A Cognitive Model of the Implicit Associations Test

在过去的二十年里，对内隐、自动态度的研究一直是社会心理学的一个主导主题。 虽然人们对这种态度的原因和后果了解很多，但对内隐态度判断背后的特定认知过程却知之甚少。 本研究的主要目的是建立内隐联想测验（IAT：Greenwald，McGhee，Schwartz，1998）的认知机制模型。 自六年前发展以来，已经有超过100篇文章使用IAT来测试各种内隐态度。 尽管它的流行，IAT分数的含义仍然存在争议。 人们对驱动内隐态度的认知过程知之甚少，对内隐联想测验具体评估的内容更是知之甚少。 为了映射IAT背后的认知过程，本项目将采用一个模型，该模型对处理进行明确的假设，并进行明确的定量预测：扩散模型（例如，拉特克利夫，1978年;拉特克利夫Rouder，1998年）。 扩散模型是一种任务决策模型，其中参与者快速做出二元选择（如IAT）。 快速决策的显式模型可以揭示一些有趣的替代解释的决策过程，可以照亮我们的理解的任务，旨在测量内隐态度。 在一系列的IAT研究中，IAT的不同因素将被操纵，以测试IAT效应背后的不同可能的认知过程。 为了做到这一点，将采用标准的IAT程序，并在此范式内的各种因素将被操纵，以测试在不同条件下的模型。 通过将定量建模工具应用于IAT，本研究可以更好地揭示一个人进行IAT时的认知过程。 这样做可以达到三个目标：（1）提高我们对参与IAT时所使用的认知过程的理解;（2）使用扩散模型来解释IAT数据，并测试不同版本的模型，这些模型锚定在关于IAT效应位点的理论中;（3）建立IAT中不同参数操作的数据集，以便未来的研究者能够根据他们希望探索的认知过程更有策略地使用IAT。通过有效地模拟IAT，科学家将能够更好地使用这种流行的方法更适当和有效。 随着越来越多的科学家对内隐联想的干预应用感兴趣，如了解与心理功能障碍有关的态度、消费者行为等，更好地了解无意识对态度和行为的贡献可以为重要的社会问题提供更深入的见解，如对种族歧视的态度。

The study of implicit, automatic attitudes has been a dominant theme in social psychology over the last two decades. While a great deal is known about the causes and consequences of such attitudes, very little is known about the specific cognitive processes underlying implicit attitude judgments. The major aim of this research is to model the cognitive mechanisms underlying one of the most widely used paradigms to test implicit attitudes: the implicit associations test (IAT: Greenwald, McGhee, & Schwartz, 1998). Since its development six years ago, there have been over 100 articles published using the IAT to test a variety of implicit attitudes. Despite its popularity, the meaning of IAT scores remains controversial. Little is known about the cognitive processes driving implicit attitudes, and even less is known about what exactly the IAT in particular assesses. In order to map the cognitive processes underlying the IAT, this project will employ a model that makes explicit assumptions about processing, and explicit quantitative predictions: the diffusion model (e.g., Ratcliff, 1978; Ratcliff & Rouder, 1998). The diffusion model is a model of decision making in tasks wherein participants make fast binary choices (like in the IAT). Explicit models of fast decision making can reveal a number of interesting alternative explanations for decision processes that could illuminate our understanding of tasks that intend to measure implicit attitudes. In a series of IAT studies, different factors of the IAT will be manipulated in order to test different possible cognitive processes underlying IAT effects. To do this, a standard IAT procedure will be employed, and various factors within this paradigm will be manipulated to test the model under different conditions. By applying a quantitative modeling tool to the IAT, this research can better reveal the cognitive processes that are at play when a person takes an IAT. In so doing, three goals can be attained: (1) to advance our field's understanding of the cognitive processes used when participating in the IAT; (2) to use the diffusion model to account for IAT data, and to test different versions of the model that are anchored in theories about the locus of IAT effects; and (3) to establish a data set of different parametric manipulations in the IAT so that future researchers can use the IAT more strategically depending on the cognitive processes they wish to explore. By effectively modeling the IAT, scientists would be better able to use this popular methodology more appropriately and effectively. This is especially urgent given that more and more scientists are interested in exploring intervention applications of the IAT, such as in understanding attitudes related to psychological dysfunctions, consumer behavior, and so on. A better understanding of unconscious contributions to attitudes and behavior could provide greater insight into important social issues, such as attitudes toward racial discrimination.