Faculty of Education - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 1 of 1
  • Item
    Thumbnail Image
    Rater consistency and judgment in the direct assessment of second language writing ability within the certificates in spoken and written English
    Smith, David R ( 1998)
    The introduction of competency-based models of language and literacy education in Australia has, to a large degree, coincided with an increased emphasis on direct assessment as the most common means of evaluating second language writing ability within the Adult Migrant English Program. The key problem in directly assessing writing ability is having two or more raters arrive at a similar judgment or rating for the same piece of writing. While there is a long tradition of research on rater consistency and judgment in the holistic assessment of writing ability, similar research on the direct assessment of second language writing ability within the context of competency-based language and literacy education is almost non-existent. This study aims to determine the degree to which the performance criteria designed to assess second language writing ability within the Certificates in Spoken and Written English can ensure acceptable levels of rater consistency, and to describe the decision-making behaviours and strategies used by raters when reading for the purposes of assessment. The think-aloud verbal reports of six experienced ESL raters assessing three texts written by intermediate level adult ESL learners were transcribed and subjected to a rigorous interpretive analysis. In terms of rater consistency, analysis of raters verbal reports indicated that while there was generally a high degree of rater consistency at the overall performance or text level there was considerably tess agreement at the level of individual performance criteria. Analysis of the data revealed that raters adopted distinctive styles or approaches to reading for the purposes of assessment and that raters interpreted and applied the performance criteria statements in a range of different ways. These findings have significant implications not only for the development of competency-based assessment procedures but also for the training of raters. v11