Teacher Education and Leadership 3-1-2018 Enhancing Citizenship Learning with International Comparative Research : Analyses of IEA Civic Education Datasets

Large-scale international databases provide valuable resources for scholars, educators and policy-makers interested in civic engagement and education in nations that are democracies or striving towards democracy. However, the multidisciplinary nature of secondary analysis of these data has created a fragmentary picture that limits educators’ awareness of relevant findings. We present a summary of research conducted across disciplines using datasets from two large-scale crossnational studies of civic education conducted by the International Association for the Evaluation of Educational Achievement (CIVED:99 and ICCS:09). The IEA studies KeywoRds citizenship education civic engagement political attitudes ICCS:09 CIVED:99

were conducted in more than 40 countries with nationally representative samples of 14-15 year olds.In a review of over 100 published articles reporting secondary analyses of these data, we identified four themes especially salient for citizenship educators: open classroom climates; teaching and learning approaches; student identity; and profiles of citizenship norms and attitudes.The review summarizes sample relevant articles to illustrate themes, emphasizes connections between education and civic engagement and suggests opportunities for future research.
Within the last two decades, the International Association for the Evaluation of Educational Achievement (IEA) has conducted large-scale civic education surveys designed to provide insights to teachers, researchers and educational policy experts.These included the Civic Education Study of 1999 (CIVED:99: Torney-Purta et al. 2001), the International Civics and Citizenship Education Study of 2009 (ICCS:09: Schulz et al. 2010) and the 2016 administration of the ICCS study (Schulz et al. 2017).These studies contained assessments of students' social and political attitudes as well as their knowledge, and included data from teachers and school leaders designed to provide additional information on the context of civic education in schools.Data from these studies, which are freely available, allow researchers to address research questions to better understand youths' civic identities and improve educational practices without having to construct an instrument, obtain permission to test a sample, and collect data.Indeed, these datasets contain reliable and valid measures from representative national samples of more than 40 countries that are larger and more generalizable than any single investigator could obtain.In short, data from these international studies of civic and citizenship education present a valuable resource for researchers to design and conduct analysis relevant to research questions of particular pertinence in their context or within their disciplinary or theoretical orientation.
Researchers working independently within a variety of disciplines from more than a dozen countries have conducted secondary analyses of these large-scale datasets.Such studies make significant contributions to broader discussions about civic education.However, the nature of these publications creates barriers to awareness of the findings and the potential of these datasets among educators and researchers.For example, political scientists, psychologists, sociologists, educators in the social studies or civics, and comparative education specialists published articles using these data in a variety of journals and described their research using different terms.These include political socialization, civic engagement or competence, political literacy, critical consciousness, and sociopolitical development.This creates difficulties for those attempting to find publications related to their interests, or to appreciate the expanding body of research stemming from these large-scale survey programmes.A summary can provide an impetus to utilize these findings, measures and data to better conceptualize effective civic engagement in youth -in particular the role of teaching and effective approaches to the process of learning in different contexts.
This article is organized around themes arising from a synthesis of published studies using CIVED:99 data and ICCS:09 data, including research conducted across many countries framed by a variety of disciplinary orientations.We focus on the IEA civic education studies for several reasons.First, the databases developed from these survey programmes have generated a considerable number of research publications, only some of which have been covered in recent reviews of citizenship education research (Castro and Knowles 2017) or of quantitative research in social and civic education more generally (Fitchett and Heafner 2017).Given their scope and complexity, these international large-scale civic education studies can be explored most appropriately in a focused review.Finally, by analysing these studies as a group, we can position the findings into larger discourses on civic education.These include understanding the impact of classroom context in promoting student learning (Hess and Avery 2008), the need for research addressing the contextualized notions of citizenship (Castro and Knowles 2017;Hahn 2010) and increased attention to the use of theory in understanding civic education (Avery and Barton 2017;Hahn 2017).With these three goals in mind we present the findings of the studies reporting CIVED:99 and ICCS:09 to inform future research and to promote more effective processes of citizenship education.
We begin by describing the IEA CIVED:99 and ICCS:09 studies.We then summarize and synthesize what has been learned about effective practices and contexts in civic education from examining approximately 100 articles reporting secondary analyses using CIVED:99 and/or ICCS:09 data (augmented with a small number of other sources to illustrate particular themes).This review started with an extensive annotated bibliography of research conducted using these datasets (Knowles and Di Stefano 2015).Two authors joined the team, which continued to examine reference and citation lists and to conduct further electronic searches to ensure complete coverage.We present findings and trends identified through this search and analysis, focusing on several substantive issues: open classroom climates, teaching and learning approaches, attitudes of students related to their social identities and students' norms of citizenship.We conclude by identifying some future research directions.In short, by presenting a summary of findings from IEA's civic education studies, our goal is to assist researchers and educators in developing a better understanding of teaching/learning processes as well as the role of national and of school contexts.Further, this review can assist them in formulating novel, interesting and feasible research questions that could be addressed either through secondary analyses of these and similar data sources (including ICCS:16) or through smaller-scale projects that could deepen understanding of the civic education process.

an oveRview of iea's laRge-sCale inTeRnaTional suRveys in CiviC eduCaTion
The IEA started in 1958 with a group of researchers in comparative education who were interested in effective approaches for student learning.A history is available at IEA (n.d.a); see also Pizmony-Levy (2013).The first study completed by IEA was a twelve-country pilot study conducted in 1960; a study in mathematics achievement was conducted later that decade.In 1971, IEA conducted a cross-national civic education study as part of their 'six-subjectarea study'; this included measures of content knowledge and of attitudes such as trust in government, support for women's political rights and participatory behaviour (Torney et al. 1975).Results from this early study indicated that an open classroom climate for discussion was a key predictor of civic knowledge and civic engagement: a finding replicated in analyses of subsequent waves of data.It is notable that to this day, the civics studies remain the only IEA studies to place as much emphasis on assessments of attitudes as on knowledge (Torney-Purta et al. 2010).
After the initial IEA study of civic education, interest in international and comparative citizenship education waned for nearly two decades until the collapse of the Berlin Wall brought new attention to preparation for living under democratic rule (Torney-Purta and Amadeo 2013a).In response to requests from post-Communist member countries, the IEA launched the CIVED:99 study.Western European countries also expressed interest because of declines in youth civic participation.The framework for the assessment of both age groups was based on input from the study's cross-national steering committee and on case studies of 24 countries (Torney-Purta et al. 1999).It focused on civic knowledge and attitudes towards democracy and citizenship, national identity and international relations, as well as social cohesion and diversity.The study considered the nested nature of young people's experiences as they develop civic identities during interactions at home, with peers and at school.The conceptual framework guiding this study, called the 'Octagon Model' (Torney-Purta et al. 2001), incorporated aspects of ecological systems theory (Bronfenbrenner 1979) and models of situated cognition (Lave and Wenger 1991).After consultation with participating country representatives and with IEA experts in measurement, data were collected from approximately 90,000 14 year olds from nationally representative samples of schools in 28 countries in 1999 (Torney-Purta et al. 2001); over 50,000 upper secondary school students from 16 countries responded to parallel test and survey instruments (Amadeo et al. 2002).In addition, surveys were administered to principals and to teachers of civic-related subjects (e.g.civic education in some countries and history in others) in the 14-year-old students' schools.
In 2009 IEA launched ICCS:09 (Schulz et al. 2010), which used a framework that shared a number of commonalities with the framework for CIVED and contained many but not all of the same measures (Barber and Torney-Purta 2012).This study included data from more than 140,000 eighth graders in 38 countries and also surveyed principals and a random sample of teachers (from a variety of subject areas not only civic-related subjects) in students' schools.The study addressed research questions about changes in content knowledge since 1999, students' interest in engaging in public and political life, perceptions of threats to civil society, the features of educational systems and classrooms related to civic and citizenship outcomes, and aspects of students' backgrounds.Some countries that had participated in CIVED:99 did not participate in ICCS:09 (including, Australia, Germany, Hungary and the United States).However, ICCS:09 included many countries that had not participated before (particularly in Asia and Latin America).To meet the needs of specific areas of the world, the ICCS:09 study also included three regional modules: the Asian Module (Fraillon et al. 2012), the Latin American Module (Schulz et al. 2011) and the European Module (Kerr et al. 2010).Finally, the ICCS:09 study also differed from the CIVED:99 study in employing more contemporary approaches to the measurement of subject-area knowledge developed in other IEA studies such as those in mathematics and science.In brief, the test designers employed a matrix-sampling strategy in which different students were administered different sets of test items in the interest of increasing content topic coverage (Schulz et al. 2013).
IEA releases the data for all researchers to use via the IEA data repository (IEA n.d.b) as soon as is feasible after primary analysis reports appear.The data from all IEA studies are available free for secondary analysis from the IEA organization and, in the case of the CIVED:99 survey of 14-year-old students, from the University of Michigan's Interuniversity Consortium for Political and Social Research (ICPSR, 2008) and in particular through the CivicLEADS Project (CivicLEADS n.d.). 1 ICCS:09 data are scheduled for archiving by early 2019.Bibliographies and analysis tools are also available on that site.Articles reporting secondary analysis usually begin to appear in books and peer-reviewed journals about two to three years after data collection.These publications are generally more focused on specific research questions than the broader primary analysis reports published by IEA, which present basic comparisons (e.g., between countries and between male and female students).

MeThod of The Review
As previously stated, the themes presented here stem from a broader systematic review of published secondary analyses of CIVED:99 and ICCS:09 data.The approach to this review was chosen in light of two perspectives contained in recent review articles.First, there has been interest in examining secondary analyses of large-scale datasets across subject areas.For example, a systematic review of publications using data from the IEA Progress in Reading Literacy Study (PIRLS) pointed to under-examined educational issues for which relevant data were available (Lenkeit et al. 2015).Second, a review by Geboers et al. (2013) demonstrated the utility of reviewing studies of civic education in particular.Small-scale studies were a focus of their review of 28 articles, which reported on the effects of civic education for secondary-school students.The authors concluded that particular aspects of schools such as student diversity and classroom contexts play an important role in citizenship education.
Our process for conducting this systematic review involved the following steps: formulating inclusion criteria, searching literature and synthesizing research around themes.Publications were included in this review if they contained published secondary analyses of data from CIVED:99 or ICCS:09 in English-language peer-reviewed academic journals.We began the review by searching for the names of the two survey programmes as keywords (including abbreviations such as 'Civ-Ed', an early formulation of the title for CIVED:99).We scanned reference lists of identified articles and used databases (Google Scholar and ERIC) to find additional articles.Knowles and Di Stefano (2015) published the first results of this search in an annotated bibliography of publications drawing from CIVED:99 or ICCS:09 data.The annotated bibliography included brief descriptions starting with the article's purpose, sample and countries analysed, key research questions and substantive findings.However, that review also included conceptual as well as empirical publications, and included some book chapters as well as journal articles.
A next step was to seek articles that were not in the annotated bibliography, some published after that search concluded.We included a widened range of keywords and search terms.Some works came to our attention when we initiated communication with scholars familiar with international citizenship research; others came from close examination of reference lists.This search identified about 50 additional research articles reporting secondary analyses of ICCS:09 or CIVED:99 data.The total number of authors and co-authors across all 101 studies was 116, and they were from universities or research institutes located in Australia, Austria, Belgium (Flemish), Chile, Denmark, Estonia, Germany, Hong Kong, Ireland, Israel, Italy, the Netherlands, Norway, Portugal, South Korea, Sweden, Switzerland, the United Kingdom and the United States.We reviewed and entered each article (and its journal source) onto a separate line of a spreadsheet and described it with a column for discipline of the senior author.Other columns coded the theoretical or conceptual approach (when present), dataset used, countries analysed, keywords, independent and dependent variables.This chart and tabulations from it will serve as the basis for the major analysis of this systematic review.The earliest entries were published in 2003, four years after the completion of the CIVED:99 study and two years after the release of the international report (Torney-Purta et al. 2001).The most recent studies included in this review are from 2017.For the purposes of identifying themes for this review, we tallied the educationally relevant aspects of the articles' analyses, results and conclusions.

findings fRoM The Review
Our review revealed that secondary analysts employ a diverse array of theoretical and methodological approaches relevant to their particular purposes and disciplinary backgrounds.For example, some political scientists used Dalton's (2008) distinctions between duty-based and engaged citizenship (Hooghe and Oser 2015;Hooghe et al. 2014), while social policy experts and others interested in social psychology employed contact theory (Collado et al. 2015;Janmaat 2014).A variety of theories from developmental psychology have informed secondary analysis of CIVED:99 and ICCS:09 data, including the developmental niche model (Barber et al. 2015;Torney-Purta and Amadeo 2011) and the communities of practice framework (Hoskins 2012).In general, the most useful theoretical approaches appear to be those falling into the general heading of social-cognitive and sociocultural models (Wilkenfeld et al. 2010).Several recent studies used theories relating to social justice and critical consciousness with the CIVED:99 data (Diemer and Rapa 2016;Godfrey and Grayman 2014;Grayman and Godfrey 2013).In terms of methodological approaches, secondary analyses have employed single-level and multi-level regression models, structural equation models, cluster analyses or latent class/ latent profile models and IRT-based approaches, among others.This variety of approaches increases the value of the conclusions (Torney-Purta and Amadeo 2013b).
Based on our review we identified major concepts relevant to teaching citizenship: open classroom climates, teaching and learning approaches and aspects of student identity related to background characteristics.Within each section, relevant dependent variables are highlighted.These include civic knowledge, expected participation and students' attitudes towards political institutions and towards rights for immigrants and political rights for women.In addition, a final section focuses on studies that did not fit within the typical independent variable/dependent variable research design; these were primarily studies using person-centered analysis (e.g.latent class analysis or cluster analysis).These studies were concerned primarily with students' citizenship norms, participation, or attitudes.While we cover an array of studies here, including each of the approximately 100 works would have been impractical.We identified four themes addressed across the articles reviewed and then focused on summarizing works that made especially substantive contributions to subsections organized around these themes.Occasionally, sources other than journal articles (including book chapters and reports) are also included to illustrate the relevance for educators of themes arising from the review of articles.

Theme 1: Open climates and classroom contexts
A single finding consistently emerged across countries, contexts, times and groups: that an open, participatory and respectful discussion climate is associated with civic knowledge and engagement.Indeed, we counted 38 studies that demonstrated a positive relation between open classroom climates and a positive civic outcome.Although the ICCS:09 and CIVED:99 studies provide several measures of civic learning opportunities (including the perceived influence of students on school decision-making), the students' perception of an open and respectful climate for classroom discussion has been most frequently studied and has shown the most consistent positive association with valued civic outcomes across countries.Research showing the value of an open classroom climate dates back to the first IEA study of civic education more than 40 years ago (Torney et al. 1975) and continues with recent publications such as Godfrey and Grayman (2014) and others.
Providing context for the role of classroom climate, Gainous and Martens (2012) analysed CIVED:99 data from the United States and identified four broad teaching approaches implemented by teachers.These included traditional teaching, active learning, use of video and fostering of an open classroom climate for discussion.Their analysis compared different combinations of these approaches and found that an open classroom climate when combined with any of the other three was associated with higher levels of civic knowledge, internal/external political efficacy and expected voting.All of the combinations that omitted open classroom climate appeared less effective.As a result, the authors posited that an open classroom climate was a necessary element of effective civic education.Later, building upon their earlier study, Martens and Gainous stated 'the unmistakable conclusion to be drawn from our research is that fostering an open classroom climate is the surest way to improve the democratic capacity of America's youth ' (2013: 18).In other words, an open and respectful classroom climate for discussion of issues was indispensable to generating democratic capacity as measured by civic knowledge and participation in the United States.
A similar analysis in a report commissioned by the Public Education Division of the American Bar Association examined open classroom climate and looked simultaneously at the use of lectures as instructional methods in the United States (Torney-Purta and Wilkenfeld 2009).The outcomes analysed included measures in the CIVED:99 instrument that corresponded to definitions of twenty-first century competencies, including civic knowledge (with special attention to media literacy) and a series of attitudinal and participatory outcomes.These authors found an open classroom climate to be an essential aspect in fostering these positive civic education outcomes in the United States.The combination of lectures with an open class climate was also effective.In other words, these were not antithetical approaches.However, reliance on frequent lectures without also having an open and respectful discussion climate appeared to be significantly less effective in promoting the knowledge and attitudinal outcomes examined.In other words, like Martens and Gainous (2013), these authors found no contradiction between structured methods and less structured open discussion.
Across a number of countries, researchers have connected this experience of open classroom climates to a variety of positive outcomes: civic knowledge, expected political behaviour and supportive attitudes regarding gender, ethnic and immigrant rights (Quintelier and Hooghe 2013, in 35 ICCS:09 countries; Torney-Purta et al. 2008, in 27 CIVED:99 countries).For example, Maiello et al. (2003) found that a climate for discussion where students' opinions are respected is positively associated with expected political participation, including expectations of voting, across 28 countries.Similar findings have been demonstrated in region-based or single country studies focusing on Latin America (Edwards 2012;Garcia-Cabrero et al. 2017;Trevino et al. 2016), western Europe (Knowles and McCafferty-Wright 2015;Torney-Purta and Barber 2005), Ireland (Cosgrove and Gilleece 2012), Italy (Alivernini and Manganelli 2011), Thailand and Hong Kong (Kennedy 2012) and the United States (Campbell 2007(Campbell , 2008;;Zhang et al. 2012).Torney-Purta and Barber (2005) analysed data from the CIVED:99 study from several European countries to look at whether several forms of democratic participatory learning were predictors of informed voting and community participation.This involved openness of classroom climate as well as students' perceived opportunity to learn about voting, perceived opportunity to learn about cooperating with diverse groups, participation in a student council and confidence in the effectiveness of school participation.They found that democratic participatory learning was positively associated with expectations of informed voting and of community participation.
Subsequent analyses have considered contextual factors based on school characteristics and students' social identity that may relate to the experience of an open classroom climate.Barber et al. (2015) explored variation of class climate perceptions among students within the same classroom in order to assess whether students' responses to questions about climate could be aggregated reliably to the classroom level.These authors demonstrate that classroom-aggregated assessments of open classroom climate were modestly reliable and did predict some civic outcomes, which demonstrates that a teacher can implement an open classroom climate with meaningful results across students.However, some students reported feeling a stronger sense of an open and respectful classroom climate than others within the same classroom; these differences in perceptions were associated with prior civic education experiences and gender.
Researchers have explored how open classroom climate may play a different role for different groups of students in the United States.Campbell (2008) found an open classroom climate was more strongly associated with higher scores on measures of appreciation of conflict and expectations of voting for lower socio-economic students than for those from higher-socio-economic environments.He concluded that an open classroom climate has an especially positive impact on students who do not expect to finish high school.Gainous and Martens (2012) built on Campbell's finding by examining whether the contribution of an open classroom climate varies based on the students' Home Environment Index, based on expected education, parental education, number of books at home, political discussions with peers and political discussions within the family.These authors found that open classroom climate has a strong association with democratic capacity for students scoring relatively low on the Home Environment Index, and a somewhat weaker relationship among more privileged students.Moreover, Wilkenfeld and Torney-Purta (2012) reported a similar study focusing on school SES obtained from US census data, and demonstrated that civics education experiences had a stronger association with civic knowledge within schools in high poverty neighbourhoods.
A number of avenues for secondary analysis in the area of school experience are open for further exploration.In addition to the measure of open classroom climate, several other scales deserve consideration.For example, CIVED:99 and ICCS:09 both included items that form a scale measuring students' sense of collective efficacy at school (also in the datasets), for example, whether respondents believe that students can band together to solve school problems.CIVED:99 and ICCS:09 also included items about students' memberships in school organizations (including student council and environmental organizations, to list just two).ICCS:09 included some additional scales that could serve a function similar to classroom climate, including the existence of positive relationships between teachers and students.Some secondary analysis has found this to be an important predictor of civic competencies (Isac et al. 2014).

Theme 2: Approaches to teaching and learning
In addition to research on an open classroom climate, several studies address alternative ways of organizing teaching and learning within school contexts.These studies provide nuance to analysis of these datasets by considering different notions of teaching, how teacher beliefs manifest in the classroom and how civic knowledge is understood.
The first group of studies considers results from the teacher surveys administered as part of the CIVED:99 and ICCS:09 studies.Of particular interest to educators, Alviar-Martin et al's (2008) analysis of CIVED:99 data from Germany, Italy, Hong Kong and the United States investigated perceived strengths and weaknesses in teachers' confidence in teaching about a variety of topics including democracy and democratic practices, national identity and international relations, and social cohesion and diversity.The study's findings, which relied on an IRT-based analysis of differential item functioning, are too complex for a detailed review here.In summary, within the United States, teachers were relatively more confident teaching about citizens' rights and obligations, environmental issues and civic virtues and relatively less confident teaching about international problems, cultural differences and minorities and the dangers of propaganda.Overall, the study provides evidence that many teachers felt unprepared to enact an issues-centered curriculum.
Other studies in this group considered the ramifications of teachers' beliefs.Chin and Barber (2010) used CIVED:99 data to explore correlations between teacher-level measures in Australia, England and the United States.They concluded that teachers viewed engagement-based civic activities as necessary components of citizenship.More recently, Gainous and Martens (2016) used US data from CIVED:99 to explore whether teachers' beliefs relate to whether teachers implement an open classroom climate and to students' civic knowledge.They found that teachers who personally felt strongly about environmental and human rights issues were more likely to endorse an open classroom climate than were teachers who felt strongly about the importance of conforming to the law, serving in the military and patriotism.Students taking courses with the former group of teachers had higher levels of civic knowledge, which the researchers attributed to the higher levels of open classroom climate.This study demonstrates that underlying dispositions may be a contributing factor in teachers' instructional decision-making.Their study is a rare example of analysis that successfully linked student and teacher data from CIVED:99.Torney-Purta et al. (2005) is another example, as is the first article covered in the next paragraph.
A second, smaller set of studies has provided analysis of students' learning of diverse aspects of civic knowledge.Zhang et al. (2012) called into question the conceptualization of civic knowledge as a unidimensional measure by examining the 38 questions designed to assess students' civic knowledge using the US data from the CIVED:99 study.They identified four distinguishable groups of items assessing (1) basic conceptual knowledge; (2) advanced conceptual knowledge and reasoning; (3) reasoning about and analysing media graphics and material about issues in the media; and (4) reasoning and analysing opinion and applying principles in synthesizing factual knowledge.Using a statistical technique called cognitive diagnostic modelling, they found that the acquisition of basic conceptual knowledge appears to be necessary to acquire advanced conceptual knowledge but appears to be less important in acquiring political skills/reasoning.The use of conceptual teaching and student engagement, such as an open classroom climate, were predictors of cognitive civic achievement.Building on this study, Arensmeier (2016) conducted interviews with Swedish students about individual ICCS:09 knowledge items, dividing them into conceptual knowledge and more basic civic literacy.Her qualitative analysis shows the need for civic education focused on conceptual teaching through questioning and discussing complex civic issues and political principles from different perspectives.These studies are relevant for educators because they suggest that conceptual teaching helps students acquire concepts and skills more effectively than rote methods of instruction.

Theme 3: Attitudes and knowledge of students related to their social identities
The studies reviewed in this section emphasize students' identities, specifically related to gender (male/female), ethnicity and immigrant status.While nearly every researcher considers these variables, they are often employed only as controls.It is helpful in addition to explore differences among groups in depth, including the extent to which school and community experiences vary in their importance depending on aspects of students' identities.

Gender
Explorations of differences between male and female students have been compelling within and across countries (Barber and Torney-Purta 2009;Campbell and Wolbrecht 2006;Hooghe and Stolle 2004).Exploring gender differences, a book chapter by Barber and Torney-Purta (2009) evaluated levels of political efficacy and attitudes towards women's rights using CIVED:99 data from 28 countries.Their study also considered national contexts such as level of economic development, national expectations for further education and per cent of students perceiving gender-based employment inequality.They found gender gaps, with female students being more supportive of women's rights and male students having higher levels of internal political efficacy.The perceptions of more equal opportunities for women and girls within a society partially reduced the size of the gender gaps in support for women's rights and in internal political efficacy.Furthermore, open classroom climate appeared to be especially effective in promoting male students' support for women's political rights.Wolbrecht and Campbell (2007) worked in a similar area using CIVED:99 data and examined the role of the country context.They found a smaller gender gap in support for women's rights in countries where there were higher proportions of women in parliament (or its equivalent).Other studies focused on expected behaviour.Hooghe and colleagues (Hooghe and Stolle 2004, using US CIVED:99 data; Hooghe and Dassonneville 2013, using ICCS:09 data from 22 European countries), explored gender differences in expected political participation.They found that female students said they were more likely to vote and support social movement activism, while male students were more likely to become a candidate, join a political party and engage in potentially illegal protest behaviours.Using US data from CIVED:99, Grayman and Godfrey (2013) found that female students expressed more support for social justice.Gilleece and Cosgrove (2012) considered civic participation at school using Irish data from ICCS:09 and found that male students had lower levels than female students, but among male students civic participation at school varied with perception of the extent of students' influence on decision making at school.
In addition, Pereira et al. (2015) explored several dimensions of knowledge using data from the ICCS:09 European module and found that female students excelled in reasoning/human rights knowledge, while male students scored higher on factual knowledge relating to the European Union.Similarly, Torney-Purta et al. (2008) analysed knowledge of human rights items in 27 countries participating in the CIVED:99 study.After controlling for overall civic knowledge, female students were more knowledgeable about the purpose of the Universal Declaration of Human Rights and the United Nations' Convention on the Rights of the Child, while male students were more knowledgeable about the overall purpose of the United Nations.
Finally, Husfeldt (2006) examined groups of students with extremely negative views of immigrants, discovering that the ratio of male to female students was 3 to 1.A similar gender pattern was identified in the person-centered cluster analysis reported by Torney-Purta and Barber (2011).About 10 per cent of students in five western countries (including the United States) formed a cluster with very low levels of trust in government and extremely negative attitudes towards women's rights, ethnic group rights and immigrants' rights.These students were predominantly male.Recent analyses that have examined CIVED:99 and ICCS:09 also find consistent gender differences (favouring female students) in attitudes towards the political rights of marginalized groups (Barber and Ross 2017).

Race and ethnicity
Another set of studies considered variations based on racial or ethnic identity.These studies typically examined data from the US sample of CIVED:99 participants, given the importance of race/ethnicity in the US context and the inclusion of questions pertaining to the respondent's race and ethnicity in the nationspecific component of the survey (ICPSR, 2004).Similar information is available in some other countries where issues of national identity are particularly salient.Torney-Purta et al. (2007) found that Latino students in the United States, in comparison with non-Latino students, demonstrated lower civic knowledge and expected civic participation.However, accounting for differences in experiences with an open classroom climate and discussion of political topics in and out of the classroom narrowed this gap.The authors suggest that ensuring equitable access to an open classroom climate that explicitly includes the study of political topics could promote more equal outcomes.In another analysis, Campbell (2007) used US data from CIVED:99 to examine the role of classroom diversity in establishing open climates for discussion that might in turn increase expectations for future political activity.Students in more racially diverse classrooms reported less open climates for discussion.Both white and African-American students perceived classrooms as most open when there were high proportions of students in class sharing their racial identity.Findings such as these could be explored using other methodologies.Turning to the Israeli context, Ichilov (2005) examined data from the upper-secondary cohort of the CIVED:99 study and compared attitudes between Israeli Palestinian Arabs and Jewish students.She found that the Arab students expressed less identification with the national anthem and flag and were not as proud of Israel's achievements as were the Jewish students.
Recent studies have introduced the concept of critical consciousness to the analysis of these data.There are three examples in secondary analysis of CIVED:99 data from the United States.Godfrey and Grayman (2014) conceptualized critical consciousness as consisting of three components parallel to those put forth by Watts and colleagues (e.g.Watts et al. 2011): the ability to critically interpret social conditions (critical reflections), feelings of efficacy to effect change (sociopolitical efficacy) and actual participation in these efforts (critical action).Diemer and Rapa (2016) also employed notions of critical consciousness to explore perceptions of societal inequality and beliefs that society should be more egalitarian among African American and Latin American students in the United States whose mothers had lower levels of education.They found that these dimensions of critical reflection predicted expected voting, conventional political action and critical action relating to protest.This study pays special attention to political efficacy, or belief in the responsiveness of government and one's own political agency.In a related study, Grayman and Godfrey (2013) found that non-white students perceived more inequality and endorsed more government economic responsibility in comparison to white students in the United States.By acknowledging that some communities perceive more discrimination and believe that the government is not responsive, this study provides a model for bringing critical perspectives to bear on analysis of large-scale datasets.

Immigrant status
Focusing on CIVED:99 data from United States and Sweden, Barber et al. (2015) compared differences in civic knowledge and attitudes towards women's rights between native-born and immigrant students.They found that immigrant students scored lower on civic knowledge and on support for women's rights.Their analysis found few differences between immigrant and nativeborn students' perceptions of school contexts (such as classroom climates).Differences in civic knowledge could be explained in part by home languagespeaking English (in the United States) or Swedish (in Sweden).There were significant statistical interactions in each country suggesting different implications for engagement in curricular and co-curricular activities among immigrant and non-immigrant students.There were also indications of peer group influence on attitudes towards women's rights.In addition, Janmaat (2008) found in the CIVED:99 data that immigrant youth are more likely to show ethnic tolerance and solidarity in social action but less likely to support women's rights, to show national pride and to have institutional trust.Finally, Rutkowski et al. (2014) focused on immigrant students in 24 countries using ICCS:09 data.They found that an atmosphere of inclusion in school including positive student teacher relations, opportunities for student voice and encouragement of participation in schools/communities was associated with trust, positive attitudes towards the country and valuing conventional citizenship norms.
A number of studies explore how immigrant or ethnic minority students are distributed among classrooms and schools and how this relates to civic outcomes for all students.Janmaat (2014) tested predictions from contact theory, which posits that positive intergroup contact increases favourable attitudes of the in-group towards the out-group members.Two studies using ICCS:09 considered contact theory with respect to immigrant populations.Native-born students in classes with a high proportion of immigrant students had more positive attitudes regarding immigrant rights than nativeborn students in classes with proportionally fewer immigrants (Isac et al. 2012;Janmaat 2014).In a similar analysis of CIVED:99 data, Janmaat (2012), using data from England, Germany and Sweden; and Campbell (2007) using data in the United States, found that studying in ethnically diverse classrooms was associated with support for immigrants' rights.
Similar approaches have been used to understand other aspects of social diversity.With a slightly different focus, Collado et al. (2015) used ICCS:09 data to explore the role of socio-economic composition of the classroom, and concluded that attending schools that are economically segregated is associated with a wider gap in civic knowledge between high and low socio-economic students.Taken together the results of these studies demonstrate support for the proposition that diversity within the classroom across immigrant, race, or economic status has the potential to promote favourable achievement and attitudes on the part of members of the in-group and can reduce academic gaps.These findings relating to contact theory support Ekman and Zetterberg's (2011) assertion that the diversity of the school has as much influence on students' civic development as the official curriculum or instructional strategies.

Theme 4: Understanding students' citizenship norms and their coherence within individuals
One innovation in the mode of analysis employed with the CIVED:99 attitudinal data took place about a decade after the original report.Much of the analysis conducted up to that point was variable-centered analysis.This included comparison of group means and reports of correlations or multilevel regression models.In these analyses, a single dependent variable or score was considered at a time.For example, researchers would predict a summary score on a single dependent variable such as positive immigrant attitudes using several independent variables.Or, they would make comparisons between groups on a set of scales considered singly.The innovation was to conduct person-centered analysis, which considers how a number of attitudes go together or are configured within individuals (Chow and Kennedy 2014).Methods of person-centred analysis, such as latent class or cluster analysis, have been utilized particularly to understand students' citizenship norms and expected participation and to generate profiles employing several attitudinal variables.
In order to demonstrate the complexity of connections between attitudes, Torney-Purta (2009) and Torney-Purta and Barber (2011) reported a person-centered analysis of twelve attitude scales from the CIVED:99 dataset within five countries sharing the western European tradition.Profiles of attitudes were identified that characterized clusters or groups of individuals.These were the clusters.First was a group of social justice supporters who express high levels of support for minority rights, immigrants' rights and women's rights but without an intention to engage in any action.Second was a group of conventional citizens who have high levels of trust in governmental institutions and are active in traditional methods of political participation.Third was a group of indifferent individuals with attitudes near the mean across the scales.Fourth, there was a disaffected group similar to the indifferent cluster except with more negative views towards minorities' rights and lower trust towards the government and, finally, an alienated cluster of students characterized by negative attitudes across the attitudinal scales (with trust in government, support for immigrants' rights and support for ethnic rights well below the international mean).Substantial numbers of students fell into the disaffected and alienated clusters across the countries.This pattern led Torney-Purta and Amadeo (2013a) to suggest that the alienated and disaffected clusters that were observed among adolescents in 1999 might have been a harbinger of the increasingly negative reactions towards immigrants observed a few years later among adults in the United States and England.Toots and Idnurm (2012) conducted another cluster analysis on CIVED:99 data focusing on measures of nationalism in Estonia, Latvia and the Russian Federation, and voiced concern about the general lack of support for democratic values expressed by one cluster of respondents in Estonia.
Three additional studies used methods as cluster analysis to explore the contexts and correlates of students' expected political participation.Chow and Kennedy (2015) explored modes of political and civic participation comparing five Asian societies and found that those with a Confucian legacy, Taiwan, Hong Kong and South Korea, were substantially less likely to hold active notions of citizenship and were minimal participators in comparison to students in Indonesia and Thailand.Toots and Idnurm (2016) utilized cluster analysis to compare gaps in expected participation among high achievers and low achievers in civic knowledge in Europe.They found concentrations of high achievers in the good citizenship group and low achievers in the alienated citizens group.Reichert et. al. (2018) analyzed ICCS:09 student data from both the open classroom climate scale and the confidence in participation at school scale to develop profiles of school experience that were related to characteristics of students, schools and communities in the Nordic countries.There are still substantial opportunities in these datasets to utilize person-centred analysis (e.g. using school or teacher characteristics or organizational memberships, to name two).

iMPliCaTions and avenues foR fuTuRe ReseaRCh
This analysis of themes stemming from the broader review of literature employing CIVED:99 and ICCS:09 has demonstrated many valuable contributions made by researchers using data from these large-scale international surveys.The accomplishments have been substantial, but tremendous opportunities remain for future analysis.These opportunities have further expanded with the release of the data from ICCS:16 (Schulz et al. 2017).The following section discusses the implications of these findings, including the importance of open classroom climates and classroom contexts, the need for contextualized notions of citizenship and a call for more interdisciplinary approaches in analysis of large-scale international datasets.
First, the findings from analyses of CIVED:99 and ICCS:09 data consistently demonstrate the importance of the context of the classroom in promoting student learning and development.The value of an open and respectful classroom climate was a frequent theme across studies reviewed for this work, and similar to broader scholarship focused on improving civic education (Hess and Avery 2008;Ho et al. 2017;Levy 2011;Parker 2010).Researchers could enhance the impact of these findings with smaller-scale quantitative, qualitative, or mixed-methods studies as well as through additional secondary analysis.Such smaller-scale studies could investigate the process for fostering such climates in various school contexts or identify barriers that teachers are likely to encounter in improving classroom and school climates (and how to overcome them).In particular, future research could consider how professional development could address these issues.Turning to opportunities for further secondary analyses, these datasets also include items and scales measuring students' perceptions of their influence on decisions about schools and of the value of student participation at school.These measures have received less attention than classroom climate and warrant additional analysis.
Further, regardless of methodological approach used, future research could consider the nature of classroom composition.Increasing economic disparities are just one source of potential tension among students.One promising avenue for future exploration in these data is to identify classrooms and/or schools with particular climate profiles.Further studies could explore whether a school/classroom climate that combines certain characteristics is more or less effective in promoting civic knowledge and engagement among students with marginalized identities or in communities with particular characteristics.Analyses such as these could provide more powerful guidance than looking at classroom features one at a time.
In addition, educators could use the results of these and future studies to prioritize the enhancement of open classroom climates for respectful discussion of issues as an important policy goal.Policy statements from professional organizations could suggest action by local educational authorities.Policy-advisory groups could encourage the fostering of more open and respectful classroom climates within schools.Indeed, the association between an open classroom climate and a host of positive civic outcomes across contexts, countries and time, is a consistent and well-established finding with relevance to civic education.
Second, this review identified the need for additional research focusing on contextualized notions of citizenship.In the recent Handbook of Social Studies Research, Castro and Knowles (2017) reviewed research on democratic citizenship and found gaps in civic knowledge, generational shifts in attitudes, individualistic notions of democracy and the salience of identities such as those related to immigrant status, gender, ethnicity, race and class.These authors also acknowledged that their focus on research within the United States was a limitation.This could be addressed by analyses using these international datasets.Data from CIVED:99 and ICCS:09 also provide an international context for interpreting the implications of student identity called for by civic education scholars (Avery and Barton 2017;Hahn 2017;Ho et al. 2017;Janmaat 2014).We identified studies exploring differences by gender, as well as membership in groups marginalized by racial, ethnic, or immigrant status and SES.These IEA survey data present rich and untapped resources for understanding the implications of several dimensions of students' identity for civic development.To phrase the problem in a different way, future research should consider student identity as more than a control variable.For example, researchers could explore whether students from different groups experience a particular classroom climate differently, and how such differences relate to connections between class participation and future political engagement.This has the potential to address intersections of student identities based on gender, race and class.It would also be important to develop theoretical approaches to conceptualize the meaning of these differences.
We support Hahn's (2010) call for additional research to consider both students' and teachers' understandings of civic engagement in different cultural and national contexts.These datasets, along with the regional modules and national options, have the potential to provide important advances in this regard.For example, Kennedy (2004) noted that the western ideological paradigm has provided the foundations for conceptualizations of democratic citizenship education, and researchers should reconsider the suitability of basic concepts of democracy and citizenship cross-culturally.With this in mind, there has been considerable work exploring the uniqueness and diversity of citizenship in Asian countries (Cho and Kim 2013;Chow and Kennedy 2015;Kennedy 2012;Kennedy et al. 2012;Kennedy et al. 2013;Knowles 2015).Additional research could also discuss the unique context in civic education in post-communist Europe: a region under-represented in this review perhaps because of its focus on publications in English-language journals.Third, while a comprehensive review of theoretical approaches used in summarizing these data is beyond the scope of this article, the research presented spans several disciplines and uses specific theories that could be useful in the future.These include the communities of practice frameworks (Torney-Purta et al. 2010;Hoskins et al. 2012), contact theory (Collado et al. 2015;Janmaat 2014), the developmental niche model (Barber et al. 2015;Torney-Purta and Barber 2011) and social justice and critical consciousness (Diemer and Rapa 2016;Godfrey and Grayman 2014;Grayman and Godfrey 2013).In addition to the theoretical approaches, the diversity of the researchers using these datasets is likely to bring a depth of understanding of the results and implications of secondary analysis (Torney-Purta and Amadeo 2013b).Future scholarship could purposefully incorporate an interdisciplinary perspective.
An earlier annotated bibliography (Knowles and Di Stefano 2015) as well as this more detailed survey of secondary analyses indicates that a number of the attitudinal and participation scales in the IEA civic datasets have received relatively little attention.Researchers should explore the complex notions of citizenship (outside the two major categories of conventional and socialmovement related), institutional trust and civic participation that includes but extends beyond voting.Measures of student religiosity are largely unexplored (although they were collected in ICCS:09).In CIVED:99 students were asked to indicate how 'good for democracy' various political activities were; in ICCS:09 these items were framed as agreement to various statements reflective of democratic concepts.There appears to have been only a few articles analysing these items (e.g.Husfeldt andNikolova 2003 andTorney-Purta et al. 2006, both with CIVED:99).An insufficient amount research has addressed how students develop a sense of collective efficacy in solving problems in their schools.In addition, future studies could deepen the understandings of local as well as national and global factors shaping civic engagement practices and outcomes.
Moreover, civic knowledge represents a frequently examined outcome in the CIVED:99 and ICCS:09 studies, but it is also often considered as a control.Future research could more fully explore how particular knowledge bases and skill sets within civic education relate to each other.Zhang et al. (2012) and the qualitative analysis conducted by Arensmeier (2016), demonstrated that civic knowledge is multidimensional.Further analysis by Zhang et al. (2016) investigated the reading load of different types of civic-related questions.These studies represent key work in understanding how students develop and demonstrate knowledge related to civic and political life.
Clearly, there has been progress in understanding the connections between education and students' civic engagement based on both primary and secondary analysis of these international studies.Looking to the future, we conclude by calling for new and rigorous secondary analysis of existing datasets to inform educational research and practice.There are substantial opportunities within the existing data sets for this further analysis.A number of disciplines are recognizing the value of researchers working on shared data (Martone et al. 2018).More narrowly targeted reviews of research going into depth within a single discipline or crossing disciplines could be beneficial both within and across nations.In addition, analyses using a variety of other methodologies could promote greater understandings of findings from secondary analysis either in a stand-alone study (such as that conducted by Levy et al. 2011) or in a mixed method framework (as proposed by Stevick 2007).The release of data from the 2016 ICCS study has provided new opportunities for analysis, to give just one example for examining trends in attitudes across time (Munck et al. 2017).Indeed, the next several years will be an exciting time to be working in the field of international and comparative civic engagement, and it holds many opportunities for researchers.