A B S T R A C TAcademic language is frequently assumed to be especially challenging for students from families of low socioeconomic status ( SES ) and even more so for language-minority students. Due to their often especially disadvantaged position regarding socioeconomic background and exposure to the language of instruction, language minority students are considered to suffer from a double disadvantage when processing complex academic language. To test this assumption, the present study investigated the relationships between various academic language features and differential item functioning ( DIF ) in a reading comprehension test for language-minority students on the one hand and German monolingual students from low-SES families on the other hand. The analyses are based on data of 19,108 fourth-grade students who took part in the reading comprehension test of the German National Assessment Study in elementary school. Our findings indicate that both lexical and grammatical features of academic language correlate with DIF disfavoring language-minority students, with especially pronounced effects for long and complex words and average sentence length. For German monolingual students from low-SES families, fewer features were associated with DIF , and the correlations were generally smaller than for language-minority students. Findings are discussed in relation to the assumed double disadvantage of language-minority students in the comprehension of academic language.I nternational large-scale assessment studies, such as PISA (Programme for International Student Assessment), have repeatedly shown that both immigrant students and students from families of low socioeconomic status (SES) are less successful in school than their peers (OECD, 2004(OECD, , 2006(OECD, , 2010. Although these effects have been found in most countries, Germany is among those with the largest achievement gaps (Ehmke & Jude, 2010 ;OECD, 2006 ). As early as the elementary school level, children with at least one foreign-born parent perform significantly more poorly in mathematics, natural sciences, reading, and listening comprehension than do children whose parents were both born in Germany (Haag, Böhme, & Stanat, 2012 ;Schwippert, Wendt, & Tarelli, 2012 ;Tarelli, Schwippert, & Stubbe, 2012 ). Similarly, elementary school children from low-SES families tend to fall behind their peers from average-or high-SES families in mathematics, sciences, reading, and listening comprehension (Richter, Kuhl, & Pant, 2012 ;Stubbe, Tarelli, & Wendt, 2012 ). Because the command of academic language is considered as playing a key role in educational success (e.g., Bailey, Francis, Rivera, Lesaux, Kieffer, & Rivera, 2006 ), limited academic-language proficiency may at least partly account for these pronounced group differences in academic achievement.
Birgit Heppt
Academic LanguageAcademic language can be defined as the language that is spoken in school or other academic settings to impart and acquire knowledge (Anstrom et al., 2010 ;Chamot & O ' Malley, 1994 )...