Ari Huhta (University of Jyväskylä): Intersections between second language acquisition and language testing research - Contributions from diagnostic assessment

The presentation discusses recent research on diagnostic assessment of second and foreign language (SFL) skills, and the ways in which such research can contribute to SLA research. The study that will be used as the key example of diagnostic assessment research in the presentation is DIALUKI, a large-scale investigation of the development of reading, writing and vocabulary in English as a foreign language and Finnish as a second language among young 10 to 17-year-old learners. Other studies on diagnosing second/foreign language skills will also be used as examples whenever relevant. Both the overall design of the studies that comprised the DIALUKI project and the development of the data gathering instruments illustrate decisions and challenges that also many SLA studies are likely to face. One of these is the role of learners with possible learning problems such as dyslexia in the study -- should they be included or not, and if yes, how does one identify them in the first place. Another is the challenge of securing the co-operation and motivation of young learners (as well as their teachers and parents) to participate in the study and, for the learners, to do their best on the tasks that are not an integral part of their school work.

When it comes to designing data gathering instruments, the key consideration is their quality: are already validated instruments available or do they have to be developed and how to go about if the latter is needed. This is obviously an area where language testing research can provide us with useful guidance. A particularly interesting aspect of the design of the data gathering instruments in the DIALUKI project was an effort to not only design measures that provide us with information about individual differences in the skills and characteristics of interest among the learners (based on ranking the learners) but also with criterion-related interpretations of the scores. In language assessment, and also more broadly in language education, the Common European Framework of Reference (CEFR) levels provide us with such a criterion. The usefulness of such interpretations and how the project approached the challenge will be described in some detail.

The procedures applied in DIALUKI to the data collection and subsequent coding and analyses, and the issues we faced in them, are also likely to be relevant to many SLA studies. These include the advantages and disadvantages of individual vs group administration of the tasks, the use of computer vs paper-based tasks, and the specific issues related to administering tasks to rather young learners. Our experiences with the data analyses may also be of interest to SLA research, including the Rasch analysis of the test scores and ratings and the findings concerning the language of administration of measures of various cognitive / psycholinguistic skills.