Dissertation/Thesis Abstract

Effects of rater characteristics and scoring methods on speaking assessment
by Matsugu, Sawako, Ph.D., Northern Arizona University, 2013, 211; 3562150
Abstract (Summary)

Understanding the sources of variance in speaking assessment is important in Japan where society's high demand for English speaking skills is growing. Three challenges threaten fair assessment of speaking. First, in Japanese university speaking courses, teachers are typically the only raters, but teachers' knowledge of their students may unfairly influence scoring. Second, in many speaking classes, teachers are often native speakers of English who may not be appropriately trained. Third, conducting practical speaking assessment is a major concern in classroom. Semi-direct testing using digital recorders could solve such issue and achieve reliability and efficiency in speaking assessment (Qian, 2009).

This study examined these sources of score variation--the effects of rater characteristics and scoring methods on speaking assessment by university instructors in Japan. Rater characteristics included scorers' knowledge of examinees and to a much lesser extent their native language differences. Scoring methods were live and delayed scoring with digital recorders. These factors were believed to be potential threats to fairness.

Participants were the researcher and the instructors (N = 9) as well as students (N = 169) in nine classes in four universities. In oral skills classes in Japanese universities, students were given one speaking task scored by instructors live and recorded. Overall raters' knowledge of examinees, despite some perceived influence of knowledge of examinees, was not a significant source of variance. For scoring methods, despite some raters' preference for delayed scoring, there were no statistically significant findings between live and delayed scoring. The two groups of raters (7 English and 1 Japanese), not including the researcher, focused on different speech characteristics. These results suggested that variations in speaking scores were not derived from differences in rater characteristics and scoring methods.

Implications include the following: instructors being the only rater depending on the purpose of assessment and circumstances, wider choices of speaking assessment methods to facilitate fairness and practicality, and possibly more involvement of Japanese instructors in teaching and assessing of speaking skills.

Indexing (document details)
Advisor: Jamieson, Joan
Commitee: Kang, Okim, McClure, John, McGroarty, Mary
School: Northern Arizona University
Department: English
School Location: United States -- Arizona
Source: DAI-A 74/09(E), Dissertation Abstracts International
Subjects: Educational tests & measurements, English as a Second Language, Foreign Language
Keywords: English language learners, Rater characteristics, Scoring methods, Speaking assessments
Publication Number: 3562150
ISBN: 978-1-303-09658-7
Copyright © 2020 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy