Department of Psychology, Ohio University, Athens, OH, USA.
Consortium for the Advancement of Cognitive Science, Ohio University, Athens, OH, USA.
Mem Cognit. 2020 Oct;48(7):1089-1111. doi: 10.3758/s13421-020-01053-1.
We explored the nature of human informativeness judgments: namely, people's judgments about the quantity of information that object stimuli convey about the category of objects to which they belong. Informativeness judgments play a key role in everyday decision-making situations involving the selection of items from groups that best represent the "group as a whole." They also provide insight into the nature of prototype formation. We investigated informativeness judgments with an experiment involving 41 category structures - the most comprehensive and rigorous examination thus far. We assess the robustness and generalizability of the results from this experiment by examining the relationship between group-level and individual-level performance. In addition, we show that in most cases (and especially in those involving relatively lower dimensionality structures), these judgments are predicted more accurately and explained more satisfactorily by Representational Information Theory (Vigo in Information Sciences 181: 4847-4859, 2011 and in Information 4(1):1-30, 2012) and its simplest core model than by standard models of prototypicality. Finally, we argue that prototypicality models are special cases of the more general "representational information" framework.
即人们对物体刺激传达其所属类别信息量的判断。信息量判断在涉及从最佳代表“整体”的群体中选择项目的日常决策情境中起着关键作用。它们还提供了对原型形成本质的洞察。我们通过一项涉及 41 个类别结构的实验来研究信息量判断,这是迄今为止最全面和严格的检验。我们通过检查群体水平和个体水平表现之间的关系,来评估该实验结果的稳健性和普遍性。此外,我们表明,在大多数情况下(特别是在那些涉及相对低维结构的情况下),这些判断可以更准确地被表示信息理论(Vigo 在 Information Sciences 181: 4847-4859, 2011 和 Information 4(1):1-30, 2012)及其最简单的核心模型,而不是标准的原型模型来更好地预测和解释。最后,我们认为原型模型是更通用的“表示信息”框架的特例。