Good assessment practice can be defined and discussed from a number of starting-points, focusing on theoretically oriented, as well as on impact or feasibility related considerations. The aim of the proposed presentation is twofold; firstly to address the issue from a process oriented and empirical perspective, referring to a Swedish as well as a European context; secondly to briefly describe and discuss a set of recently developed Guidelines for Good Practice, currently being implemented at the European level. The theoretical framework for the presentation is a unified and expanded view of validity, as expressed, for example, by Messick (1989), emphasizing construct based interpretation and inference, as well as possible consequences of test use. In addition, a somewhat more pragmatic approach to the concept of validity and validation, referring, e.g., to Kane (2006), is applied. The examples given stem from the field of language assessment, at local as well as at national and European levels. Principles and practice characterizing the development of national assessment materials for foreign languages in Sweden are presented. The approach adopted can be characterized as distinctly collaborative and reciprocal, with systematic involvement of teachers, teacher trainers, researchers within different fields, and, perhaps most importantly, large numbers of students of different ages (Erickson, 2006, 2007; Erickson & Lander, 2007). Examples of contributions from three of these categories, viz. students, teachers and researchers, are given. Special attention is paid to the large-scale collection of test-taker feedback routinely done in pre-testing phases, the principles of which are related to literature within the field (Alderson, Clapham & Wall, 1995; Bachman & Palmer, 1996; Gipps, 1994), and to studies of dimensionality of test performance (Åberg-Bengtsson & Erickson, 2006). Furthermore, a European survey of students’ views on language testing and assessment is briefly presented and related to the issue of Good Practice (Erickson & Gustafsson, 2005). Ca. 1 400 adolescent students in 10 European countries completed a questionnaire focusing on issues of content, format and frequency, as well as on students’ definitions of, and reflections on, ‘good’ vs. ‘bad’ language testing and assessment [practice]. Qualitative analyses of the latter are presented, and examples are given of the relationship between the perceptions and reflections of the students, and validity theory. The implications and impact of the survey are briefly touched upon, inter alia, with reference to possible needs for guidelines to improve the quality of assessment and testing, and to raise the level of assessment literacy in different contexts, and among different groups of stakeholders. Finally, the EALTA Guidelines for Good Practice in Language Testing and Assessment, adopted in 2006, and to some extent influenced by the questionnaire study described above, are used as an example of an operational definition of the concept of Good Practice (EALTA = European Association for Language Testing and Assessment). In this document, currently translated into 32 languages, different categories of agents and situations are addressed. Initially, however, it is emphasized that a number of general principles apply, regardless of type of testing or assessment context. These concern respect for students/examinees, responsibility, fairness, reliability, validity and collaboration among the parties involved.

In spite of the fact that the examples in the proposed presentation are related to language assessment, it is claimed that they demonstrate a considerable degree of generality and may thus also be applied to other [subject] domains. One of the conclusions drawn is that collaboration among stakeholders, including test-takers, may be defined as a crucial element of good practice in testing and assessment. Furthermore, the results obtained in the projects briefly touched upon, give substantial evidence to the claim that students often demonstrate impressive insights and analytic competence when it comes to defining quality of assessment processes and materials. Finally, based on theoretical notions, interpretations and definitions of validity, as well as on empirical examples of well functioning processes and practices, a tentative conclusion is that, concerning Good Practice, traditional boundaries, e.g. between formative and summative assessment, or indeed between assessment and testing as such, need to be questioned, and, if not abolished, at least toned down to a considerable extent.