Age differences in reported recollective experience are due to encoding effects, not response bias.

Three experiments examined whether reduced recollective experience reported in old age is due to a criterion shift towards more cautious responses by older subjects. In Experiment 1 Young and Old subjects took a recognition test without specific instructions on how they should encode the presented words. For recognised items subjects indicated whether they recollected the item's previous occurrence or whether they just knew it had been on the list. They then rated their confidence that the word came from the study list. Although overall recognition levels were equivalent, older adults recollected less and reported more know responses than the younger subjects. However, there was no overall difference in confidence, contrary to a criterion shift explanation. In Experiment 2A specific encoding instructions removed the age-related change in recollective experience entirely. Experiment 2B reproduced the test conditions of Experiment 2A, but without specific encoding instructions, and replicated the pattern of know responding found in Experiment 1. Thus the three experiments together suggest that the amount of recollection experienced by the elderly is not explicable in terms of cautiousness, but is driven by the encoding carried out by the elderly at presentation.