Values in language assessment: Validity’s elephant in the room
Tim McNamara

Nearly a quarter of a century after Messick’s definitive account of validity, the ‘sleeper’ issue of values in assessments, the third cell in Messick’s famous validity matrix, has still barely been acknowledged in language testing. While the social and political context of assessment is by now widely discussed, this has been mainly in terms of the consequences of test use, rather than the values embodied in test constructs. Messick himself naïvely assumed that test developers were able to determine, or at least could vouch for, the values in the tests they were developing, but the increasing policy-determined nature of test constructs in the form of standards and frameworks shows this not to be the case. In fact test constructs often are sites of struggle over competing social values. This paper considers the values implicit in widely-used policy-related assessment frameworks including the CEFR, workplace-related proficiency scales such as that of the International Civil Aviation Organization (ICAO), and tests in school settings such as PISA. It argues that engaging with values in assessment shows the limitations of current theories of validity and poses fundamental challenges for language testers.

Building on a career as EFL/ESL teacher and teacher trainer in Australia and the United Kingdom, Tim McNamara is now Professor in the School of Languages and Linguistics at The University of Melbourne, where he teaches and supervises in applied linguistics. His research interests include language testing and assessment, language and subjectivity, especially from a poststructuralist perspective, language teaching, particularly the teaching of languages for specific purposes, language and refugee issues, and language and citizenship. He is the author of Language Testing (OUP, 2000) and (with Carsten Roever) of Language Testing: The Social Dimension (Blackwell, 2006).

Validity in classroom assessment

Angela Hasselgreen

If validity in testing is a complex issue, then validity in classroom assessment is infinitely more so. Classroom assessment may include tests, but it takes on a wide range of forms, such as everyday tasks, dialogue, feedback and teacher questioning, and involves both the teacher and the learners in the assessment process. In fact it encompasses all the day to day ‘things’ that go on in a language classroom which shed light on what has been learnt or needs to be learnt. Thus to be valid, classroom assessment has to be judged not only on the basis of more familiar issues, such as whether tasks and criteria reflect what is being assessed, but also on the tricky areas of how useful feedback is, whether the dialogue in the classroom really reveals what is learnt, and indeed, how far the pupils are actually participating in this dialogue. Given that this type of assessment is what teachers rely on most in the important business of ensuring that students learn what is appropriate for them, getting it right is important. This paper will confront some of the theoretical and practical issues that need to be considered to ensure that our classroom assessment does the job it is supposed to be doing.

Angela Hasselgreen is Professor of English Didactics at the Bergen University College, Faculty of Teacher Education, Norway. She has a research background in language testing and assessment, particularly involving younger learners. She has worked with European partners in carrying out research into relating language assessment of school children to the Common European Framework of Reference for languages (Council of Europe). She has been involved in the development the new Norwegian school curriculum (Kunnskapsløftet) for English, as well as the national tests in English for schools. She is co-founder of the Norwegian national network for research into language learning in the primary school.

Reading Skills in the Information Age: Cognitive processes and implications for assessment
Jean-François Rouet

The advent of digital technologies has raised new issues concerning the nature of reading processes and reading skills, new challenges to defining the construct of reading literacy and thereby new requirements to the validation of reading assessment. Electronic information systems enable the production of new types of texts, new relations across texts, and new forms of reading. In order to effectively take part in these contemporary forms of print-based communication, students arguably need to acquire specific pieces of knowledge and skills. In this presentation, I review some key demands of electronic reading. I propose a framework aimed at identifying the key processes and knowledge structures involved in skilled use of complex information systems (Rouet, 2006; Rouet & Britt, 2011). I focus on processes of information search, information evaluation, and information integration across multiple sources. I discuss the similarities and differences with more traditional forms of reading. Finally, I point out some implications for the assessment of functional reading skills.

Jean François Rouet is a research scientist specialized in psychology, research director with the Centre National de la Recherche Scientifique in Poitiers, France. His work deals with the learning and use of complex information systems: text, hypertext, and other complex documents. He is interested in finding out how children acquire knowledge about documents, as well as strategies of information search; in adults, he studies the cognitive resources involved in searching and evaluating text information. He is also interested in cognitive ergonomics and the design of learning technologies.