Scientific interest concerning the impact of school’s compositional effect on student performance has intensified throughout the current century, as the subject has been vastly explored by researchers in various national contexts (Agirdag, Van Houtte & Van Avermaet, 2012; Jensen & Rasmussen, 2011; Van der Slik, Driessen & De Bot, 2006; Schnepf, 2007; Goldsmith, 2003, Lleras, 2008). These effects have usually been studied taking student outcomes as measured by standardised tests on different school subjects and levels of education. In Portugal we conducted a research using an extensive database (16 269 students and 417 schools) with the results of fourth grade students on Portuguese Language national standardised tests in the Lisbon Metropolitan Area (2009-10). We intent to reveal what are the main factors behind students results in public schools giving special attention to the effect of ethnicity, considered both at the student level (its national origin) and at the school level (the schools ethnic composition). A multilevel analysis was developed involving both individual level variables and school level variables to answer the following research questions: Does the schools’ ethnic composition effect on Portuguese Language scores stand when students’ gender, social and ethnic origins are taken into account? Does this effect stand when the schools’ social composition is taken into account? How does the schools’ ethnic composition moderate the relation between having/not having an immigrant background and students’ Portuguese Language scores, when the socio-economic status (SES) of both students and schools is controlled?

--