Although the debate about school effect has around 5 decades (since Coleman Report, 1966), there remain difficulties in assessing such effect and, above all, in understanding the conditions under which it varies. In other words, we need to deepen our knowledge regarding, for example, the effect of schools social composition, or the effect of pedagogical work of teachers, in the results of students considering their social and national origins. While studies on the school effect have a long tradition in Anglo Saxon countries (e.g. Jencks et al., 1972; Rutter et al., 1979; Smith and Tomlinson, 1989; Entwistle et al., 1997; Thrupp, 1999; Oakes, 2005), other countries only recently have conducted studies on this subject (e.g. Cousin, 1998; Cervini, 2006; Szulkin & Jonsson, 2007; Portela et al, 2007; Pereira, 2010; Seabra et al., 2014). Concerning the specific situation of students with immigrant background, the school effect can be different for these students when compared with those who are natives. Generally speaking, we know that students living in disadvantaged social conditions are more sensitive to the effects of the school environment and benefit when they are in more favored environments from a social point of view. But what happens for students who have an immigrant background? Portes and MacLeod (1999) find that in the USA "Mexican-American students do significantly worse than their peers when they are in private schools mostly white, but their scores on the tests are still superior to their co-ethnics in the public schools of minorities"(p.389). On the contrary, "students of Chinese / Korean origin appear as impervious to potential handicaps of the schools they attend: they also have good results in relation to their peers, whether attending high status schools and poor schools with a lot of minority population, [which] once again highlights the vulnerability of children from state of disadvantage "(p.391). To contribute to this debate, a research was conducted in Portugal, using an extensive data base with the results of fourth and sixth grade students in the Portuguese and Mathematics national exams in Lisbon Metropolitan Area (LMA) in 2009/10. In Portugal about 5% of pupils have a foreigner nationality but in this regional area we can find about 25%. The main objective was to explore the differences in the results of national exams between students with immigrant background and autochthonous, considering the composition of schools. We have analyzed the school composition in three dimensions (social background, ethnic origin and the school performance) and compared these effects in Math’s results and Portuguese’s results. The theme of the school composition effect on student outcomes has been the subject of recent research (Agirdag, Van Houtte & Van Avermaet, 2012; Dumay & Dupriez (2008); Jensen & Rasmussen (2011); Van Houtte & Stevens, 2009) revealing the importance of their role. Methodology, Methods, Research Instruments or Sources Used Data Our data covers approximately 50,000 pupils, from almost 800 schools (611 in the fourth- grade; 162 in the sixth-grade), representing the universe of public schools and students in Lisbon Metropolitan Area. The database that supports this research was made available by the Ministry of Education Portuguese and contains variables about students’ sociodemographic characteristics and also about their results in national exams in each school in Mathematics and Portuguese. Measures A dichotomous variable was considered to measure the national origin of students (descended from immigrants versus autochthonous) and was computed at the individual level (independent variable). The dependent variables were Mathematics and Portuguese achievement, assessed by the results obtained by the students in the National Attainment Tests on fourth and sixth-grades. The three moderated variables were computed at the school level. The school social composition was assessed trough the average of students' social background, the ethnic composition by the % of students with immigrant origin and the academic performance by the % of levels 4 or 5 (the highest marks in the Portuguese scale on fourth and sixth-grades). Data analysis Firstly, a Multiple Correspondence Analysis (MCA) was used for the definition of student’s socio-economic index with three input variables: social class (occupation and employment status), parents’ educational level and economic support (ES). In order to explore if school conditioned the relation between national origin of students and their results in Mathematics and in Portuguese national exams, different moderated models were tested. We tested the effects of three moderated school variables: social composition, academic performance and ethnic composition. The schools variables were computed considering the 800 schools and then they were imputed in student’s database. The conditioned effects were tested using a parametric method (ALS) through PROCESS macro for SPSS version 21 (Hayes, 2013). Post-hoc probing were also conducted to interpret significant moderator effects (Hayes and Matthes, 2009; Preacher, Curran and Bauer, 2006). The graphical display of moderation analyses joined to the results in order to represent the configurations of the moderated results. Conclusions, Expected Outcomes or Findings As the school composition was analyzed in three dimensions– social background, ethnic origin and school performance – we were be able to compare the results and deepen our knowledge about the way school composition affects the academic results of students descended from immigrants. One of the main conclusions was that students with an immigrant background have worse outcomes when they attend schools with a strong presence of students with this profile. The contrast between descendants of immigrants and natives is emphasized in schools with high percentages of descended of immigrants. Other conclusion is that the autochthonous students have, on average, better math results. However, this difference is not significant when the presence of descendants of immigrants in schools is lower. In other words, when students with an immigrant background attend schools with a low percentage of descendants of immigrants, they tend to have similar results to those of autochthonous students.

--

school effect, ethnic origin, moderation effects