Evaluating the validity of the simplified conjoint-recognition model.

The distinction between verbatim and gist traces has been central to various fields of cognitive research, especially false memory, judgment and decision-making, cognitive development, and aging. The conjoint-recognition paradigm was developed to quantify the contributions of verbatim and gist traces to various tasks in those fields. A simplified version of the original paradigm has been devised and has been implemented in several recent studies. However, the validity of this methodology and the memory processes it measures remain uncertain. To address the problem, we first conducted factor analyses at the level of experiment conditions and then a posterior factor analysis to determine whether the simplified paradigm decomposes verbatim and gist traces in the same manner as the full paradigm. Second, we evaluated the sensitivity and specificity of verbatim and gist parameters to a series of experimental manipulations to examine whether the simplified model satisfies convergent and discriminant validity criteria. Two conclusions emerged. On the one hand, the simplified conjoint-recognition paradigm is a reliable and valid method of separating the contributions of verbatim and gist memory in the tasks to which it has been applied. On the other hand, the specific test format increases the difficulty of recollection rejection.