Constructing and Validating Parallel Forms of Performance-Based Writing Prompts in an Academic Setting

"Based on the findings of the investigation on prompt comparability, it was concluded that the degree of comparability among the three new EPT writing prompts was generally quite high in terms of both the placement reproducibility and the skill profile consistency, despite their slight differences in terms of the ""fit-to-specification"". In addition, it was found that the three new writing prompts functioned similarly when responded to by ESL examinees from different subgroups. Furthermore, the results of the outlier analysis on both examinees and raters revealed that the three new writing prompts are indeed valid test instruments for the EPT test-takers and, in general, the EPT raters do interpret the rating rubrics consistently. The findings on prompt comparability and differential prompt rubrics consistently. The findings on prompt comparability and differential prompt functioning imply that the multi-faceted measurement model provides valuable diagnostic information concerning how the various individual aspects of a writing assessment program are functioning. Furthermore, criterion-referenced-based test specifications provide an important basis for developing and validating parallel forms of performance-based items in the sense that test specifications include not only the specification of the construct(s) to be measured but also the specification of the attributes of the test instruments that are to measure the construct(s)."

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Education

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Liu, Hui-Chun

Contributors dc:contributor

• Davidson, Fred

• Education, Reading

Language dc:language

eng

Identifier

(MiAaPQ)AAI9737179

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/80178