Guntre Raymond W, Bodner Glen E, Azad Tanjeem
Department of Psychology, University of Calgary, Alberta, Canada.
Mem Cognit. 2007 Jul;35(5):1083-92. doi: 10.3758/bf03193480.
Encoding tasks that increase memory accuracy are appealing from both practical and theoretical perspectives. Within the Deese/Roediger-McDermott (DRM) paradigm, we found that generating list words from anagrams (relative to reading) produced a mirror effect: enhanced recognition of studied words coupled with a reduction in false recognition. Signal detection analyses suggest that the increase in correct recognition was due to enhanced item-specific encoding of the list words, whereas the reduction in false recognition was due to enhanced strategic monitoring at test (i.e., a distinctiveness heuristic), rather than to reduced relational encoding at study. Further support for a distinctiveness heuristic account was obtained using both "theme judgment" instructions and within-group conditions. In our final experiment, we replicated this mirror effect using a purely mnemonic (self-referential) encoding task, showing that extra perceptual cues are not necessary to induce participants to adopt a successful memory-improvement strategy at test.
从实践和理论角度来看，能够提高记忆准确性的编码任务都很有吸引力。在迪斯/罗迪格 - 麦克德莫特（DRM）范式中，我们发现从变位词生成列表单词（相对于阅读）会产生一种镜像效应：对所学单词的识别增强，同时错误识别减少。信号检测分析表明，正确识别的增加是由于列表单词的项目特定编码增强，而错误识别的减少是由于测试时策略性监控的增强（即独特性启发式），而不是由于学习时关系编码的减少。使用“主题判断”指令和组内条件都获得了对独特性启发式解释的进一步支持。在我们的最后一个实验中，我们使用纯记忆（自我参照）编码任务复制了这种镜像效应，表明在测试时不需要额外的感知线索来促使参与者采用成功的记忆改进策略。
