Abstract: People often rely on formal knowledge as well as experience about how memory works when evaluating the source of remembered information (Johnson, Raye, Foley, & Foley, 1981; Johnson, 1997). Such reliance may lead to the picture preference bias, a tendency to misattribute perceived words as pictures rather than misattributing perceived pictures as words in a word-picture monitoring task (Foley, Durso, Wilder, & Friedman, 1991; Riefer, Hu, & Batchelder, 1994; Foley, 1998). Previous research has suggested that picture preference bias in source monitoring may have played a role in the misinformation effect in eyewitness memory research.
In the present study, the effect of repeated exposure on the picture preference bias was examined in a standard source monitoring paradigm. A 2 (Item type: word vs. picture) × 2 (Exposure: once vs. twice) × 2 (Test: source monitoring vs. recognition) within-subjects design was employed. The participants studied 144 words or pictures either once or twice during the study phase. They then completed a mixed source monitoring and recognition test.
The results revealed a significant interaction between source misattribution type and exposure so that the picture preference bias emerged only after repeated exposure: The error of misattributing perceived words to pictures was significantly higher than that of misattributing pictures to words in the repeated exposure condition, but not in the one exposure condition.
Findings from the present study suggest that attempting to enhance eyewitness memory by repeatedly exposing the witnesses to the remnants of the original event could be counterproductive.

Key words: source monitoring, picture preference bias, eyewitness testimony, exposure times