Context reinstatement in recognition: Memory and beyond

• In a recognition test, context familiarity is attributed to a test item.
• The exact study context reinstated at test may also aid recollection.
• We show that context familiarity affects confidence for recognition decisions.
• Changes in confidence are linked to volunteer/withhold decisions.
• Reinstated contexts further boost confidence and reduce response withholding.

Context effects in recognition tests are twofold. First, presenting familiar contexts at a test leads to an attribution of context familiarity to a recognition probe, which has been dubbed ‘context-dependent recognition’. Second, reinstating the exact study context for a particular target in a recognition test cues recollection of an item-context association, resulting in ‘context-dependent discrimination’. Here we investigated how these two context effects are expressed in metacognitive monitoring (confidence judgments) and metacognitive control (‘don’t know’ responding) of retrieval. We used faces as studied items, landscape photographs as study and test contexts and both free- and forced-report 2AFC recognition tests. In terms of context-dependent recognition, the results document that presenting familiar contexts at test leads to higher confidence and lower rates of ‘don’t know’ responses compared to novel contexts, while having no effect on forced-report recognition accuracy. In terms of context-dependent discrimination, the results show that reinstated contexts further boost confidence and reduce ‘don’t know’ responding compared to familiar contexts, while affecting forced-report recognition accuracy only when contribution of recollection to recognition performance is high. Together, our results demonstrate that metacognitive measures are sensitive to context effects, sometimes even more so than recognition measures.