Estimating the frequency of nonevents: the role of recollection failure in false recognition.

Participants studied lists of multiply presented converging associates (e.g., bed, dream, pillow, etc.) and were timed as they estimated how often they saw list items, related foils (e.g., blanket), and nonpresented critical items (SLEEP). Average number of repetitions (few [3] vs. many [6]) and repetition variability (fixed vs. variable) were manipulated between subjects. Participants responded more slowly to critical items (3.18 sec) than to list items (2.45 sec) or foils (2.22 sec). In addition, critical-item judgments of frequency (JOFs) were about as large as list-item JOFs, and false recognition (i.e., nonzero JOFs) of critical items was most likely in the few-fixed condition (96%) and least likely in the many-fixed condition (74%). These findings suggest that people can use recollection failure--the absence of an anticipated recollective experience, coupled with strong familiarity--to distinguish critical items from list items and that recollection failure is weighted most heavily when people expect familiar probes to access episodic information.