Sidman M
New England Center for Autism, Southborough, Massachusetts 01772.
J Exp Anal Behav. 1992 Jul;58(1):173-82. doi: 10.1901/jeab.1992.58-173.
In a conditional discrimination procedure, samples appeared in a center key, and comparisons appeared in two of four outer keys. The location of comparison keys varied from trial to trial. Separate learning curves for each of the six possible pairs of comparison keys were plotted in a signal-detection space, revealing different patterns of progress on each pair. Also, when learned conditional discriminations were disrupted, pairs of keys differed in their patterns of disruption. Varying the location of comparison stimuli among six different pairs of keys had not eliminated key position as a controlling aspect of the stimuli. The variations simply increased the number of stimulus compounds--key position and experimental stimuli--that the subject learned. Plotting conditional-discrimination learning curves in a signal-detection space reveals relations among hits, false alarms, accuracy, and comparison preference that help to define a subject's progress.
在条件辨别程序中，样本出现在中央按键上，比较刺激出现在四个外部按键中的两个上。比较按键的位置在每次试验中都不同。在信号检测空间中绘制了六种可能的比较按键对中每一对的单独学习曲线，揭示了每一对上不同的进展模式。此外，当习得的条件辨别被打乱时，按键对在其打乱模式上存在差异。在六种不同的按键对之间改变比较刺激的位置并没有消除按键位置作为刺激控制因素的影响。这些变化只是增加了被试学习的刺激组合——按键位置和实验刺激——的数量。在信号检测空间中绘制条件辨别学习曲线揭示了命中、误报、准确性和比较偏好之间的关系，这些关系有助于定义被试的进展情况。
