ABSTRACT
The emergence of cognitive psychology has greatly reduced the separation between the two disciplines of scientific psychology, experimental and correlational, described by Cronbach in 1957. Cognitive psychologists have developed experimental methods to analyze the knowledge and mental processes that underlie performance on many types of items from achievement and aptitude tests (see Pellegrino & Glaser, 1979; Snow & Lohman, 1989). They also have explored which processing abilities are differentially related to levels of performance on tests. Thus the links between performance on assessment tasks and individual variation in knowledge and skills have been more clearly articulated. In turn, this type of analysis can inform item generation as illustrated in many of the chapters in this volume. However, some gaps that remain between experimental and correlational approaches are differences between depth and breath of focus, between our ability to build generative models for small sets of items, and our desire to include varied and unique items in assessments. Tensions exist between cognitive modeling of narrow sets of problems and broadly conceived assessment frameworks, between attention to a part of a system and appreciation of the complexity of the system as a whole.
