Smith Andrew R, Windschitl Paul D
Department of Psychology, Appalachian State University, Boone, NC, 28608, USA.
Department of Psychology, University of Iowa, Iowa City, IA, USA.
Mem Cognit. 2015 Oct;43(7):1071-84. doi: 10.3758/s13421-015-0524-4.
The biasing influence of anchors on numerical estimates is well established, but the relationship between knowledge level and the susceptibility to anchoring effects is less clear. In two studies, we addressed the potential mitigating effects of having knowledge in a domain on vulnerability to anchoring effects in that domain. Of critical interest was a distinction between two forms of knowledge-metric and mapping knowledge. In Study 1, participants who had studied question-relevant information-that is, high-knowledge participants-were less influenced by anchors than were participants who had studied irrelevant information. The results from knowledge measures suggested that the reduction in anchoring was tied to increases in metric rather than mapping knowledge. In Study 2, participants studied information specifically designed to influence different types of knowledge. As we predicted, increases in metric knowledge-and not mapping knowledge-led to reduced anchoring effects. Implications for debiasing anchoring effects are discussed.
锚定对数值估计的偏差影响已得到充分证实，但知识水平与易受锚定效应影响之间的关系尚不清楚。在两项研究中，我们探讨了某一领域的知识对该领域易受锚定效应影响的潜在缓解作用。至关重要的是区分两种形式的知识——度量知识和映射知识。在研究1中，研究过与问题相关信息的参与者（即知识水平高的参与者）比研究过无关信息的参与者受锚定的影响更小。知识测量的结果表明，锚定效应的降低与度量知识的增加而非映射知识的增加有关。在研究2中，参与者研究了专门设计用来影响不同类型知识的信息。正如我们所预测的，度量知识的增加而非映射知识的增加导致了锚定效应的降低。我们还讨论了消除锚定效应偏差的意义。
