The dual strategy model of reason: Integrating counterexample and probabilistic reasoning

推理涉及在给定前提的基础上进行推理，是高级认知的重要组成部分。不幸的是，研究表明，人们很难做到合乎逻辑，而且经常做出不合逻辑的推论。目前，人们如何推理有两个主要的模型，心理模型理论和概率理论，它们提出了非常不同的潜在过程和机制。然而，我实验室的最新工作表明，一种混合推理模型，即双重策略模型，可能会整合这些不同的方法。这个模型表明，这些理论中的每一个都可能代表一种不同形式的推理策略，并确定了两种这样的策略。反例推理者通过构造给定问题的关键方面的表示来进行推理，并且对潜在的反例很敏感。统计推理者使用对信息的直观访问，以便对可能的结论进行似然估计。一些研究使用了先前设计的诊断工具，表明人们更倾向于使用这两种推理策略中的一种。这预示着个体在各种各样的推理和判断任务上的差异。本计划将继续更详细地研究这一模型，以便提供更有力的经验基础，并更精确地确定哪些维度表征了战略差异。第一组研究将考察时间限制对推理的影响。先前的结果表明，统计推理者正在使用一种复杂的快速、直观的处理形式，这应该相对不受时间限制的影响，而反例推理者使用的是更密集的工作记忆过程，对于这种过程，快速思考应该比缓慢思考的逻辑性更低。第二个系列的研究来自一些初步结果，这些结果表明，反例推理者更关注有限的可用信息子集，而统计推理者则更分散，关注更大的信息阵列，但方式不那么集中。这将在一系列使用标准和非标准注意力衡量标准的相关研究中进行检验。第三个系列的研究将考察不同策略使用对儿童学习成绩的潜在影响。最近对成年人的一项研究表明，策略使用可以预测认知能力之外的表现。这表明，策略使用的差异可能是理解儿童和青少年处理学术信息的潜在方式的差异的重要工具。推理的双重策略模型表明，它利用了人们思维方式的一个主要区别，这种区别有可能改变我们对人类推理中的可变性的理解。拟议中的研究将更深入地探讨这种区别的认知、注意和发展成分。

Reasoning involves making inferences on the basis of given premises and is an important component of advanced cognition. Unfortunately, research has shown that people have real problems being "logical" and often make illogical inferences. Currently, there are two major models of how people reason, mental model theories and probabilistic theories, which suggest very different underlying processes and mechanisms. However, the latest work from my lab suggests that a hybrid model of reasoning, the Dual strategy model, could potentially integrate these various approaches. This model suggests that each of these theories might represent a different form of reasoning strategy, and identifies two such strategies. Counterexample reasoners make inferences by constructing representations of key aspects of a given problem and are sensitive to potential counterexamples. Statistical reasoners use intuitive access to information in order to generate likelihood estimates of possible conclusions. Several studies, using a previously elaborated diagnostic instrument, suggest that people preferentially use one of these two forms of reasoning strategy. This predicts individual differences on a wide variety of reasoning and judgment tasks. This program will continue to examine this model in more detail, in order to provide a stronger empirical base and to determine with more precision what dimensions characterize the strategy distinction. A first set of studies will examine effects of time constraints on reasoning. Previous results suggest that Statistical reasoners are using a complex form of rapid, intuitive processing, which should be relatively impervious to time constraints, while Counterexample reasoners use more working memory intensive processes for which thinking fast should be less logical than thinking slow. A second series of studies are derived from some preliminary results that suggests that Counterexample reasoners are more focused on a restricted subset of available information while Statistical reasoners are more diffuse and attend to a larger array of information, but in a less focused way. This will be examined in a set of correlational studies using both standard and non-standard measures of attention. A third series of studies will examine the potential implications of differing strategy use on academic performance in children. A recent study with adults has shown that strategy use predicts performance over and above cognitive capacity. This suggests that differences in strategy use could be important tool in understanding differences in the underlying way that children and adolescents process academic information. The Dual strategy model of reasoning has shown that it taps into a major distinction in the way that people think, one which has the potential to change our understanding of the variability in human reasoning. The proposed studies will delve deeper into the cognitive, attentional and developmental components of this distinction.