To draw conclusions about students´ understanding solely based on test results is not an uncomplicated endeavour. An incorrect response does not necessarily indicate a lack of knowledge. A lack of vocabulary, deficit in knowledge of the language of the test, poor reading skills or misinterpretations can also be the cause (Harlow and Jones 2004; Oakland and Lane 2004). Another problem is that some students may have necessary skills and strategies to solve a problem, but may still interpret the question in another way than the intended (Schoultz 2000; Harlow and Jones, 2004). The reading demand the question has on students can influence what is being measured (Haladyna et al., 2002).

The study presented in this paper is part of a larger study of the academic language of school science, and its significance for students’ reading and solving of questions in science. Each sub-study will approach various aspects of the academic language in school science such as linguistic features, vocabulary and coherence. Focus in this first sub-study is linguistic features concerning questions in science in the international study The Trends in International Mathematics and Science Study (TIMSS). Using a large scale study like TIMSS, it is possible to investigate the significance of language in a systematic manner for different science subject areas and also different groups of students.

This study aims to:

• Develop a method for characterizing science questions regarding linguistic features.
• Examine which -and to what extent- linguistic features influence how different groups of Swedish 8th grade students succeed in answering questions posed within different science subject areas in TIMSS.

To develop a method for characterizing science questions regarding linguistic features this study is based on a social semiotic perspective as it is developed in systemic functional linguistics (SFL) (Halliday 2004). A text and its features are seen as a realization of the context of situation, and thus reflecting the social order. In different social contexts, different registers of language are used. In SFL a register is described as a configuration of what the text is about (field of discourse), personal relationships in the text (tenor of discourse), and how the text is structured (mode of discourse) (Halliday and Hasan 1989). Each discourse could be described as a continuous scale. The continuum for field of discourse spans from common-sense and everyday aspects to technical and specialized. The register tenor of discourses goes from personal, informal and subjective aspects to formal and objective. The mode of discourse continuum covers less complicated and unpacked language to more complicated and packed language. By using register as a base for the linguistic analysis the results will also be more generally applicable for different languages.

To discuss results of different groups of students this study is also based on a reception theoretical perspective (for a school context see e.g. Graesser et al. 2011; Langer 2011; Liberg et al 2012; Luke & Freebody 1999). In such a perspective the text as such or the language of the text is not seen as inherently cognitive demanding. The cognitive demands are relative the encounter between reader and text within a specific context, i.e. no task can be called cognitively demanding out of context (Schleppegrell 2004). Of importance are the reader’s experiences and knowledge of the language used, the reader’s pre-knowledge of and interest in the content of the text, and of reading such texts and participate in practices where such texts are read (e.g. Liberg 2001). Results of six different groups of students will be investigated here, concerning language background and thereby presumed skills in the language of the test, sex (boys/girls), and high/low performance on the test.

Graesser, Arthur C. & McNamara, Danielle S. & Kulikowich, Jonna M. (2011). "Coh-Metrix: Providing Multilevel Analyses of Text Characteristics." Educational Researcher 40(5): 223-234. Harlow, A. and A. Jones (2004). "Why Students Answer TIMSS Science Test Items the Way They Do." Research in Science Education 34(2): 221-238. Haladyna, T. M. & Downing, Steven M. & Rodriguez, Michael C. (2002). "A Review of Multiple-Choice Item-Writing Guidelines for Classroom Assessment." Applied Measurement in Education 15(3): 309-334. Halliday, M. A. K. & Hasan, Ruqaiya (1989). Language, context, and text: aspects of language in a social-semiotic perspective. 2. ed. Oxford: Oxford Univ. Press Halliday, Michael. A. K. (2004). An introduction to functional grammar. 3. ed. London: Arnold. Hyland, Ken & Tse, Polly (2007). "Is There an "Academic Vocabulary"?" TESOL Quarterly 41(2): 235-235. Langer, Judith A. (2011). Envisioning knowledge: building literacy in the academic disciplines. New York: Teachers College Press Liberg, Caroline (2001). Svenska läromedelstexter i ett andraspråksperspektiv - möjligheter och begränsningar. In: Nauclér, Kerstin (red.) Symposium 2000 – Ett andraspråksperspektiv på lärande. Nationellt centrum för sfi och svenska som andraspråk. Stockholm: Sigma Förlag. S. 108-128. Liberg, Caroline, af Geijerstam, Åsa, Wiksten Folkeryd, Jenny, Bremholm, Jesper, Hallesson, Yvonne & Holtz, Britt-Maria. (2012). Textrörlighet - ett begrepp i rörelse. nI: Synnøve Matre och Atle Skaftun (red.), Skriv! Les! 1. Paper presented at Skriv! Les! Nordisk forskerkonferanse om skriving og lesing. Trondheim: Akademika forlag. S. 65-81. Luke, Allan & Freebody, Peter (1999) Further notes on the Four Resources Model. Reading online. http://www.readingonline.org/research/lukefreebody.html [downloaded 2012-02-13]. Oakland, Thomas & Lane, Holly B. (2004). "Language, Reading, and Readability Formulas: Implications for Developing and Adapting Tests." International Journal of Testing 4(3): 239-252. Schoultz, Jan (2000). Att samtala om/i naturvetenskap : kommunikation, kontext och artefakt. [diss.]. Linköping: Univ; 2000. http://stp.lingfil.uu.se/timss/#Publications