Signal Detection-Based Analyses of Stimulus Control

项目3将重点关注刺激控制发展中尚未完全理解的刺激变量和动机变量之间的相互作用。利用歧视学习中的经典范式，我们的研究将调查变量和变量相互作用，这些变量和变量相互作用可能有助于解释智力残疾者经常观察到的缓慢和/或不一致的学习结果。鉴于这种缺陷历来被归因于注意力不集中、过于宽泛和/或狭隘的关注和/或缺乏动机，我们的假设是，长期的歧视学习失败可能发生在那些按照传统标准被归类为注意力集中和动机良好的个体身上。我们认为，当教师/实验者从学习者的角度规划不可区分的偶然性时，就会导致长期失败。该项目的灵感来自于我们实验室的许多发现，以及当前基于信号检测理论（SDT）的学习过程的定量理论分析。这种定量方法在残疾研究中几乎没有被研究过，但我们相信它们至少有两个显著的优势。首先，它们提供了一种高度结构化的、经过经验验证的方法，用于分析程序化偶发事件的关键要素可辨别的程度。其次，有有效的定量技术来评估程度

Project 3 will focus on as yet incompletely understood interactions between stimulus variables and motivational variables in the development of stimulus control. Making use of classic paradigms in discrimination learning, our research will investigate variables and variable interactions that may help to account for the slow and/or inconsistent learning outcomes that are frequently observed in persons with intellectual disabilities. Whereas such deficits have historically been attributed to inattention, overly broad and/or narrow attending, and/or lack of motivation, our hypothesis is that protracted discrimination learning failures can occur in individuals who would be classified as both attentive and well-motivated by traditional standards. We propose that protracted failures result when teachers/experimenters program contingencies that are indiscriminable from the perspective of the learner. The proposed project is inspired by numerous findings from our laboratories and by current quantitative theoretical analyses of learning processes that have their roots in signal detection theory (SDT). Such quantitative approaches are virtually unstudied in disability research, but we believe that they have at least two significant advantages. First, they offer a highly structured, empirically verified method for analyzing the degree to which cdtical elements of programmed contingencies are discriminable. Second, there are effective quantitative techniques for assessing the degree to which learning failures are the result of biases due to stimulus and/or reinforcer variables. Together, we believe that these two features of SDT-based analyses will help to shed light on certain longstanding problems in the analysis of learning processes of individuals with intellectual disabilities. We suggest also that a quantitative SDT-based methodology has natural advantages for supporting more effective teaching interactions for such individuals. To demonstrate these advantages, we will employ another quantitative method, the percentile-scheduling approach. By combining SDT-based analyses with percentile approaches, our goal is to establish a foundation for effective computer-based management of discrimination learning procedures of the type that are now used in educational intervention for persons with disabilities.