NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED469173
Record Type: Non-Journal
Publication Date: 1999-Aug
Pages: 13
Abstractor: N/A
Reference Count: N/A
ISBN: N/A
ISSN: N/A
Impact of Local Item Dependence on True-Score Equating. LSAC Research Report Series.
Reese, Lynda M.; Pashley, Peter J.
This study investigated the practical effects of local item dependence (LID) on item response theory (IRT) true-score equating. A scenario was defined that emulated the Law School Admission Test (LSAT) preequating model, and data were generated to assess the impact of different degrees of LID on final equating outcomes. An extreme amount of LID was induced for a dataset that was analyzed by applying the Rasch model, and a typical amount of LID was induced for a dataset that was analyzed by applying the three-parameter logistic model. Results indicate that the effects of the extreme LID on the conversion line derived after scaling was carried out were negligible. Effects on the observed score distribution were noted for the extreme LID case, however, and such effects would have implications for score reporting. For the typical amount of LID, both the equating conversion line and the observed score distribution were essentially unaffected. (Author/SLD)
Publication Type: Numerical/Quantitative Data; Reports - Research
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: Law School Admission Council, Newtown, PA.
Identifiers - Assessments and Surveys: Law School Admission Test