The integration of migrants is at the centre of current social, economic and education policy debates in Europe. Substantial national policy differences within the European Union create different settings and patterns of integration of migrants across Europe. Education can be understood as a process through which integration is achieved. The overarching aim of this paper is to study interconnections between various institutional characteristics of migrant accepting societies in their relationship to the educational integration of migrant children. This paper addresses the following research questions: 1. Does empirical evidence demonstrate that the educational integration of migrants is more successful in national systems with particular education system, labour market, welfare and migration policy characteristics? 2. What is the analytical relationship between welfare regimes, migration and education systems in their relationship to the educational integration of migrants? 3. Can the current large system used to capture the national system level effects on the educational integration of migrants be replaced with a small number of factors reflecting essential institutional dimensions of migrant educational integration? 4. Do combinations of these factors produce a unique pattern of educational integration? The following theories which relate between the educational integration of migrants and institutional features of national systems provide us with an analytical framework for identifying the connections between the key dimensions of educational integration and key elements of educational systems welfare regimes and migration policy: This study draws on the theoretical and empirical research which developed typologies of welfare regimes (Esping-Andersen, 1990; Ragin 1994; Bambra, , 2007; Hill, 2009; Kautto 2002). Existing typologies and the criticism on these typologies is used as a basis and a motivation for creating and testing empirically a new typology through the use of the robust statistical methodologies. Furthermore, this study is based on research which links the type of welfare regime to migrant integration (Kogan, 2007; Sainsbury, 2006; Fleischmann and Dronkers, 2007) as well as the theories and the empirical research regarding the nature of the relationship between education systems and the welfare state (Castles, 1993, 1989; Hega and Hokenmaier, 2002; Heidehheimer, 1981). The present study takes this research forward and through elaborated empirical analysis tests empirically the relationship between the welfare state and the education system ((Wilensky,1975; Wilkens, 2005). The present study also draws on research that relates between migration policy in terms of the degree of its selectivity affects both the absolute and relative educational levels of migrants and hence the patterns their integration in host countries (Cerna, 2009, Belot and Hatton, 2008). The potential link between migration policy, welfare regimes and educational systems and migrant integration draws on theories which relate types of labour markets and types of education systems (Messina, 2006; Medina, 2009; Marsden and Ryan, 1986; Müller and Gangl, 2003), and also between the type of the labour market and consequent level of socio-economic inequality which it creates (Esping-Andersen, 1999; Iverson, 2005).

The 2006 Programme for International Student Assessment, database on education systems in Europe (Eurydice, 2009) the comparative welfare state dataset (Huber et al, 2004), UN 2006 Human Development Index , the Civic Citizenship and Inclusion Index (Geddes et al 2004) and other administrative data. Multilevel regression modelling is used to capture the effect of the variables measured at the macro-level of countries and the meso-level of schools, on the dependent variables, measured at the level of individuals. Factor and cluster analyses are used to study the relationship between characteristics of national systems and reduction of their number to a small set of factors. Both first and second generation of migrants are considered. Non-migrant population is used as a reference level. The educational integration is measured through educational attainment, family resources, motivational and attitudes to learning, access to various kinds of the educational resources available through schools and education systems.

This is ongoing study. The analysis to date suggests the following broad findings: The amount of variance of educational attainment explained by individual level variables and by the country level variables differs cross-nationally, by subject, and between migrants and non-migrants; (a) Individual level: • Thus, family background and family resources have better explanatory power for non-migrants and affect their attainment more strongly; • In contrast, variables such as motivation and regular studies have better explanatory power for migrants and contribute more to their attainment than to the attainment of non-migrants. (b) National level: • The attainments of second and first generation migrants are higher in countries with: (i) selective migration policies; (ii) less differentiated education systems and higher levels of state involvement in education; (iii) broader social welfare provision. Various characteristics of education policy and systems, the characteristics of the type of welfare state and migration policy correlate highly and cluster well with each other in a meaningful way, and allow for the building of dimensions of national systems regarding education and educational attainment. They are thus very promising in their potential for the cross-national study of the impact of institutional characteristics of national systems on educational integration.