投票中的锚定鲁棒性
Anchor-proofness in Voting
AI总结：
本文研究了投票中锚定偏差对集体决策的影响，发现只有极限制的投票规则才能保证结果不受呈现顺序影响。
摘要：
本文为经济理论中的基础问题做出了贡献：个体层面的认知偏差如何与集体选择机制相互作用？我们研究了一个设定，其中选民对一组替代方案持有内在偏好排序，但通过批准投票来决定集体结果。选票受锚定偏差影响：替代方案按顺序由社会规划者呈现，选民只在某个替代方案可接受且严格优于之前遇到的所有替代方案时才批准它。我们首先分析哪些基于批准的投票规则是锚定鲁棒的，即无论呈现顺序如何，它们总能选出相同的结果。我们显示这一要求极为严格：只有非常限制性的规则才能满足。然后我们转向社会规划者可能产生的影响。在积极方面，当规划者不了解选民的内在偏好时，她无法操控结果。
英文摘要：
We study how order-dependent individual evaluations affect collective decision-making. We consider a voting environment in which voters have intrinsic preference rankings and acceptability thresholds, but form approval ballots sequentially as alternatives are presented by a social planner. Because alternatives encountered early on serve as anchors, approval ballots may depend on the presentation order even when intrinsic preferences remain unchanged. We ask when approval-based voting rules are robust to this order dependence, a property we call anchor-proofness. We first show that no non-constant approval-based voting rule is anchor-proof. We then study weaker forms of robustness and characterize anchor-proof profiles for the classical approval voting rule and the nomination rule. Finally, we examine the strategic role of the social planner under different information structures. We show that, while a planner with no information about voters' preferences cannot safely manipulate any rule satisfying weak and total unanimity, even limited information about voters' acceptability or plurality assessments can restore manipulation. Overall, we highlight presentation order as an important design feature of collective decision-procedures when voters are subject to anchoring.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
