PDF pending restoration
ERIC Number: ED170328
Record Type: RIE
Publication Date: 1978-Nov
Reference Count: 0
Estimating Item Characteristic Curves. Final Report for Period March 1978 through September 1978.
Ree, Malcolm James
Item characteristic curve (ICC) theory describes the relationship between the ability of individuals and the probability of their answering a test question correctly; it is useful in estimating test scores, equating the scores of various tests, and scoring responses during adaptive testing. A simulation study of the effectiveness of the following four computer programs for estimating ICC was conducted: ANCILLES, OGIVIA, LOGIST, and simple transformation to the item-test biserial correlation. Using the three-parameter logistic model, which relates item responses to subjects' ability, three groups of 2,000 simulated subjects were administered 80-item tests. These simulated item responses were then calibrated using the four programs. The estimated item parameters were compared to the known item parameters in four analyses for each program in all of the three data sets--each set representing a different level of ability. Results indicated that no one program functioned best in all situations posed by the data sets. In conclusion, the choice of an ICC estimation program should be based on the planned use of the estimates, the characteristics of the distribution of ability in the groups available for item administration, the necessity to calibrate all items, and the computer resources available. (Author/CP)
Publication Type: Reports - Research
Education Level: N/A
Authoring Institution: Air Force Human Resources Lab., Brooks AFB, TX.