A cross-national comparison of teachers' beliefs about the aims of civic education in 12 countries: A person-centered analysis

This article examines teachers' beliefs about the aims of citizenship education in 12 countries from Europe and Asia. A latent class analysis of the International Civic and Citizenship Education Study of 2009 identified three distinct profiles of teachers' beliefs about the goals of citizenship education. These profiles are associated with teachers' characteristics and with national indicators of democratic development. Profiles can be more useful than single beliefs in understanding how teaching contributes to students' civic development. Teachers across countries thought it far more important to foster students' participation in the school or local community than to foster future political participation. © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Democratic societies need citizens who are politically engaged and knowledgeable about their governments. Adolescence is a key period for preparing democratic citizens, and schools are expected to foster knowledge, skills and dispositions supporting future engagement and participation (Hahn, 1998). School-based civic and citizenship education (CCE) e either as a crossdisciplinary area or as a designated subject e has been recognized as an important factor in this process (Galston, 2001;Niemi & Junn, 1998).
In schools young people learn about how to contribute to society through formal and informal learning experiences (Parker, 2001;Reichert & Print, 2018). Teachers across subject areas play a role, with those who specialize in civics or in history and social science having particular responsibilities (Losito & Mintrop, 2001). However, few researchers have considered how teachers in both categories see their roles. What priorities do they place and what goals animate their teaching? To what extent is students' later participation a goal? What differences exist across countries in teachers' civic education aims? Thornton (2005) has characterized teachers as "curricularinstructional gatekeepers" implementing the curriculum according to their own epistemological and ideological beliefs about learning. In a review, Fives and Buehl (2012) distinguished between teachers' beliefs about self, context, content, specific teaching practices, teaching approach, and students. Park and Oliver (2008) characterized teachers' beliefs about the purposes and goals for teaching as an aspect of pedagogical content knowledge guiding instructional decisions. For example, Manzel, Hahn-Laudenberg, and Zischke (2017) argue that topics specified in the curriculum are more likely to be taught when they align with a teacher's beliefs about civic education. More generally, teachers' beliefs can act as filters in selecting topics and classroom activities (Fives & Buehl, 2012). Consequently, this study makes the assumption that what teachers believe about CCE goals matters to their teaching and the quality of students' learning (Martens & Gainous, 2013;Torney-Purta, Richardson, & Barber, 2005); teachers' beliefs directly and indirectly affect teaching-related decisions (Patterson, Doppen, & Misco, 2012;Thornberg, 2008).
Although the curriculum often defines goals for CCE, teachers must be selective; there is variation in how closely they adhere to either localized expectations or pronouncements from the state or national levels. Illustrating this, Kerr (2002) examined case study materials describing CCE in twenty-four countries (compiled by Torney-Purta, Schwille, & Amadeo, 1999) and concluded that teachers' values and beliefs have more influence in countries where the curriculum is less explicit. Teachers' decisions are also influenced by students' needs and school context (Thornton, 2005). Civic education considers the individual's role in a society with particular cultural values (Alexander, 2000). Unsurprisingly, these values influence the extent to which teachers aim to help students adopt a role in the existing social order or be prepared to change society (Thornton, 2005). Historical, political, economic and social contexts as well as the availability of educational resources are influential. An Israeli study found that teachers reported discussing global poverty differently in schools with high and low levels of socio-economic status and diversity (Goren & Yemini, 2017). When the school or community is experiencing tensions, education related to conflict resolution may assume enhanced importance.
These are important reasons for taking a comparative perspective to understand the beliefs of those who organize students' civic learning experiences. However, little research has described profiles of teachers' beliefs about CCE. The present analysis focuses on the content beliefs about the aims of CCE that teachers endorse, which constitute their subjective perceptions of its goals. Understanding cross-national differences in teachers' approaches can lead to insights about educational improvement. These insights may differ in regions with different histories of democracy. Therefore, the beliefs of teachers from four regions were examined (Asia, Eastern Europe, the Nordic area, and Western Europe), guided by four inter-related research questions: (1) What is the relative frequency among teachers of specific beliefs about the goals of CCE, particularly in relation to dimensions of knowledge, thinking or interaction skills and participation of various kinds? (2) Are there distinct groups of teachers characterized by distinct patterns or profiles in their beliefs about the goals of CCE? (3) How can these distinct groups be described with respect to characteristics of teachers? (4) How are these distinct groups distributed in countries from four regions, and what associations exist between distributions of group membership and indices of democratic and economic development? Our analyses provide a research basis for educational personnel and curriculum designers to identify effective approaches to civic and citizenship education in particular contexts.

Literature review
There is considerable agreement that teachers' beliefs influence classroom practices (Fives & Buehl, 2012;Patterson et al., 2012;Sim, 2011). However, relatively little is known about differences among teachers in beliefs about CCE. A well-supported finding from student data is that classroom climates that are open for discussion and allow respectful questioning of ideas are effective in promoting positive civic development (Geboers, Geijsel, Admiraal, & Dam, 2013;Knowles, Torney-Purta, & Barber, 2018;Reichert, Chen, & Torney-Purta, 2018). However, the overall cross-national evidence about teachers' beliefs or practices is sparse. Studies can be categorized by the type of data e qualitative, quantitative or mixed methods. First, we consider studies that derived typologies of teaching from qualitative methods using small samples. Then we consider studies with moderate-sized samples that used factor analysis of quantitative survey data to develop several scales representing different beliefs about teaching. The researchers then compared groups of teachers who were high and low on these scales. A few studies combined qualitative and quantitative approaches. Review of these variable-centered analyses and the "typologies" of teachers identified led to our research plan employing person-centered analysis of large-scale quantitative data from twelve countries to examine profiles or typologies of teachers in depth.

Qualitative and mixed methods studies
An example of qualitative research is a study of citizenship pedagogy with eight teachers in Singapore that identified three distinct approaches towards instruction linked to teachers' concepts of citizenship. One group conceptualized CCE as promoting good, moral persons; a second group believed that students should take social responsibility and become active in their communities; the third group of nationalistic teachers was concerned about transmitting knowledge and values to support the nation (Sim & Print, 2009). Sim, Chua, and Krishnasamy (2017) in another small study in Singapore identified character-driven teachers, socially participatory teachers, and critically-reflexive teachers. Kenyon (2017) used narrative inquiry with three US social studies teachers, concluding that personal experiences formed their ideas about authority, which in turn shaped three teaching approaches: for personally responsible citizenship, for a social justice orientation, or for reflective inquiry. An interview study with teachers of civicsrelated subjects in the UK and in Denmark about their citizenship concepts confirms that cultural and educational contexts need to be considered; the British teachers' emphasized knowledge and citizens' action while the Danish teachers stressed the experience of democratic decision-making and participation in a community (Hahn, 2015). Finally, a qualitative study in the Netherlands by Willemse, Dam, Geijsel, van Wessum, and Volman (2015) concluded that teachers are often unaware of how their beliefs influence their teaching practice, but involvement in curriculum development can increase this awareness.
Using a sorting technique together with interview and survey data from several US samples, Anderson, Avery, Pederson, Smith, and Sullivan (1997) explored social studies teachers' perspectives. Despite common beliefs about the goals of citizenship education, such as encouraging tolerance and addressing controversial issues, these researchers also found differences. Social studies teachers could be classified as cultural pluralists, communitarians, legalists, critical thinkers, and assimilationists, though not all types were found in all samples. Importantly, the authors identified significant correlates of group membership. A focus on cultural pluralism was more common among teachers working in schools with students from several cultural backgrounds. Patterson et al. (2012) in a mixed methods study using surveys and interviews from social studies teachers in the US state of Ohio found that all teachers emphasized knowledge as an important CCE component, but few saw it as enhancing active participation. Most teachers focused on a traditional understanding of citizenship (e.g. being knowledgeable, obeying the law, voting); about one quarter emphasized civic understanding and community participation; less than four percent of the teachers believed that citizens need a critical understanding of the world in order to promote social justice. Lee (2005) conducted a non-representative teacher survey in five countries that identified commonalities and differences among civics teachers; these findings were substantiated by semistructured interviews. The welfare of others and ethical behavior were important concerns among the majority of teachers surveyed. Knowledge transfer was especially important in the Chinese and Russian samples, while Western teachers emphasized tolerance. Rogers and Westheimer (2017) also employed mixed methods showing that liberal US teachers more frequently than conservative teachers addressed economic inequality in their classes usually because of personal concern about this issue. Further, civically and politically engaged teachers were more likely than other teachers to encourage students to consider the complexity of problems such as economic inequality. Hess and McAvoy (2015), in their mixed methods study in three states in the US, also concluded that the approach taken and climate of the classroom varies with differing political views of teachers.

Quantitative studies
Knowles (2017) factor analyzed survey data on the beliefs of teachers in the US state of Missouri demonstrating links between beliefs and teaching practice. He identified a conservative, teachercentered approach emphasizing transmission of civic knowledge and the moral standards of society with students assumed to be passive recipients; a liberal approach in which students are active participants and constructors of meaning; and a critical, inquirybased approach focusing on skills and aiming to help students problematize and deconstruct an issue (often countering injustice). In another study, Thornberg and O guz (2016) compared scale scores from a survey about moral and citizenship education administered to about 200 student teachers in Sweden and Turkey. They identified two groups: one traditional-conservative and the other critical-progressive emphasizing social justice. These studies present an interesting but fragmentary picture.
A large-scale quantitative study is the international 1999 Civic Education Study (CivEd) of the International Association for the Evaluation of Educational Achievement (IEA) (Torney-Purta, Lehmann, Oswald, & Schulz, 2001). The researchers analyzed group means (a variable-centered analysis) and found that teachers of civic education across countries emphasized knowledge transmission, although they also believed that CCE ought to put more emphasis on critical thinking and values (Losito & Mintrop, 2001). The variable-centered analyses from the IEA International Civic and Citizenship Education Study (ICCS) in 2009 and 2016 also suggest that teachers across subjects and school principals prioritize student knowledge and skills (Schulz, Ainley, Fraillon, Kerr, & Losito, 2010;Schulz et al., 2017).
More extensive cross-national analyses of CivEd data showed differences among teachers of civics-related subjects in perceptions of their roles (Torney-Purta et al., 2005). This variable-centered analysis found that teachers' educational experiences were associated with their confidence in teaching specific civic topics. Some differences between countries were identified. For example, Nordic teachers were likely to negotiate with students about what was to be learned, while teachers from the Czech Republic and the US were more restrictive. Torney-Purta et al. (2005) concluded that teacher development should account for the cultural and educational contexts within a society, a claim similar to that from a comparison of Western democracies using extensive mixed methods data that served as a basis for the design of the CivEd study (Hahn, 1998).
In an analysis of teachers' confidence in teaching civic topics in Germany, Hong Kong, Italy, and the US, Alviar-Martin, Randall, Usher, and Engelhard (2008) also used CivEd data. Patterns of confidence in different content domains varied across national contexts. The researchers concluded that the status of civic education and the availability of guidelines and teacher preparation are associated with teachers' confidence. When civics is a distinct school subject and when policies or curricula provide guidance, teachers feel more confident in teaching this subject matter.
Finally, a variable-centered analysis of the CivEd data from civics-related teachers in the US was conducted by Gainous and Martens (2016). Students who perceived their classrooms as more open for discussion were likely to be taught by teachers with more liberal (rather than conservative) political convictions. Liberal teachers were defined as those who emphasize environmental and human rights issues. Conservative teachers were those who emphasize obeying the law and loyalty to the country, as well as respecting government representatives and serving in the military. Both an open classroom climate and being taught by liberal civics teachers were positively associated with students' political knowledge.
A smaller teacher survey in the US concluded that "Democrats are more likely than Republicans to value teaching students to be tolerant, to be global citizens, and to be activists who challenge the status quo. On the other side, Republicans are more likely to value teaching students to be respectful of authority and to know facts and dates" (Farkas & Duffett, 2010, p. 31). Complementary evidence comes from the study by Knowles (2017) previously discussed. Using a structural equation model, he concluded that teacher-and text-centered modes of instruction (e.g. lectures, completing worksheets) are more commonly used by more conservative teachers. Furthermore, a more critical ideology was associated with using participatory modes of instruction (e.g. debates, roleplaying).

Summary
Despite different terminologies, there is considerable overlap suggesting three major categories. Several studies identified a traditional, somewhat conservative type of teacher emphasizing knowledge transfer, support for the nation, and obedience to the law. Another type emphasizes the needs of the community and social participation. These teachers sometimes also highlight social justice, a concern that is primary in a third type of teachers who focus on critical thinking and inquiry. Teachers whose primary concern is students' individual growth and character development were identified in studies in Singapore (Sim et al., 2017;Sim & Print, 2009).
Most published research, with the exception of the variablecentered analyses using the CivEd data from 1999, has relied on small geographically concentrated groups of respondents. The cross-national evidence remains scarce, despite the fact that the beliefs about CCE's objectives have been shown to vary among countries with different democratic traditions (Lee, 2005;Schulz et al., 2010Schulz et al., , 2017Torney-Purta et al., 2005;Zyngier & Carr, 2012). Furthermore, some studies suggest that beliefs about the purposes of CCE are coherent. What is missing is a quantitative approach analyzing the existing large-scale quantitative comparative datasets using cluster analysis or latent class analysis. This could expand upon the qualitative studies describing teachers with contrasting conceptualizations of CCE goals. These qualitative small-scale studies have examined distinctive teacher profiles in approaches to CCE, but they cannot draw general conclusions about how common or rare different types of teachers are. Personcentered quantitative analysis has not yet been employed to identify profiles of teachers' beliefs using data from representative samples across countries.
Person-centered analysis of large-scale datasets also has methodological advantages and is especially suitable when variance in the constructs may not be equally distributed among countries (Eid, Langeheine, & Diener, 2003). A person-centered approach, such as latent class analysis, has the goal of classifying individuals into "distinct groups or categories based on individual response patterns so that individuals within a group are more similar than individuals between groups" (Jung & Wickrama, 2008, p. 303; emphasis added). These analyses "can take on a comparative perspective within a sample to explore both commonality and difference in persons' various characteristics simultaneously" (Chow & Kennedy, 2014, p. 473). Person-centered analysis based on large and representative samples is ideal for examining group or individual differences as it enables researchers to describe patterns of similarities and differences among individuals. It can also examine the characteristics of individuals of different types more comprehensively than qualitative studies (Reichert, 2016a).
Moreover, the findings of person-centered research using largescale assessments tend to be easier to grasp for policy-makers, educators and the public than the results of variable-centered analyses of these data (Torney-Purta & Barber, 2011); this means more potential contribution to educational improvement. These findings can be supplemented by follow-up analyses providing deeper descriptions of the correlates of membership in each response profile (Reichert, 2016b(Reichert, , 2017. The present study addresses the lack of cross-cultural research on teachers' beliefs about the goals of CCE using large-scale data. It moves beyond tables of item frequencies or means to person-centered quantitative analyses of profiles of teachers' beliefs.

Present study
This study was conducted to understand teachers' beliefs about the aims of CCE in societies with differing democratic histories. It identifies differences and similarities among teachers by classifying them into groups with specific and distinct profiles in their endorsements of the aims of CCE. The analysis used an innovative research method and was guided by four research questions.
First, what is the relative frequency among teachers of specific beliefs about the goals of CCE, particularly knowledge, thinking, interaction skills and participation of various types?
Second, is it possible to identify distinct groups of teachers characterized by different patterns (i.e. "teacher profiles") in their beliefs about the goals of CCE? Prior research suggests that many teachers emphasize knowledge acquisition, but there is variation in teachers' beliefs (Patterson et al., 2012). Such heterogeneity is especially likely in cross-national research such as this study (Eid et al., 2003). We expected to identify distinctive teacher profiles of beliefs about the goals of CCE; at least one group of teachers should prioritize the acquisition of civics-related knowledge. Previous small-scale research suggesting a category of teachers stressing community participation and another group of teachers highlighting critical thinking led us to expect other patterns of emphasis as well.
Third, how do these distinct groups of teachers differ from each other with respect to other characteristics? Are there differences between teachers of CCE-related subjects (who may have more subject-matter knowledge) and other teachers, for example? Other characteristics of teachers, such as experience and gender, might also be related to profile group membership.
Fourth, how are the teacher profiles distributed among countries with different levels of economic development and distinct democratic traditions? National goals of education and curricula vary among countries, and it is reasonable to expect more homogeneity of teacher profiles within countries than between countries (Hahn, 2015). Furthermore, we expect that distributions of teacher profiles will be more similar among countries within the same region than among countries from different regions, owing to commonalities in democratic development, cultural values and educational systems.
Finding answers to these questions is vital in the context of the pressing societal issues that civic education is expected to address. These include globalization, the environment, immigration, and citizens' participation. Teachers' beliefs about what is important to learn are influenced by their own experiences and knowledge, the curriculum, students' needs, and values. Comparative research can help to understand these patterns of belief and to assist schools, teachers and curriculum designers in improving education for democratic citizenship.

Data and samples
The International Association for the Evaluation of Educational Achievement (IEA) International Civic and Citizenship Education Study (ICCS) database was utilized to study teachers' beliefs about the goals of CCE (see Brese, Jung, Mirazchiyski, Schulz, & Zuehlke, 2014;IEA, 2011;Schulz, Ainley, & Fraillon, 2011;Schulz et al., 2010). 1 Conducted in 2009, ICCS is a large-scale assessment of fourteen-year-olds' civic knowledge and understanding, dispositions and attitudes in 38 countries. ICCS used a stratified two-stage probability sample design, where schools were sampled with probability proportional to size during the first stage. One intact class of target-grade students and a fixed number of target-grade teachers of all regular school subjects were randomly selected during the second stage (Zuehlke, 2011). According to a review of published secondary analyses by Knowles et al. (2018), the data collected from ICCS teacher surveys have rarely been utilized by researchers.
For the present analysis, the samples from three societies from each of four regions were analyzed: Taiwan (Chinese Taipei), Hong Kong and the Republic of Korea for Asia; the Czech Republic, Poland and Slovenia for Eastern Europe (post-Communist countries); Denmark, Finland and Sweden for the Nordic region; and England, Ireland and Italy for the Western part of Europe. The survey question about aims was not administered in Norway. Since this left only three countries in that region, three countries were selected from each of the remaining regions.
This set of countries enables us to compare those that share many similar cultural and political values within regions but vary between regions in cultural norms and democratic traditions. For example, Nordic countries are relatively homogeneous, are successful welfare states and have long-established stable democracies. Western European societies are also long-standing democracies but have less similar historical developments. Eastern European (post-Communist) countries became democracies more recently. These countries are new to the European Union and face challenges due to political, economic and societal transitions. It has been suggested that CCE in Asian societies is influenced by Confucian principles emphasizing hierarchy and collectivism (e.g. Knowles, 2015); Ho (2017) argues that seeking harmony through negotiation and mediation rather than direct confrontation is valued. Previous studies have also suggested that CCE in Asian societies incorporates the cultivation of moral virtues (Lee, 2010).
Hard copies of teacher questionnaires were sent in 2009 to each school for distribution to sampled teachers, except for teachers in Slovenia and Sweden where teachers received a cover letter with instructions on how to complete the questionnaire online (Brese & Jung, 2011). In total, 24 127 teachers from the 12 selected countries participated in ICCS (ranging from 928 teachers in Denmark to 3 023 teachers in Italy), and 7 378 of these teachers indicated that they taught civic education, social sciences or humanities (ranging from 356 teachers in Poland to 807 teachers in Slovenia). 2 However, 4.16% of all teachers did not respond about the aims they believe most important for civic education. These teachers were excluded, leaving 23 123 teachers in the total sample, of whom 7 145 were civics-related teachers. Use of "senate weighting" ensured that none of the countries would be dominant in the analyses.

Aims of civic and citizenship education
All teacher respondents were asked: "What do you consider the most important aims of civic and citizenship education at school?" Ten choices were offered; each teacher was asked to indicate the three aims they considered most important. The choices and overall frequencies are listed in Table 1. These frequencies are interesting in themselves.
"Promoting students' critical and independent thinking" was most often endorsed (overall 60%, 89% in the highest country and 19% in the lowest country). Another aim with high variation was "developing students' skills and competences in conflict resolution," (overall 37% with 63% in the highest country and 15% in the lowest). "Knowledge of citizens' rights and responsibilities" garnered support from 55% overall while "knowledge of social and political institutions" garnered support from 34% overall. The least endorsed aim was "preparing students for future political participation," endorsed overall by only about 6% (with the highest country being 16%).

Predictors of teacher profiles
Several covariates were introduced as potential predictors of the teachers' 3 selection of particular aims of CCE. Teachers' age (median age: 45 years) provides an indicator of experience, 4 which may be associated with beliefs and teaching approaches (for example, Knowles (2017) found that conservatism was positively associated with years of teaching). Teachers' gender (male/female; 32% male teachers) was utilized, because men and women tend to engage in different types of political activities (Coff e & Bolzendahl, 2010) and they have been found to be knowledgeable in different areas of the political domain (Stolle & Gidengil, 2010). Male teachers appear to support more conservative ideologies of CCE than female teachers (Knowles, 2017;Thornberg & O guz, 2016). Another important teacher-level characteristic was school subject, that is, whether a teacher taught a civics-related subject (no/yes; 31% civics-related teachers). Teachers of civics-related subjects might feel obligated to cover specific civic topics or be more likely to encourage participation than teachers of other subjects. In addition, school sector (public vs. private school; 80% public school teachers) was included. A country indicator was used as a predictor of latent class membership.

Indicators of democratic quality and economic development
One aim was examining teacher profiles in relation to democratic and economic country contexts. Four country-level indicators were used ( Table 2). The Gross Domestic Product (GDP) and a country's public expenditure on education as the percentage of the GDP were indicators of economic development (the higher these scores, the higher the national income or/and educational expenditure). Indicators of democratic quality included the Economist Intelligence Unit (EIU)'s (2008) Index of Democracy that captures the quality of the electoral process and pluralism, the functioning of government, political participation, political culture, and civil liberties. Also used was Transparency International's (2009) (Lack of) Corruption Perception Index (CPI). The higher these scores, the Table 1 Relative frequencies of choosing each of the ten aims of civic education among teachers from twelve countries (ordered from most frequently chosen to least frequently chosen objectives). higher the democratic development/the less corruption.

Data analysis
The research questions were approached in steps. First, several joint choice latent class analyses (LCA) were estimated using the Latent Gold 5.1 Syntax Module (Vermunt & Magidson, 2016). Latent class analysis is a probabilistic method to identify groups of teachers characterized by distinct response profiles. Teachers with similar responses group together while teachers with different responses are found in different profile groups (or "latent classes"). Joint choice LCA can model data that do not satisfy the criterion of independent choices. This method was chosen because teachers selected three out of ten possible choices (i.e. the selection of one objective was not independent of the probability of choice for remaining objectives). The clustered nature of the data and the stratification of schools due to the sampling strategy were accounted for using the complex modelling options in Latent Gold.
Relative fit measures, such as the Akaike information criterion (AIC) and the Bayesian information criterion (BIC), were used to decide how many profile groups (latent classes) are needed to describe heterogeneity of the responses in the data. Several relative fit measures were triangulated to decide upon the number of groups most suitable to describe the response profiles of teachers' beliefs about the most important aims of CCE. Furthermore, covariates associated with group membership can improve the accuracy of the assignment of teachers to profile groups. Predictors were added subsequently to improve the classification in a step-by-step procedure. Finally, bivariate correlations at the country-level examined associations between the indicators of democratic tradition or economic development and the distributions of teacher profiles in specific countries.

Results
Elaborating on the presentation in Table 1 here we summarize how many teachers on average prioritized (as among the top three) each of the ten objectives of CCE (Table 3). The three most frequent choices in every country were among the top five choices in the pooled sample, though variation existed. "Promoting students' critical and independent thinking" was among the top three objectives in all countries except Korea. It is important to note that "preparing students for future political participation" was among the three most infrequently chosen aims of civic education in all countries except Denmark. The low percentages of "preparing students for future political participation" are very similar to what was found in CivEd 1999, where teachers reported relatively little emphasis on actual participation. However, in that study many teachers prioritized participation over knowledge when asked what the emphasis of civic education should be (Losito & Mintrop, 2001). In the recently released ICCS 2016 preparing students for participation again ranked low (Schulz et al., 2017). The following section discusses how distinct teacher profiles were identified across countries and regions.

Number of distinct groups
First, LCA were performed separately for each country (no table). The model fit indices indicated that in most countries, at least one and no more than three latent classes would be suitable to describe the teachers. That is, the national teacher samples were relatively homogenous with respect to their beliefs about the most important aims of CCE. However, this also means that some groups may be relatively small in some countries.
LCA of the pooled data from all twelve countries were performed subsequently and several fit measures were triangulated (Fig. 1), keeping in mind interpretability of the solutions. Whereas BIC-related measures are more conservative in model selection, AIC-related measures often favor models that are too complex (Dziak, Coffman, Lanza, & Li, 2012;Tein, Coxe, & Cham, 2013). Fig. 1 shows that in the pooled sample some fit measures increase with more than three latent classes and others level off (Collins & Lanza, 2010). Therefore, the three-class model was selected for further examination. Most of the previous mixed methods studies also identified two or three groups.
Covariates were added step-by-step to ascertain whether these variables could reduce classification errors and improve the accuracy of assignments of teachers to those three groups. The results in Table 4 suggest that a model with the country indicator and the subject taught by the teachers as predictors of group membership generally performed better than latent class models without covariates or models with additional covariates. All regression parameters of the chosen model were fixed (cf. Bray, Lanza, & Tan, 2015), and then these additional covariates were entered in the equation to predict group membership. 5 5 Due to space limitations and because most of the covariates did not significantly reduce the classification reliability, in this paper we only describe the profiles in relation to the mean levels of these variables and with respect to country differences. The in-depth analysis of the remaining covariates is reserved for future reports.

Teacher profiles
The identified groups of teachers were labelled "teaching for dutiful school participation and consensus building," "teaching for knowledge and community participation" and "teaching for independent thinking and tolerance." Fig. 2 shows the teacher profiles as the probabilities of choosing a CCE aim conditional on groupmembership. The probabilities sum to one for each group and need to be interpreted in view of the overall relative frequencies of choice for each objective. These descriptions can be contextualized by referring back to Table 1 and Fig. 2. Very few teachers chose students' "future political participation" as an objective for teaching, and the profiles showed uniformly low choice. So that objective could not be used to distinguish between the profiles. Likewise, although "respect the environment" was chosen relatively frequently there was little distinction between the profiles in choosing this aim. For all the remaining teaching objectives there was at least one profile where the objective was relatively high or low.

Teaching for dutiful school participation and consensus building
The first group of teachers stands out because they are likely to believe that CCE should primarily support the development of skills in conflict resolution and they promote students' participation in school life. Teachers in this group were unlikely to believe that CCE should primarily promote critical and independent thinking, or the Table 3 Ten aims of civic education (country percentages) (sorted from most frequent to least frequent aim in the pooled sample).

Asia
East Europe Nordic West Europe

Teaching for knowledge and community participation
The second profile group of teachers put a relatively strong emphasis on knowledge acquisition (about both rights/responsibilities and about institutions), and these teachers were also very likely to believe that CCE should promote students' participation in the local community. They were moderately likely to believe in promoting independent thinking. On the other hand, these teachers were relatively unlikely to believe that skills in conflict resolution or in defending one's own point of view were among the three most important aims of CCE. This group combines elements found in two types of teachers identified in previous studies through its emphasis on knowledge and community participation.
Forty percent of the teachers of civics-related subjects were in this group, compared to 30% of all other teachers (differing from the previous profile). Male teachers (37%) were somewhat more likely than female teachers (31%) to be found in this group. The same applies to teachers at private (46%) versus public schools (30%). Also, the youngest teachers were likely to be in this group (60%), while the oldest teachers were comparatively unlikely to be in this group (28%). Some older teachers may anticipate facing challenges if they were expected to organize students' participation outside the school.

Teaching for independent thinking and tolerance
Teachers in the third group were highly likely to select the aim to promote critical and independent thinking, an objective identified as typical for some teachers in previous research. The teachers in this group also thought it important to develop skills in conflict resolution as well as students' capacity to defend their own points of view. Furthermore, they were somewhat more likely than those in other profile groups to support the development of effective strategies for the fight against racism and xenophobia (although this was a small percentage except in Sweden; see Table 3). In fact, discouraging racism was not a particularly important aim in any of the profile groups. Teachers in this profile were relatively unlikely to choose participation at school or participation in the community as primary aims of CCE. They also infrequently chose the knowledge aims, relative to the other groups. All profile groups had low frequencies of promoting political participation, but this group was the lowest.
Approximately equal numbers of teachers of civics-related subjects (49%) and of other subjects (45%) were in this group. Likewise, there were minimal gender differences (48% of the female teachers and 44% of the male teachers). Half the teachers at a public school (48%) believed that the main objective of CCE is to teach for independent thinking and tolerance, whereas it was only about a third of teachers at private schools (37%). This is difficult to interpret because the nature of private schools differs across countries (e.g. independent schools predominating in some and religious schools in others). The oldest group of teachers was especially likely to be in this group (64% compared to less than 44% of younger teachers).

Country-level covariates of teacher profiles
Particularly interesting is the distribution of teachers across the three profile groups in each country. In most societies one type of teacher predominated (Table 5 and Fig. 3). The Nordic countries (Denmark, Finland, Sweden) as well as Slovenia had very large proportions of teachers who believed that CCE should primarily develop independent thinking and tolerance. Unlike the Nordic teachers, a substantial number of Slovenian teachers were also in the group that prioritized dutiful participation at school. It is worth noting that the educational and political trajectory of Slovenia differs in many respects from Poland and the Czech Republic, whose profiles are considered below (see also Krzywosz-Rynkiewicz, Zalewska, & Kennedy, 2018). Most teachers in Hong Kong, England, Ireland, and Italy believed that CCE should primarily facilitate knowledge acquisition as well as students' participation in their communities. In England and to some extent in Italy a considerable number of teachers also believed that independent thinking and tolerance should be primary aims of CCE.
Some countries' results are noteworthy. More than 85% of the teachers in the three Nordic countries chose independent thinking and tolerance. Substantial groups of teachers from the Czech Republic were found in all three profile groups. However, those who believed in the objective of independent thinking and tolerance were the most common. Polish teachers were concentrated in the two groups other than independent thinking and tolerance. In fact, Table 5 Distribution of teachers across groups per country (posteriori probabilities in percent).

Country
Dutiful school participation and consensus building Knowledge and community participation Independent thinking and tolerance  in these two post-Communist countries no profile held more than 45% of the teachers, perhaps because of changes in curriculum that have been incorporated more fully by some educators than by others. The Asian societies were quite dissimilar to each other: Almost all Korean teachers were in the dutiful participation at school group; teachers in Hong Kong prioritized the acquisition of knowledge and community participation; in Taiwan a majority of teachers believed CCE should promote independent thinking, but there was also a large number of Taiwanese teachers who prioritized CCE for dutiful school participation and consensus building. This group was less than one percent in Hong Kong.
The bivariate associations between the percentages of teachers per national group and four indicators of economic and democratic development are shown in Table 6. The higher the per capita income, the public expenditure on education, or the democratic development, and the less corruption prevailing in a country, the higher was the percentage of teachers for whom independent thinking and tolerance were the primary aims of CCE. In contrast, the lower the per capita income and the more corruption was perceived to be prevalent, the greater was the likelihood of teachers being primarily concerned about teaching to promote dutiful participation at school and consensus building. The public expenditure on education and the democratic quality of a country were relatively low in countries where substantial numbers of teachers prioritized knowledge acquisition and community participation.

Discussion
This research set out to identify distinct groups of teachers based on their beliefs about the primary objectives of CCE. By using large-scale representative teacher data from countries with differing democratic traditions and through a person-centered statistical approach, we systematically examined teachers' beliefs about CCE cross-nationally in a way that augmented previous small-scale qualitative and quantitative studies. Studying these beliefs is important because they function as filters and guide instructional decisions (Fives & Buehl, 2012). Following a brief summary of results, we consider how societal contexts might shape teacher beliefs by discussing each profile group in relation to relevant contextual factors. After a brief discussion of a few general observations, we then address limitations of the present study, followed by concluding remarks on the implications of the findings.

Identification of teacher profiles
Three groups of teachers with distinct profiles of beliefs about the aims of CCE were identified; these were "teaching for dutiful school participation and consensus building," "teaching for knowledge and community participation," and "teaching for independent thinking and tolerance." It is important to note that participation at school was the focus of the first group, while the second group focused on participation in the community. Importantly, all three profile groups supported teaching respect for the environment to about an equal extent. However, in none of the three groups did a substantial number of teachers believe that encouraging students' political participation was among the top three aims of CCE. Relating this to previous research, we found a group emphasizing independent thinking as did most of the studies reviewed (e.g. Anderson et al., 1997;Patterson et al., 2012;Sim et al., 2017). However, there was also a group combining knowledge transmission and community participation, which earlier research often found in separate groups. Furthermore, relatively few teachers in these countries see the role of CCE in enhancing social justice (e.g. Patterson et al., 2012). In the current analysis promoting strategies against racism or xenophobia was rarely chosen, and community participation did not appear to have an activist focus. The first group showed little emphasis on independent thinking and moderate support for knowledge acquisition, which may reflect a more traditional approach to citizenship teaching identified by others (e.g. Knowles, 2017;Patterson et al., 2012;Sim et al., 2017). An encouraging note is that teaching about respect for the environment was an aim of substantial numbers of teachers in all countries (less so in Denmark) and about equally in all three profiles.
Our results provide vivid descriptions of the teachers and make it easier for educators, policy-makers and stakeholders to grasp the nature of teachers' beliefs. The prevalence of these three teacher profiles varies significantly when teachers of civics-related subjects and other teachers are compared, as well as among societies with different democratic traditions. The patterns of teachers' beliefs about the primary goals of CCE differed significantly among the three Asian countries, thus suggesting limited homogeneity within the region. Countries in the other three regions showed considerable similarities (except for some divergence between Slovenia and the other two post-Communist countries).

Teacher profiles, societal context and teacher development
Following the characterization of teachers as curricularinstructional gatekeepers whose aims are influenced by their epistemology, students' needs, curricular goals and contexts (Thornton, 2005), we identified different groups of teachers varying in their beliefs about the aims of CCE. Knowledge and thinking skills were dominant among the three (out of ten) most important goals of CCE according to the teachers in this study. Encouraging political participation and diminishing racism were the least endorsed aims for CCE across countries. Other studies in the US and internationally also suggest that teachers and school principals prioritize students' knowledge and skills rather than future political participation (Losito & Mintrop, 2001;Patterson et al., 2012;Schulz et al., 2010Schulz et al., , 2017. However, there was considerable heterogeneity among teachers. The results also showed that teachers of civics-related subjects, who should be most familiar with the civics-related curricular goals and who may have specialized training, differed from other teachers in their distribution across profiles. Furthermore, we expected that each country would have one dominant group reflecting within-country similarities in Table 6 Correlations between group membership (country percentages) and country-level indicators.

Indicator
Dutiful school participation and consensus building Knowledge and community participation Independent thinking and tolerance Gross Domestic Product À0.70** 0.07 0.52* Expenditure on education À0.37 À0.50* 0.81*** Democracy Unit Index À0.23 À0.49 0.69** Lack of Corruption Perception Index À0.57* À0.08 0.55* Note: The lack of corruption perception index is scaled in a way such that high numbers mean less corruption.*p < .10; **p < .05; ***p < .01. Data sourced from ICCS 2009 and Table 2. recommended curricular goals, among other factors. We anticipated considerable heterogeneity in the distribution of these profiles between countries, particularly among countries from different regions. In fact, although there were several countries dominated by one of the three profile groups, in other countries teachers were distributed across more than one group (especially in the post-Communist countries where reforms in teacher training may have been unevenly implemented). Within regions there was also considerable homogeneity reflecting common social values and democratic developments; Asian societies were less similar to each other than countries in other regional groups. Research on Asian students has shown diversity across the region similar to these findings (e.g. Chow & Kennedy, 2014). We discuss possible explanations below.

Dutiful school participation and consensus building
On the whole teachers showing this profile were quite unlikely to choose critical and independent thinking among the three most important goals of CCE. They were relatively likely to think that CCE should support the development of skills in conflict resolution and promote participation in the school setting. Teachers of civicsrelated subjects were significantly less likely than teachers of other school subjects to show this profile, which may reflect their training or perceptions of students' or schools' needs.
Almost all teachers in Korea were in this uncritical and schoolfocused group, but the post-Communist countries and Taiwan also were frequently included. Some of these countries have relatively recently become democracies, and independent thinking may not be the priority of teachers trained in earlier periods. Skills in conflict resolution are important in CCE in some places. For example, in the Republic of Korea conflict with the Democratic People's Republic of Korea (North Korea), has been omnipresent, and more than half of the teachers in (South) Korea said skills in conflict resolution are important. This profile of teacher beliefs was relatively rare in the Nordic and Western European countries.
In addition, the Confucian notion of governance emphasizes hierarchy and conformism (Knowles, 2015). Ho (2017) offers an alternative interpretation referring to Confucian ideas of harmony, emphasizing communal trust and social cohesion. Other scholars indicate that the influence of Confucianism in society is strongest in Korea and weakest in Hong Kong (Geun, 2014;Lin & Ho, 2009). These are possible explanations for the fact that teachers in Korea appear quite homogeneous in their beliefs about the priorities of CCE.

Knowledge and community participation
A third of all teachers prioritized the acquisition of civic and political knowledge as well as community participation as goals of CCE. These teachers predominated in the Western European countries as well as in Hong Kong. Significant numbers of teachers in two of the post-Communist countries (Czech Republic and Poland) were also in this group. Hong Kong was under British rule for 150 years though citizenship education was not promoted. Moral education was a focus, but recent resentment about the imposition of national education by the Chinese government might indicate a recognition that democratic citizens need knowledge of democratic institutions and a willingness to engage in volunteering and community service. 6 Future research could clarify how democratic reforms as well as cultural and religious influences might be causally associated with teachers' emphasis on knowledge acquisition and community participation.
Although both aspects of knowledge (of rights/responsibilities and of institutions) appeared in this profile, the locus of participation (in the school and in the community) appeared as defining characteristics in different profiles. The promotion of school participation appears in the previous profile along with uncritical approaches, while community participation appears in this profile along with the two aspects of knowledge.

Independent thinking and tolerance
Finally, almost half of the teachers prioritized independent thinking and to a lesser extent tolerance as among the three most important aims of CCE. In the three Nordic countries teachers with these beliefs were by far the most common. Note that Torney-Purta et al. (2005) using CivEd data also found that Nordic teachers were especially positive about negotiating with students concerning teaching objectives and processes. Teachers in England, Taiwan and the Czech Republic were also found in this profile. In fact, this third group aligns in some respects with the critical epistemological approach in civic and moral education pursued by teachers who aim to promote independent thinking and social justice according to Knowles (2017) and Thornberg and O guz (2016). However, teachers in this profile were not especially likely to prioritize future political participation to accompany this independent thinking.
The other two profile groups show signs of a conservative approach (e.g. prioritizing knowledge acquisition over independent thinking). Future studies need to examine how profiles of teachers' beliefs about the goals of CCE align with their teaching practices. Small-scale research suggests some possibilities. Teachers aiming at encouraging dutiful citizens or knowledge transmission may adopt a teacher-centered approach, while those concerned about social injustice may utilize an interactive or constructive approach (e.g. Knowles, 2017;Sim, 2011). Yet, comparative research is scarce. Mechanisms explaining potential differences among teachers within the same group and how teachers adjust their teaching approaches conditional on their goals with specific classes deserve exploration. Linking these profiles of aims to teachers' concepts of good citizenship could clarify whether the emphasis on critical thinking shown by this group is associated with a particular understanding of what it means to be a citizen in a democracy. Patterson et al. (2012) in their US study suggested discrepancies between teachers' definitions of citizenship and teachers' beliefs about the general purpose of the social studies.
Furthermore, the goals prioritized by members of this third profile group who emphasize independent thinking and tolerance were more likely to be found in democracies where there was an absence of corruption, as well as relatively high national income and public expenditure on education. Although it is not possible to disentangle cause and consequence, Nordic education policies do emphasize the value of democratic participation; education is also less centralized in some of these countries (Blossing, Imsen, & Moos, 2014). In addition, economic resources and relatively trustworthy governments enable teachers and citizens in general to pursue "post-materialist" values (Dalton & Welzel, 2014) or to be "stand-by citizens," who have low levels of habitual participation but are ready to be mobilized when important issues arise (Amnå & Ekman, 2014).
Finally, Taiwanese society e although influenced by Confucianism e is more culturally diverse than other Asian societies. Taiwan has also become more pluralist over the previous decades, and its education system has been decentralized (Ainley, Schulz, & Friedman, 2013). This process has meant that curriculum and teaching are less likely to be permeated by Confucian ideas in the civic and moral education curriculum (Hung, 2015). The split between teachers who emphasize dutiful CCE versus independent thinking might reflect this development.

Further observations
The moderate to high emphasis on protection of the environment among teachers in all twelve countries is encouraging. However, the low emphasis placed on combatting racism and xenophobia will be discouraging to many, especially given the influx of immigrants in the Western European and Nordic countries. Finally, teachers are much more comfortable encouraging participation at school or in the community than they are with encouraging future political participation. Remember that the teachers sampled were from a range of subject matters (some more appropriate for explicit civics-related instruction). Not only were teachers of civics-related subjects more likely members of the group favoring knowledge and community participation, they also more commonly selected future political participation as an important aim of civic education compared to other teachers.
In conclusion, this analysis provides evidence that teachers' beliefs are not only associated with personal characteristics, but with the cultural and institutional contexts of a country (Hahn, 1998;Torney-Purta et al., 2005). Teachers' content beliefs may function as filters and guide their practices (Fives & Buehl, 2012;Thornton, 2005). Teacher preparation and development programs could be strengthened if they considered these contexts as well as the culture-related goals of CCE. In an inter-connected world with increasingly diverse societies, these programs also need to extend beyond the nation state to address global issues in which young people often show more interest than in their own countries' political processes (Barrett, 2017;OECD, 2016). Future studies should examine with mixed methods and perhaps in longitudinal perspective how cultural and institutional contexts influence teacher development and classroom practices in CCE as Singer, Braun, and Chudowsky (2018) have suggested for international assessments in general.

Limitations and future research
Limitations of the present study should be noted and provide avenues for future research. First, the ICCS provides cross-sectional data; this research was correlational and causality cannot be established. Furthermore, the indicators used to measure economic and democratic development may be correlated with other characteristics of a country or education system. Although large-scale longitudinal data collections are challenging at an international level, future waves of quantitative large-scale studies could make the collection of longitudinal data a priority to provide more possibilities for researchers to conduct comparative causal analyses. Mixed methods studies involving the analysis of curriculum documents in these countries and interviews could be helpful to explore these findings.
Second, this study examined three countries in four regions with different cultural and democratic traditions. These countries do not represent the whole spectrum of democracies, and rapid social and political change may have altered some perspectives since these data were collected. Although choosing countries as the highest sampling units could limit the potential for implications about regional and global dynamics (as argued by Beck, 2007), teacher education and curricula possess country-specific characteristics. Therefore, studying teachers' beliefs using samples from different countries and education systems is valuable. It is quite likely that teachers in still other democracies could be characterized within one or more of these profiles.
Third, due to the sampling strategy in ICCS, it is not possible to link teachers to their students. Although it is commendable that the ICCS collected data from representative teacher samples, many researchers in this field are interested in the associations between teacher beliefs, classroom practices, and student learning e associations about which firm conclusions have not yet been drawn (Torney-Purta et al., 2005). However, the role of teachers of unrelated subjects is difficult to discern. It would be valuable for future researchers to examine connections between teachers' beliefs and practices and how both are associated with students' learning outcomes.

Conclusions
This study aimed to shed light on an under-researched topic that has been recognized as important to the democratic health of societies. Generalizing results from cross-cultural research is challenging. However, this study using an innovative analytical approach showed that teachers' beliefs about the purposes and goals of CCE cohere in meaningful ways and are associated with individual teachers' characteristics and contextual factors such as the economic and democratic development characterizing a society. Furthermore, teachers can be classified in relation to their beliefs about the goals of CCE, and we expect that these groups of teachers will provide different types of instruction that then have the potential to influence students in different ways.
On one hand, this analysis showed that teachers rarely prioritize political participation even when they can select as many as three priorities of CCE. They are more likely to suggest students be participants in their schools or in their local communities. This replicates previous analyses of international data (Losito & Mintrop, 2001;Schulz et al., 2010Schulz et al., , 2017 and suggests that participation involving political goals is something teachers believe they should avoid. We speculate that political participation may be thought to be inevitably partisan, causing teachers to be wary about promoting participation. On the other hand, this study suggests that enhancing teacher development and student learning requires researchers, school leaders and curriculum designers to understand heterogeneity among teachers. Different aims of CCE may be associated with different pedagogical approaches (e.g. Torney-Purta & Amadeo, 2011). Those who hope to improve teacher pedagogy need to first understand what teachers' goals are. Some methods may be better suited to achieve a broader set of goals, others may only be helpful in achieving a particular goal. Furthermore, some teachers who emphasize the transmission of civic knowledge and the principles of society may see students as passive recipients of instruction; others may view students as active participants who co-construct the meaning of the topics studied; still others may focus on critical thinking or inquiry-based approaches to promote social justice (Knowles, 2017;Thornberg & O guz, 2016). Thus, teachers provide different learning opportunities which are also influenced by students' needs and affect student learning in diverse ways.
This study also provided insights regarding associations between teacher beliefs and national contexts. For example, the acceptance of democratic rules seems to be an important factor associated with teachers' endorsement of independent thinking and tolerance. Where democracy is under pressure or not fully developed, such as in Hong Kong or some of the post-Communist countries, instilling knowledge and active membership in the local community seem to be more important. Furthermore, cultural factors such as Confucian values appear to be relevant, yet other characteristics and moderating factors may be important.
Finally, teachers' beliefs have previously been found to influence decisions about teaching contents and classroom practices (shown with different methodologies by Patterson et al., 2012, andThornberg, 2008). Contextual factors appear to indirectly shape opportunities for student learning and participation. Future research should examine the mechanisms through which multiple contexts influence adolescents' civic development. The interactions between teachers' beliefs, school climates, and teachers' preparation as well as their civic and political engagement should be a focus. Eventually, this analysis raises questions. To give just one, how can the apparent disinclination to teach against racism be addressed? The factors to be considered include cultural and societal values, curricular prescriptions, teacher training, and personal epistemologies of teachers.