The main purpose of this dissertation, and the four studies it includes, is to investigate some ways in which the concept of writing ability is rendered concrete in different national writing tests. The dissertation combines two theoretical approaches. Assessment theory serves as a framework for the approach to national tests. Tests are understood as a practice aiming at providing evidence for the inferences teachers need to draw about students’ writing ability, in order to make decisions that are part of school practice, such as giving grades. I examine the tests from the perspective of an anthropological theory of text that takes dialogism as its point of departure and views texts as culturally valued utterances. The analyses focus on the relations between writing task, students’ texts and assessment criteria, and how these can be both in alignment and in conflict with the claims on score meaning made by different test constructions. In the first study, four students’ texts from the national writing test in upper secondary school are closely analyzed to show how a writing task regulates students’ scope for action in their texts as regards global inner structure and perspective on content. The second study investigates how a national writing test was carried out in two classrooms in primary school, and highlights the relations between the task, students’ texts and assessment criteria. In the third study, I argue that a pursuit of reliability in writing tests becomes misguided unless it takes as its starting point the definition of writing ability the tests are made to measure. In the fourth study I examine the writing tasks of the tests in upper secondary school from 2007–2012. The analysis brings four distinct and recurrent task types to the fore, from which a reconstruction of the target domain of writing within the school subject of Swedish is made. Overall, the dissertation contributes to critical reflection on the understandings of writing expressed by test constructions, and points at possible further development of writing tests, given the theoretical perspective on writing that the dissertation employs.

This thesis presents a case study from an authentic school practice, where seven students, in their second year of a social sciences program in an upper sec­on­dary school, use internet texts in various learning situations. The aim of the study is to map the reading practices of students encountering internet texts. The main data con­sists of obser­vations, audio and screen recordings, writ­ten instruc­tions, and screen ­shots of the sites visited. Reading practices are ana­lysed, draw­ing on concepts from New Literacy Studies and Systemic Func­tional Grammar, inclu­ding literacy events, literacy in terms of text cul­ture, text­ual norms, abstraction, auth­o­rity and mod­ality as a scale of reliability. The results reveal that meaning making resources such as colours, amount of writing and images and choice of fonts all seem to be parts of students’ con­ceptions of reliability. These textual norms result in learning situations in which students search for texts with pre­dominantly dense writing promoting ency­clopaedic know­ledge. These highly auth­orit­ative texts can be hard to under­stand for the students, something that the text analyses indicate. In com­parison to text books, the internet texts used show, a higher level of auth­ority and abs­trac­tion, rein­forced by gram­matical meta­phors. Most situ­ations in the study include peer interaction, but the most obvious learning poten­­tial resides in situations with a clear reading goal, where stu­dents work in groups and where negotiation is part of the meaning mak­ing pro­cess. The pedagogical implications of the study suggest the potential for students to achieve a higher degree of understanding of the encountered inter­net texts, through group work, and discussions concerning the impact of different layouts and the demands of verbal language. Another potential con­cerns methods for avoiding critical literacy being reduced to trivial visual scanning, via dis­cussions focusing on criteria for reliability evaluations. It is suggested that increased teacher awareness concerning the types of internet texts the students will encounter in authentic situations may contribute to students’ field and genre insight.