True and false memory for colour names versus actual colours: Support for the visual distinctiveness heuristic in memory for colour information

Abstract

In a colour variation of the Deese–Roediger–McDermott (DRM) false memory paradigm, participants studied lists of words critically related to a nonstudied colour name (e.g., “blood, cherry, scarlet, rouge … ”); they later showed false memory for the critical colour name (e.g., “red”). Two additional experiments suggest that participants generate colour imagery in response to such colour-related DRM lists. First, participants claim to experience colour imagery more often following colour-related than standard non-colour-related DRM lists; they also rate their colour imagery as more vivid following colour-related lists. Second, participants exhibit facilitative priming for critical colours in a dot selection task that follows words in the colour-related DRM list, suggesting that colour-related DRM lists prime participants for the actual critical colours themselves. Despite these findings, false memory for critical colour names does not extend to the actual colours themselves (font colours). Rather than leading to source confusion about which colours were self-generated and which were studied, presenting the study lists in varied font colours actually worked to reduce false memory overall. Results are interpreted within the framework of the visual distinctiveness hypothesis.

Keywords:

• DRM
• False memory
• Visual Distinctiveness
• Colour imagery

Acknowledgments

The authors would like to thank Elizabeth Marsh, Barbie Huelser, Lisa Fazio, and the participants in the Cognitive Psychology and Perceptual & Brain Sciences brownbag at Colorado State University for useful comments on an earlier version of this manuscript. This research was supported by National Science Foundation (NSF) Grants 0349088 and 0638486 to Anne M. Cleary.

Notes

¹Due to a misprint, the word “food” was presented at test instead of “foot” in Conditions 1 and 2. Therefore, “food” was treated as a nonstudied word. This problem was corrected for Conditions 3 and 4, and “foot” was tested as a nonstudied word.

²Some might be concerned that presenting the critical colour more often than the comparison colours creates a potential confound (whereby facilitative priming effects for the critical colour may be due to perceptual priming from the higher frequency of the critical colour itself within the experiment). To address this concern, we carried out an additional experiment in which for standard (noncolour) DRM lists, a repeating “high-frequency” colour-filled dot appeared on approximately half of the trials for each word list while for the remaining trials, the colour for the colour-filled dot was randomly selected from among the remaining “low-frequency” colours. If the facilitative priming shown in Experiment 2b had been due to the repetition of the critical colour itself within the experiment, facilitative priming should have been shown for the “high-frequency” colours in the standard (noncolour) DRM lists. However, such facilitative priming was not shown. “High-frequency” colour trials yielded a mean RT of 308 ms (SD = 67.14), and “low-frequency” colour trials yielded a mean RT of 308 (SD = 66.71), t(17) = 0.18, SE = 2.76, p = .86. This suggests that the facilitative priming shown for critical colour dots in Experiment 2b was caused by the presentation of colour-related words from the colour DRM lists.

³Because of a misprint in the program, “fast” appeared twice on the recognition test as an associate for the “slow” list, and its second occurrence appeared in place of the noncritical associate for the “soft” list. This error occurred across all conditions (studied vs. nonstudied and black font vs. coloured font), and the computations of hit and false alarm rates were adjusted accordingly.