ERIC Number: ED236165
Record Type: Non-Journal
Publication Date: 1983-Apr
Reference Count: N/A
A Comparative Analysis of Two Order Analytic Techniques: Assessing Item Hierarchies in Real and Simulated Data.
Chevalaz, Gerard M.; Tatsuoka, Kikumi K.
Two order theoretic techniques were presented and compared. Ordering theory of Krus and Bart (1974) and an extended Takeya's item relational structure analysis (IRS) by Tatsuoka and Tatsuoka (1981) were used to extract the hierarchical item structure from three datasets. Directed graphs were constructed and both methods were assessed as to how well they reproduced the theoretical structure of the data. It was discovered that the Krus and Bart (1974) procedure more adequately represented the complex interrelationships among test data than did the extended IRS method. Simulated data were found to present many problems and to be inappropriate for research in this area. Research in this area should include a large scale sampling distribution study to determine the distribution properties of simulated data. A more sophisticated method of generating binary responses which accounts for the distribution of theta needs to be developed. Also, a significance test and possibly a test of the differences between two item characteristic curves should be investigated. (Author/HFG)
Publication Type: Reports - Research
Education Level: N/A
Sponsor: National Inst. of Education (ED), Washington, DC.
Authoring Institution: Illinois Univ., Urbana. Computer-Based Education Research Lab.