Dixon L S
J Appl Behav Anal. 1981 Winter;14(4):465-78. doi: 10.1901/jaba.1981.14-465.
Matching-to-sample procedures were used to assess picture representation skills of severely retarded, nonverbal adolescents. Identity matching within the classes of objects and life-size, full-color photos of the objects was first used to assess visual discrimination, a necessary condition for picture representation. Picture representation was then assessed through photo-object matching tasks. Five students demonstrated visual discrimination (identity matching) within the two classes of photos and the objects. Only one student demonstrated photo-object matching. The results of the four students who failed to demonstrate photo-object matching suggested that physical properties of photos (flat, rectangular) and depth dimensions of objects may exert more control over matching than the similarities of the objects and images within the photos. An analysis of figure-ground variables was conducted to provide an empirical basis for program development in the use of pictures. In one series of tests, rectangular shape and background were removed by cutting out the figures in the photos. The edge shape of the photo and the edge shape of the image were then identical. The results suggest that photo-object matching may be facilitated by using cut-out figures rather than the complete rectangular photo.
匹配样本程序被用于评估重度智力迟钝、不会说话的青少年的图片表征能力。首先使用物体类别内的身份匹配以及物体的真人大小、全彩照片来评估视觉辨别能力，这是图片表征的一个必要条件。然后通过照片与物体匹配任务来评估图片表征能力。五名学生在两类照片和物体中表现出了视觉辨别能力（身份匹配）。只有一名学生表现出了照片与物体匹配能力。四名未能表现出照片与物体匹配能力的学生的结果表明，照片的物理属性（平面、矩形）和物体的深度维度对匹配的控制可能比照片中物体与图像的相似性更大。进行了图形-背景变量分析，为图片使用中的程序开发提供实证依据。在一系列测试中，通过裁剪照片中的图形来去除矩形形状和背景。此时照片的边缘形状与图像的边缘形状相同。结果表明，使用裁剪后的图形而非完整的矩形照片可能会促进照片与物体的匹配。
