE 2003, 6). Several scholars focus on a traditional conception of citizenship, which considers how youth are civically engaged at the community level (Adler and Goggin 2005;Janmaat 2011). Civic engagement at the community level refers to collective actions taken in the neighbourhood context, and can include everything from helping a neighbour to attending a town meeting (Adler and Goggin 2005).
Little of the work conducted in the citizenship education field, however, acknowledges an important civic mission goal that relates to helping students develop civic engagement at the school level (Carnegie and CIRCLE 2003). In fact, Hoge (2002) and Althof and Berkowitz (2006) acknowledges the relationship between citizenship education and school-level civic engagement, which refers to addressing problems in the school. Based on Dewey's (1916) notion that schools can serve as mini-polities or public spaces for collective action to occur, Flanagan et al. (2007) believe that students can engage in various civic activities related to school involvement. These civic activities can include volunteering, voting in school elections and joining extracurricular clubs. Several scholars argue that civic engagement should be broadened to include prosocial behaviour and school bonding (Callahan, Muller, and Schiller 2008;Obradović and Masten 2007). Prosocial behaviour reflects an early form of civic engagement that includes: students' interpersonal skills, conflict resolution, and group problem solving skills (Duke et al. 2009;Osterman 2000). School bonding or attachment is widely used to measure students' sense of school as a community (Diaz 2005;Flanagan et al. 2007). Based on the National Longitudinal Study of Adolescent Health, school bonding measured at high school strongly predicted students' voting in the national election after they became eligible voters (Callahan, Muller, and Schiller 2008). Taking these perspectives together, it is important to understand the degree to which citizenship education programmes can help students develop civic engagement at the community and school level.

Method
This study aims to analyse the research literature specific to citizenship education in American K-12 classrooms and evaluate how these practices foster civic engagement. Educating young people about citizenship is a multifaceted process that requires learning to develop a certain set of knowledge and dispositions related to participating in a democracy (Haste 2004;Janmaat 2011). Although there are evaluations of civic courses related to improving students' civic knowledge or recall of factual knowledge about government processes (Brabeck et al. 1994;Schultz, Barr, and Selman 2001), this paper is primarily concerned with evaluations that focus on students' civic engagement or behaviour related to collective action in the communities.
For this review, several criteria were used to select studies for inclusion in this sample: (1) interventions designed for K-12 students; (2) American school-based programmes; (3) at least one dependent variable related to civic engagement; (4) studies published after 1985. In a comprehensive search of peer-reviewed journals using the databases Education Research Complete (ERIC), Education index, Google search and PSYCINFO, the following keywords were used: citizenship education, character education, service-learning, political simulations, and civic engagement. References in these works were surveyed for additional relevant citations, which were then obtained from various disciplinary publications and evaluated. Thus, a total of 29 articles, chapters, and books are reviewed in this study.

Studies reviewed
Having established a framework for understanding citizenship education and its role in developing civic engagement, the next section is a review of the research. Table 1 is an outline of the research organised by citizenship education programme with a description and summary of the findings.

Character education programmes
Promoting Alternative Thinking Strategies, Open Circle Programme and Raising Healthy Children Various studies provide evidence suggesting that character education programmes, such as Promoting Alternative Thinking Strategies (PATHS), Open Circle Programme (OCP), and Raising Healthy Children (RHC) can help students develop civic engagement at the school level (Catalano et al. 2003 Kelly et al. 2004;Riggs et al. 2006). Table 2 is a summary of each character education programme with information about students' characteristics, programme features, and implementation scale. It is evident that most programmes require teachers to use curriculum material with an option for them to participate in teacher-education programmes. Further, these curriculum material feature lesson plans organised around interpersonal problem-solving and conflict resolution skills (Kam, Greenberg, and Kusché 2004).
The PATHS programme informs teachers on strategies related to helping students manage their emotions in various social situations (Curtis and Norgate 2007;Kelly et al. 2004;Riggs et al. 2006). For example, students have practice in verbally identifying and labelling feelings to help understand their peers' emotions through a curriculum lessons that use "Feeling Face" cards. Teachers are required to incorporate these 20-40 minute lessons three to five times a week (see Table 2). Riggs et al. (2006) compare second and third grade students who were enrolled in schools from middle and upper class communities that were randomly assigned to the PATHS programme. The comparison group did not receive exposure to the programme or similar character education programme. Regression analyses of teacherreported data reveal that students enrolled in the PATHS programme demonstrated greater school bonding in comparison with the control group. Curtis and Norgate (2007) and Riggs et al. (2006) estimate that the PATHS programme has an effect size (Cohen's d) of 0.09 to 0.19 on students' school bonding (see Table 1), which represents a small to medium effect size (Cohen 1988).
The OCP is a curriculum requiring teachers to implement weekly circle time activities (three times a week for 20-30 minutes a day) to engage students in discussion of various social problems. In a study of fourth grade students who were in classrooms that were randomly assigned to the OCP, programme students had a significant gain of over one standard deviation on measures related to prosocial behaviour in comparison with those who did not receive OCP or any similar interventions (Hennessey 2007). The RHC is another programme based on schoollevel reforms to promote school bonding (Catalano et al. 2003). First and second grade students participated in a study where schools were randomly assigned to implement the RHC programme (Catalano et al. 2003). Students in the control group had teachers who did not receive training in using the RHC curriculum. Results from the study indicate that students enrolled in the RHC programme reported greater school bonding in comparison with those in the control group. However, the lack of pre-and post-test data complicates these finding because of the possibility that students may naturally grow to like school, regardless of programme exposure.

Seattle Social Development Programme
The Seattle Social Development (SSD) programme has mentoring programmes that assign students to their peers for academic and social support. A number of studies report positive findings based on longitudinal data collected from a quasiexperimental trial of the SSD programme (Catalano et al. 2009;Hawkins, Doueck, and Lisner 1988;Hawkins et al. 2001). Catalano et al. (2009) assess data collected from a longitudinal study of students who were followed from first to sixth grade. Using regression analyses, the authors observe that the intervention had positive effects on school bonding, after controlling for students' race, socio-economic status, and residential mobility. The researchers (Catalano et al. 2009) report an estimated effect size (d) of 0.10 on school bonding (see Table 1), which is considered to be a small effect size (Cohen 1988). Hawkins et al. (2001) find that positive long-term effects on school bonding continued after students reached ages 13 and 18. However, these positive effects did not reach significance until students advanced to the fifth grade.

Child Development Project
Similar to the SSD programme, the Child Development Project (CDP) provides mentoring programmes called Buddies, which matches students from the upper grades to work with younger peers. In addition, the CDP promotes family and school bonding through its parent workshops, which is one key feature missing from other character education programmes. Families can participate in sessions that provide strategies on creating positive learning environments at home (Berkowitz and Bier 2005).
Various studies provide evidence showing that teachers trained to use the CDP can help students improve on various measures related to school-level civic engagement, including: school bonding and prosocial behaviour (Watson, Battistich, and Solomon 1998). The CDP was evaluated in a quasi-experimental study that randomly assigned elementary schools to treatment and control conditions (Watson, Battistich, and Solomon 1998). Solomon et al. (2000) report findings from a study that focuses on fidelity issues from three years of programme implementation. Based on an analysis of observations conducted in treatment and control schools, the authors report that teachers in the control schools were less likely to use cooperative learning activities and promote active discussion in comparison with teachers in the treatment schools. Most important, Solomon et al. (1996) report the highest effect size in comparison with any evaluations conducted on the programme. In a study conducted in schools located in low and high-income neighbourhoods, the researchers (Solomon et al. 1996) estimate that the CDP (see Table 1) has a medium effect size (d) that ranges from 0.40 to 0.49 on students' sense of community. Another study on the CDP has evidence suggesting that positive gains in students' civic engagement may persist into middle school (Battistich, Schaps, and Wilson 2004). Table 2 reviews the implementation scale and costs associated with the reviewed character education programmes. The CDP's strength lies in its use of teachereducation programmes that educate teachers on incorporating cooperative learning activities such as class meetings and discussions (Solomon et al. 1988). Although the start-up cost is not particularly high in comparison with other character education programmes ($6,000 USD per school), implementing extensive teacher and administrator training may take up to three years (Centre for the Study and Prevention of Violence 2009). The programme also requires school-parent workshops, which can help parents become more involved with their children's school life (Solomon et al. 1996). However, Cunningham, Bremner, and Boyle (1995) report that school-parent workshops entail substantial investments for schools and parents. These workshops require schools to train leaders and facilities to meet with parents individually, and parents must manage issues relating to travel time, extra child care, and home educational resources.
In sum, the reviewed studies provide evidence suggesting that character education programmes are directed towards helping students develop school-level civic engagement, as indicated by school bonding and prosocial behaviour. Moreover, it is noteworthy that character education programmes are more commonly found in the elementary grades. Among all the character education programmes reviewed, the research reflects strongest support for the CDP. The CDP has a moderately higher effect size on students' school bonding in comparison with all other programmes (see Table 1). Although the SSD programme may take up to three years to improve students' commitment to school (Hawkins et al. 2001), the CDP is effective for students within two years of programme exposure (Solomon et al. 1996).

Role play and political simulations
Curricula based on local government Role play and political simulations can help improve students' civic engagement at the community level (Kahne, Chi, and Middaugh 2006;McDevitt and Chaffee 2000). Table 3 is a summary of each political simulation programme with information regarding instructional practices, costs, and implementation scale. All the programmes require teachers to use curriculum material based on teaching about government with an option for them to attend teacher-education programmes. In particular, CityWorks and Project Citizens are year-round programmes that provide supplemental lesson plans for existing high school government courses (Hartry and Porter 2004;Vontz, Metcalf, and Patrick 2000). Teachers organise lesson plans (10-20 minutes per class period) to engage students in researching and debating policy issues relevant to their local governments. Kahne, Chi, and Middaugh (2006) report that students enrolled in government courses that were randomly assigned to implement the CityWorks curriculum reported greater inclination to volunteer in their communities in comparison with students in the control groups. Hartry and Porter (2004) analyse survey data from high school students who enrolled in classrooms that were randomly assigned to use the Project Citizens programme. Analyses of pre -and post-test data reveal that programme students reported higher levels of community-level civic engagement in comparison with those who enrolled in regular government courses. However, the assessment lacked specific details about students' engagement in certain types of civic activities, such as volunteering and voting.

Curricula based on national and state elections
Although the CityWorks and Project Citizens are year-round programmes that concentrate on local government, the Kids Vote and Student Voices programmes focus on the political action surrounding local, state, and national elections. In comparison with the year-round programmes, the Kids Vote and Student Voices programmes require students to participate in high levels of simulated legislative functions related to preparing arguments for debates, obeying certain speech-making procedures and engaging in voting processes (see Table 3 for more details). Additionally, these election-based curricula are designed to connect students with media and news outlets to access information on political candidates and issues (Syversten et al. 2009). McDevitt and Chaffee (2000) analyse survey data collected from high school students who enrolled in civics courses randomly assigned to use the Kids Vote programme. Controlling for students' demographic and academic performance, the authors find that programme enrolment was positively related to community-level Sources based on Middaugh (2002, 2006) and Vontz, Metcalf, and Patrick (2000). civic engagement (effect size of 0.08 to 0.14), in particular to students' initiating political discussion at home (see Table 1). In support of these findings, the authors explain that students increased their use of newspaper and Internet to gain knowledge about politics. In addition to mock debates and simulated congressional hearings, the Student Voices programme provides students with unique opportunities to communicate with political campaigns through the Student Voices website (Syversten et al. 2009). Two studies report positive findings on the Student Voices programme (Feldman et al. 2007;Syversten et al. 2009). Within one particular school district, schools were randomly chosen to implement the Student Voices programme. According to the authors (Syversten et al. 2009), random assignment at the school level was used to avoid situations where comparison teachers and students in the same school might be inadvertently exposed to the Student Voices programme. Social studies teachers in the control schools continued to use lessons consistent with their usual practices. In the first study, Syversten et al. (2009) find that programme students significantly improved on a number of community-level civic engagement measures. The researchers (Syversten et al. 2009) report an estimated effect size of 0.03 and 0.21 on students' voting interest and engagement in electoral politics, respectively (see Table 1). However, these positive impacts were not evident in less conventional forms of civic engagement, such as demonstrating and boycotting. The authors observe that the Student Voices programme can help students participate in conventional forms of civic engagement like voting, but not necessarily unconventional forms like protesting and boycotting. In another study, Feldman et al. (2007) pay particular attention to the programme's impact on students enrolled in diverse and urban schools. In addition to programme gains on students' civic engagement (effect size of 0.09 to 0.17), the study provides evidence suggesting that these programme gains were equal amongst White and non-White students.
One reason why the Kids Vote and Student Voices programmes seems to be more effective than the other programmes may be due to its emphasis on teacher development and media technology (see Table 3). The Student Voices programme requires significant investments in media-based support that includes providing computer terminals with Internet access and teachers knowing how to facilitate online discussion/debates amongst students regarding political issues. Research on information and communication technology in schools report that teachers' decision to implement technology in the classroom can be influenced by a number of different factors including: access to resources, quality of software, incentives to change and commitment to teacher-education programmes (Hobbs 2004;Mumtaz 2000). These costs can be substantial for schools to consider, especially for those located in low-income neighbourhoods.
In sum, these studies suggest that role play and political simulations are more commonly found in the high school grades. Also, these citizenship education programmes are more focused on helping students develop civic engagement at the community level, rather than at the school level. Comparisons of these four programmes reveal stronger evidence in support of the Student Voices programme. Evaluations of the Student Voices programme suggest slightly higher effect sizes on various measures of students' civic engagement (Syversten et al. 2009). Also, these programme gains on civic engagement were found to be equal amongst White and non-White students (Feldman et al. 2007). One reason why the Student Voices programme may be particularly effective is its unique strength of providing students with classroom Internet access to communicate with other students and political candidates (Feldman et al. 2007;Syversten et al. 2009). However, Syversten et al. (2009 caution that students enrolled in the Student Voices may be more politically efficacious because of the excitement and controversies that naturally surround national elections.

Service-learning programmes
Primarily offered for students at the secondary level, service-learning programmes are carefully monitored service experiences integrated in the academic curriculum and directed towards meeting community needs (Billig 2000). Service-learning programmes offer students a variety of projects related to addressing the following needs: academic (e.g. tutoring), health (e.g. serving food at a soup kitchen), and community (e.g. participating in neighbourhood clean-ups) (Reinders and Youniss 2006). Most of the studies in this literature report that service-learning programmes can improve civic engagement at the community level (Hart et al. 2007;Koliba, Campbell, and Shapiro 2006;Metz and Youniss 2005;Morgan and Streb 2001;Niemi, Hepburn, and Chapman 2000;Reinders and Youniss 2006;Schmidt, Shumow, and Kackar 2007). For example, Hart et al. (2007) analyse data collected from the National Educational Longitudinal Study with evidence suggesting that service-learning enrolment is related to a standard deviation increase in voting interest amongst high school students. In the study, high school students who participated in service-learning programmes were more likely to vote and volunteer in comparison with those not having exposure to these programmes. Also, several studies indicate that students enrolled in service-learning programmes reported increased civic engagement at the school and community level (Allen et al. 1997;Blyth, Saito, and Berkas 1997;Switzer et al. 1995). In a randomised study of middle school students assigned to a service-learning programme, Scales et al. (2000) report that the service-learning programme had an effect size of 0.07 to 0.09 on students' civic engagement (see Table 3). Students not only reported increased community-level civic engagement as indicated by neighbourhood volunteering, but also on a school-level measure based on school commitment.
Recent studies devote particular interest to comparing civic engagement outcomes for students enrolled in service-learning programmes on a voluntary basis with those enrolled on a school-required basis (Hart et al. 2007;Metz and Youniss 2005;Schmidt et al. 2007). Bennett (2003) hypothesises that mandated service might demotivate a student from engaging in future volunteering because he or she is required to help others. However, students enrolled in mandated service-learning programmes still experience improvements in civic engagement, as indicated by volunteering and voting interest (Hart et al. 2007;Schmidt et al. 2007). Although students who voluntarily join service-learning benefit more than those who enrol on a school-required basis (Metz and Youniss 2005), mandated service-learning programmes do not necessarily have adverse effects on students' future volunteering (Schmidt et al. 2007).
Other studies examine the extent to which effectiveness can depend on the type of service projects offered by the programme (Koliba, Campbell, and Shapiro 2006;Reinders and Youniss 2006). In an analysis of high school students enrolled in various service-learning programmes, Reinders and Youniss (2006) examine variations in the service projects based on contact with the people being assisted (direct/abstract) and community needs addressed (academic/health). For example, the authors discover that students who worked with homeless people at the shelter benefitted more in terms of civic engagement than those who engaged in raising money for charities. In addition, the degree of student reflection (reading, writing, and discussion before and after the service experience) may also influence the effectiveness of service-learning programmes. Scales et al. (2000) find that students who participated in service-learning programmes with more than average levels of reflection reported greater understanding of how schools are connected with their communities. Last, various studies indicate that an average exposure of more than 30 hours per year of service-learning programmes may be necessary to find positive effects on students' civic engagement (Allen et al. 1997;Scales et al. 2000).
In sum, the reviewed studies suggest that service-learning programmes are more commonly found in middle and high schools. Also, it is evident that servicelearning programmes are not only capable of helping students develop civic engagement at the school level, but also at the community level (Allen et al. 1997;Blyth, Saito, and Berkas 1997). Service-learning programmes that are particularly effective possess certain features, which include: reflection activities for participants (Scales et al. 2000), at least 30 hours of programme exposure (Allen et al. 1997), and direct contact with those being served (Reinders and Youniss 2006).

Discussion
This review of studies provides evidence suggesting that various citizenship education programmes from kindergarten to high school have the capacity to develop students' civic engagement. Consistent with theories on citizenship education theories, schools can provide civic learning opportunities that are appropriate to students' grade levels (Hoge 2002). A summary of these findings will explain how certain programme and teaching practices can help students develop civic engagement.
More commonly found in the elementary grades, character education programmes rely on activities related to cooperative learning and parental involvement to promote school bonding. Findings from these studies are consistent with theories regarding character education programmes and their capacity to improve students' school-level civic engagement (Althof and Berkowitz 2006). Among the programmes reviewed, there is stronger evidence in support of the CDP. Not only has research shown that the CDP is particularly effective for students enrolled in elementary schools located in low and high-income neighbourhoods (Solomon et al. 1996), but that these civic engagement gains may persist through middle school Wilson 2004 2004).
Studies on role play and political simulations suggest that these programmes can help students develop civic engagement at the community level. In addition to providing classroom discussion, students engage in opportunities to research issues and participate in mock debates. The Student Voices programme seems to be more effective in improving students' civic engagement in comparison with the other political simulations (Syversten et al. 2009). Syversten et al. (2009 conclude that the media richness of the curriculum (e.g. Internet and television) is critical to helping students gather information on current events. Programmes that offer a diverse number of media outlets can help students become informed on various problems and issues in the community. However, the Student Voices programme is timed to coincide with national and state elections occurring every two and four years. Thus, students should also have opportunities to participate in other political simulations (e.g. CityWorks), which can be used in government courses throughout the school year.
Although character education and political simulations appear to influence civic engagement at the school or community level, service-learning programmes can promote civic engagement at both levels (Allen et al. 1997;Blyth, Saito, and Berkas 1997;Switzer et al. 1995). One explanation is that students have opportunities to venture outside the classroom and experience direct connections with their communities. Hart et al. (2007) argue that service conducted in the community site involves collective action towards a mutual end, where youth and adults can work together. In terms of best practices for service-learning programmes, students need opportunities for reflection (written, oral or visual) to help place their actions within the context of their personal development of civic engagement (Eyler and Giles 1999;Koliba, Campbell, and Shapiro 2006).

Future research
Future research on character education programmes in schools should attend to a number of considerations. One concern among the reviewed studies relates to measuring school-level civic engagement, in particular to focusing on how students feel connected to school. This paper suggests that future studies should consider the quality of school-level civic engagement. For example, studies should consider how students engage in certain civic activities, which may include the type (e.g. volunteering and mentoring) and purpose (e.g. academic and social).
In terms of evaluations on role play and political simulations, future studies should examine the extent to which students' racial and social background influences their programme experience. For example, Flanagan et al. (2007) report that ethnic minority youth are less likely to trust government institutions in comparison with their ethnic majority peers (Gonzalez and Padilla 1997;Hepburn and Popwell 1992;McLaughlin 1993;Sanchez-Jankowski 2002). Thus, it will be important to consider how students' cultural background and understandings about democracy may influence their experience in an American model of citizenship education.
Several evaluations on service-learning rely on analysing students' self-reports that were administered before and after service-learning experience (Morgan and Streb 2001;Reinders and Youniss 2006). Self-reports leave open the possibility that more civically engaged students are inclined to join service-learning programmes or report higher civic engagement (Reinders and Youniss 2006). Although students' self-reports can be useful, Celio, Durlak, and Dymnicki (2011) suggest that this data can be complemented with other information collected from peers, teachers and independent observers.

Conclusion
The emerging research on citizenship education in American schools demonstrate important findings regarding how certain curricula can foster civic engagement amongst students. For primary grade students, character education programmes can help students learn the importance of community at school. Political simulations and service-learning programmes are more common in high schools, which have the potential to improve students' civic engagement at the school and community level. The strength of this study lies in addressing citizenship education theories, and evaluating the extent to which citizenship education may be effective in helping students develop civic engagement. Additionally, this study contributes understandings about specific programmes such as the CDP and Student Voices that may serve as models for future citizenship education programmes.
This study, however, is limited in understanding how citizenship education programmes may reflect certain values about citizenship that can influence students' experiences with democracy. In particular, civic education scholars have engaged in a long-standing debate about how certain instructional strategies promote assimilationist and multiculturalist values to educate students about citizenship (Gay 1997;Parker 1997). Also, this study is limited in addressing developmental theories in regards to civic engagement. Drawing on developmental theories may be particularly useful in providing a growth context of how youth progress from receiving civic values to the later stages of advocating for social change in their communities (Niemi and Sobieszek 1977). Future work is needed to explore how these perspectives can inform research in the citizenship education field.
Students enrolled in schools located in middle and high-income neighbourhoods may receive more quality opportunities for citizenship education in comparison with those in diverse and urban schools (Parker 2001). On speaking about this civic empowerment gap, Levinson (2010) believes that all schools have a role in ensuring that all students develop the knowledge and dispositions necessary for effective participation in the political process. The democratic health of a nation relies on the full participation of a large and diverse representation of citizens and not merely from a narrow segment of society. Notes 1. Primary or elementary education in the United States typically refers to the first six years of formal education. The first year of primary education is generally referred to as kindergarten (age five), with subsequent years referred to as first grade, second grade, and so forth. Students graduating in the last year of primary education (sixth grade) are normally 12 years old (Sen, Partelow, and Miller 2005). 2. Secondary education in the United States refers to the last six or seven years of formal education (middle and high school). Secondary education typically starts in middle school with students starting in sixth or seventh grade (age 11 or 12). High school begins with students in ninth grade (age 14) and progressing to twelfth grade (ending around age 18) (Sen, Partelow, and Miller 2005).