A variety of statistical and logical procedures can help us in identifying items that fail to function as intended in a typical population. Chapter 12 dealt with four major item analysis procedures/statistics : item difficulty (p i), item discrimination (D), distracter analysis, and item homogeneity. It also described how item descriptive statistics can be evaluated in relation to the total score statistics to improve item quality. Of the procedures in Chapter 12, pi, D, and distracter analysis, are best applied when developing written, structured-response items for achievement/ability tests. Descriptive item statistics and item homogeneity indices are also applicable in attitudinal or psychological scale-development contexts with structured, Likert-scale items.

Page 3