Department of Psychological Medicine, Institute of Psychiatry, London, UK.
Behav Res Methods. 2013 Jun;45(2):344-54. doi: 10.3758/s13428-012-0255-4.
Although many visual stimulus databases exist, none has data on item similarity levels for multiple items of each kind of stimulus. We present such data for 50 sets of grayscale object photographs. Similarity measures between pictures in each set (e.g., 25 different buttons) were collected using a similarity-sorting method (Goldstone, Behavior Research Methods Instruments & Computers, 26(4):381-386, 1994). A validation experiment used data from 1 picture set and compared responses from standard pairwise measures. This showed close agreement. The similarity-sorting measures were then standardized across picture sets, using pairwise ratings. Finally, the standardized similarity distances were validated in a recognition memory experiment; false alarms increased when targets and foils were more similar. These data will facilitate memory and perception research that needs to make comparisons between stimuli with a range of known target-foil similarities.
尽管存在许多视觉刺激物数据库，但没有一个数据库具有每种刺激物的多个项目的项目相似性水平的数据。我们为 50 组灰度物体照片提供了此类数据。使用相似性排序方法（Goldstone，行为研究方法仪器与计算机，26（4）：381-386，1994）收集了每组图片（例如 25 个不同的按钮）之间的相似性度量。验证实验使用了 1 组图片数据，并比较了来自标准成对度量的响应。这表明了密切的一致性。然后使用成对评分对图片集进行标准化相似性排序度量。最后，在识别记忆实验中验证了标准化的相似性距离；当目标和诱饵更相似时，误报增加。这些数据将促进需要在具有一系列已知目标-诱饵相似性的刺激物之间进行比较的记忆和感知研究。
