ERIC Number: ED468759
Record Type: Non-Journal
Publication Date: 1993-Jun
Reference Count: N/A
The Effect of COMC on the Stability of LSAT Item Parameter Estimates. LSAC Research Report Series.
Reese, Lynda; McKinley, Robert
In item calibration using LOGIST (M. Wingersky, R. Patrick, and F. Lord, 1987), when the program determines that it cannot accurately estimate the c-parameter for a particular item due to insufficient information at the lower levels of ability, an estimate of the c-parameters, called COMC, is obtained by combining all such items. The purpose of this study was to evaluate the effect of using COMC on the stability of item response theory (IRT) parameter estimates. In this study, two Reading Comprehension pretest sections of the Law School Admission Test (LSAT) were administered twice, with samples of 1,00 and 1,030 for the June administration and 1,625 and 1,610 in December. and new parameter estimates were obtained for each administration. Different items were set to COMC for the two administrations due to sample fluctuations. In comparing the item characteristic curves for the different patterns of c-estimation, it seems that the most discrepant item characteristic curves are produced when the c-parameter is fixed for only one of the administrations. When the c-parameter is fixed for both administrations, item characteristic curves are more stable. However, results show that the use of COMC is not an acceptable procedure for dealing with the instability of the c-parameter. A means of deriving a more reasonable estimate of the c-parameter should be identified. (Contains 2 tables, 11 figures, and 2 references.) (SLD)
Publication Type: Numerical/Quantitative Data; Reports - Research
Education Level: N/A
Authoring Institution: Law School Admission Council, Newtown, PA.
Identifiers - Assessments and Surveys: Law School Admission Test