Language Testing

In this article I argue that a growing awareness of the fundamentally social character of language assessment challenges us to rethink our priorities and responsibilities in language testing research. This awareness has been brought about by the treatment of the social character of educational assessment in Samuel Messick’s influential work on validity, and by the intellectual changes triggered by postmodernism, where models of individual consciousness have been reinterpreted in the light of socially motivated critiques. The article concludes by arguing that the institutional character of assessment often means that the needs of learners are not well served by much language assessment theory and practice, and calls for a reexamination of our research priorities.

This article examines language assessment from a critical perspective, defining critical in a manner similar to Pennycook (1999; 2001). I argue that alternative assessment, as distinct from testing, offers a partial response to the challenges presented by a critical perspective on language assessment. Shohamy’s (1997; 1999; 2001) critical language testing (CLT) is discussed as an adequate response to the critical challenge. Ultimately, I argue that important ethical questions, along with other issues of validity, will be articulated differently from a critical perspective than they are in the more traditional approach to language assessment.

The article describes the strong power of tests and the fact that tests lead to far-reaching and high-stakes decisions and consequences about individuals and groups. Further, there is evidence that tests are often introduced by those in authority as disciplinary tools, often in covert ways for the purpose of manipulating educational systems and for imposing the agendas of those in authority. Yet, such uses of tests as instruments of power violate fundamental values and principles of democratic practices. The article proposes a number of assessment strategies which are based on democratic principles so that society can guard and protect itself from such undemocratic practices. The principles include the need:

• for citizens in democratic societies to play a participatory and active role and transfer and share power from elites to and with local bodies;

• for those who develop powerful tools to be responsible for their consequences;

• to consider voices of diverse and different groups in multicultural societies; and

• to protect the rights of citizens from powerful institutions.

These lead to assessment practices which are aimed at monitoring and limiting the uses of tests, especially those that have the potential to exclude and discriminate against groups and individuals. Specifically, assessment practices include the need:

• to examine the uses of tests through critical language testing (CLT);

• to develop assessment models that are based on shared and collaborative models;

• to assume a growing responsibility for those who are engaged in test development and use;

• to examine the consequences of tests;

• to include different voices in assessment, especially in multicultural societies; and

• for test-takers to protect and guard their rights from the authority and misuses of tests.

The implementation of outcomes-based assessment and reporting systems in educational programs has been accompanied by a range of political and technical problems, including tensions between the summative and formative purposes of assessment and doubts surrounding the validity and reliability of teacher-constructed assessment tasks. This article examines ways in which these problems have been manifested and addressed, using two recent examples from school and adult immigrant education in Australia. The first example concerns a recent controversy surrounding the use of national literacy benchmarks for primary school learners. Analysis of the issues suggests that some learner groups may be disadvantaged by the practice of reporting aggregate outcomes in terms of minimum standards, but that government policy is unlikely to change as long as the accountability function of assessment remains paramount in the public eye. The second example discusses the teacher-developed assessment tasks that are used to assess the achievement of language competencies in the Australian Adult Migrant English Program (AMEP). It is argued that problems of consistency and comparability that have been identified by research can be addressed through the development of fully-piloted task banks and the provision of appropriate forms of professional development. Greater attention needs to be given to the role of the teacher if outcomes-based assessments are to provide high quality information.

Within the context of accountability for US schools, standardized achievement tests are being used for increasingly `high stakes' decisions for all students including those for whom English is a second language, even when their English language skills are not adequate for the task. This article discusses approaches to the standardized assessment of content knowledge for English language learners (ELLs),1 including testing in the student's first language, the use of test accommodations, and measuring growth in English as an alternative for accountability until student control of English is sufficient to assure validity of test scores. Limitations of current research on the use of standardized content assessments with ELLs are presented and alternative approaches suggested.

In a recent issue of Language Testing, Rea-Dickins and Gardner (2000) reported on their work in classroom-based assessment, and provided examples of different ways in which information from learner assessments was used by teachers working with learners for whom English is an additional/second language (EAL). The research reported here is also concerned with classroom assessment in an EAL school context. As in the previous article, it is presented from the perspective that issues of classroom assessment and, in particular, formative assessment require further detailed analysis. Using data from teacher interviews, classroom observations, video and audio recordings of learners, and lesson transcripts, the starting point for this investigation is the concept of the assessment cycle. Taking a grounded approach, it traces different stages in the teacher assessment process and presents a working model for the analysis of teacher decision making in relation to assessment practices. At the same time, it identifies classroom assessment as a multifaceted phenomenon with distinct identities linked to learning, teaching and bureaucratic functions.

Authenticity is now firmly established as a central concern in test design and test validation (Bachman, 1990). However, there is disagreement about what authenticity is and about the degree of authenticity that can realistically be achieved. This article explores the theoretical and practical issues surrounding authenticity in course-based assessment, drawing on data from a university-level Japanese language course in Australia. It examines a teaching and assessment activity based around interviewing native speakers outside the classroom, which was designed to optimize authenticity. Using tapes that students made of the interview for assessment and retrospective interviews, the study examines various dimensions of authenticity and reveals a wide diversity in individual experiences. The article argues that the addition of an assessment dimension fundamentally changes the nature of a task, and thus compromises authenticity. It further suggests that authenticity must be viewed in terms of the implementation of an activity, not its design, and examines the various factors that affect the degree of authenticity experienced by individual students. The implications of this diversity for the validity of the activity are discussed.