Abstract
This article uses an educational effectiveness approach to model the impact of student, school, and educational system characteristics on several cognitive and non-cognitive student outcomes related to citizenship education. Using multivariate multilevel analysis, data from the International Association for the Evaluation of Educational Achievement (IEA) International Civic and Citizenship Education Study (ICCS) 2009 were analyzed, encompassing 102,396 lower secondary-school students (14-year-olds) in 4,078 schools in 31 countries. The results indicate that schools have a small influence on students' civic knowledge and hardly an impact on civic attitudes and intended civic behavior. Civic competences are mainly explained by individual student characteristics and out-of-school factors. Factors at the school level that were found to make a difference in students' civic competences are related to stimulating a democratic classroom climate in which free dialogue and critical debate on controversial political and social issues are encouraged, nurturing positive interpersonal relationships and creating opportunities for students to learn and practice democracy.
Notes
1. For further information on the factors covered by Creemers' model as well as empirical evidence of their effects, we refer the reader to Creemers (1994) and Creemers and Kyriakides (2008). Please note that especially due to the diversity of curricular approaches to citizenship education across countries (rarely a separate subject) as well as the conceptualization of citizenship education as the result of efforts made by the entire school team rather than by the individual teacher, some important factors related to student characteristics (e.g., time on task) and in particular teacher behavior in the classroom (e.g., orientation, structuring etc.) could not be covered or measured.
2. To control for the basic variations of the economic, democratic and educational contexts (see also Dogan, 2004; Hoskins, Barber, Van Nijlen, & Villalba, 2011; Schulz et al., 2011), we considered several indicators at the country level: Human Development Index, Human Development Index adjusted for inequality, Democracy Index, expenditure on education (secondary), and gross domestic product (GDP) per capita. We decided, however, to only include the Human Development Index in the analysis, because it highly correlates with all the other measures (r ≥ .74) and especially with the Democracy Index (r = .78).
3. While acknowledging the importance of educational system characteristics, we decided not to use the information from the ICCS national context questionnaire (one expert's opinion) on the characteristics of the educational policy with respect to the civic and citizenship education. In additional analyses, we have included the effects of such factors (e.g., priority given to citizenship education in the national curriculum; curricular guidelines for implementing citizenship education; support provided to teachers and school leaders in terms of the provision of initial and in-service teacher training for civic and citizenship education; assessment of students and schools in relation to civic and citizenship education) in the model specified in the current paper. These preliminary analyses showed that their effects were either not statistically significant (see also, Hooghe & Quintelier, 2011) or confounded with indicators of development. In fact, results reported here showed the level of development to explain most of the variance at the country level. Nevertheless, a more adequate measurement of these factors would likely yield more relevant knowledge concerning this research topic.
4. In interpreting these results, one must keep in mind that the reported parameters are correlations between the average school results across the countries. In fact, when conducting a more in-depth country-specific analysis, we could observe that in a limited number of countries schools deviated slightly from the overall pattern (e.g., in some countries the correlations between civic knowledge and conventional citizenship were small but positive).
5. When the country level of development was included, the correlations decreased by .20 on average.