Abstract
The mirror effect refers to the expectation that a rise in hit rates (e.g., through strengthening of memory traces) is mirrored by a fall in false alarm rates and a fall in hit rates is mirrored by a rise in false alarms. Currently, many models of memory predict the mirror effect, but not all measures of recognition memory show this pattern. Typically, the method used to increase hit rates in recognition memory experiments is to strengthen entire lists of items (e.g., through repetition, increased exposure duration, or a levels-of-processing manipulation). However, laboratory manipulations that strengthen entire lists may exaggerate list strength, consequently misrepresenting rates of mirror effects across different studies, because realistically, memory for some items in a list (e.g., category) is stronger than others (e.g., memory for my house is stronger than my neighbor’s house). Here, we aim to see if the mirror effect is observed in a recognition-induced forgetting paradigm when the list with increased hit rates is comprised of mixed-strength items, mimicking real-world memory, rather than all strong items. Across three experiments we manipulated list-strength through repetition, exposure duration, and a judgement of learning task. We found no evidence of a mirror effect, suggesting that the mirror effect may only be observed in the laboratory. The increased hit rates due to experimental manipulations without a corresponding decrease in false alarms may indicate a unique benefit for objects held in memory.