Test developers have to attend to all aspects of validity throughout test development and implementation. As one of the major aspects, scoring validity has to be established for the dependability of scores assigned to a test performance. This study is an investigation into the scoring validity of a speaking test of Turkish as a second language (TSL). For this purpose, in this study, six tasks and a rating scale were developed and administered to twenty-four L2 learners of Turkish whose performance was evaluated by four raters. The score dependability was investigated through Generalizability (G) and Decision (D) analyses. The results indicated that most of the score variation could be attributed to test takers, and not to error variance, i.e. raters and tasks.
Assessing speaking in Turkish as a second language scoring validity generalizability analysis
Primary Language | English |
---|---|
Journal Section | Original Articles |
Authors | |
Publication Date | December 17, 2018 |
Published in Issue | Year 2017 Volume: 34 Issue: 1 |
This work is licensed under a Creative Commons Attribution 4.0 International License.