NotesFAQContact Us
Search Tips
ERIC Number: ED410291
Record Type: Non-Journal
Publication Date: 1997-Mar
Pages: 24
Abstractor: N/A
Reference Count: N/A
ISSN: ISSN-0892-9580
Raters and Single Prompt-to-Prompt Equating Using the FACETS Model in a Writing Performance Assessment.
Du, Yi; And Others
The FACETS equating model meets the complex requirements for equating writing performance assessment across both raters and prompts. This study is based on an equating of the 1996 writing performance assessment in the Minneapolis Public Schools (Minnesota). Raters and prompts were equated simultaneously using the FACETS model. About 3,000 fifth graders and 3,000 seventh graders participated in the writing assessment. Three prompts were assessed in each grade, and each student wrote to one of the prompts. About 30 raters were selected from Minneapolis Public Schools teachers to score the papers using a uniform rubric. An extension of the Rasch model to include multiple facets (FACETS model) was used in equating to determine the transformation rules. The four facets were student, item (scoring component), rater, and prompt. Overall results show that the FACETS model calibrates raters, students, topics, and scoring dimensions so that all facets are positioned on a common scale. The scale is in log-odds, or "logit," units that constitute an equal-interval scale with respect to appropriately transformed probabilities of responding in particular categories. The advantages of the FACETS model include: (1) sample independence; (2) calibration invariance; (3) equating more than one facet at the same time; and (4) flexibility in the sample size for examinees and items. (Contains 9 figures, 8 tables, and 13 references.) (SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A