|Title||A Comparison of Item Selection Techniques for Testlets|
|Publication Type||Journal Article|
|Year of Publication||2010|
|Authors||Murphy, DL, Dodd, BG, Vaughn, BK|
|Journal||Applied Psychological Measurement|
This study examined the performance of the maximum Fisher’s information, the maximum posterior weighted information, and the minimum expected posterior variance methods for selecting items in a computerized adaptive testing system when the items were grouped in testlets. A simulation study compared the efficiency of ability estimation among the item selection techniques under varying conditions of local-item dependency when the response model was either the three-parameter-logistic item response theory or the three-parameter-logistic testlet response theory. The item selection techniques performed similarly within any particular condition, the practical implications of which are discussed within the article.