In the past fifteen years, many countries have reformed their educational systems with reference to the PISA studies (Breakspear 2012). Also, with respect to educational language policies, a substantial influence is exerted by these studies (Cummins 2008). Such language policies are often informed by data that compare the academic achievement of native speaking (NS) students and language minority (LMi) students (e.g. Schnepf 2007). Some PISA reports have claimed that NS students outperform LMi students. Based on this finding, far reaching inferences have been drawn about the consequences of speaking and exposure to a minority language (ML). Many policymakers have responded to these PISA reports with a monolingual reflex by increasing political pressure on linguistic minorities to abandon their heritage languages (Agirda, 2010; Pulinx, Van Avermaet & Agirdag 2016). However, previous PISA assessments were not appropriate to examine the impact of students’ language use because very limited information was available about their actual language use. That is, only a dichotomous language background variable was assessed. The most recent PISA 2012 data provides a unique opportunity to fill this research lacuna as it includes a more elaborated questionnaire on linguistic issues (OECD 2013). As such, the purpose of this study was to contribute to the scholarship in this field by investigating how students’ language background and various indicators of language use are related to academic achievement. To examine the impact of language use and academic achievement, we conducted three-level multivariate regression analyses. The results revealed that there is indeed an achievement gap between LMi and NS students for both reading and mathematics. After taking account of students’ background characteristics, students’ academic profiles and school characteristics, the LMi-NS achievement gap narrows but remains significant. However, the reason for the underachievement of LMi is not their language use. LMi students who speak a ML more often with their parents do not achieve less. On the contrary, speaking a ML more often with the father is positively related to mathematics and reading achievement of LMi students. Nevertheless, exposure to the instruction language (IL) remains important, in particular in school context. Indeed, we found that speaking the IL with the best friend and/or schoolmates is positively associated with mathematics and reading achievement. This might reflect the negative effect of within-school segregation, and not so much related to between-school segregation, as we found that school LMi composition is not related to academic performance.

Agirdag, O. (2010). Exploring bilingualism in a monolingual school system: insights from Turkish and native students from Belgian schools. British Journal of Sociology of Education, 31(3), 307-321. Breakspear, S. (2012). The policy impact of PISA: an exploration of the normative effects of international benchmarking in school system performance (No. 71).Paris: OECD Publishing. Cummins, J. (2008). Essay review [Review of the book Where immigrant students succeed: A comparative review of performance and engagement in PISA 2003, by P. Stanat & G. Christensen]. Curriculum Inquiry, 38(4), 493–497 OECD (2013). PISA 2012 Results: Excellence through Equity (Volume II): Giving Every Student the Chance to Succeed. Paris: PISA, OECD Publishing. doi: http://dx.doi.org/10.1787/9789264201132-en Pulinx R., Van Avermaet, P. & Agirdag O. (2016). Silencing linguistic diversity: The extent, the determinants and consequences of the monolingual beliefs of Flemish teachers. International Journal of Bilingual Education and Bilingualism. doi:10.1080/13670050.2015.1102860. Schnepf, S. V. (2007). Immigrants’ educational disadvantage: an examination across ten countries and three surveys. Journal of Population Economics, 20(3), 527-545.