Doctoral Dissertation Research: Ballot Structure and Voting Decisions

本项目侧重于选票设计和投票程序在影响投票站选民作出的决定以及由此影响选举结果方面所起的作用。PI认为，向选民展示选举方案的不同模式对于理解人们如何投票，以及某些候选人或政党在不同类型的投票中获得不成比例的优势或劣势的程度至关重要。为这个项目选定的国家（阿根廷、智利、厄瓜多尔、萨尔瓦多和洪都拉斯）提供了机会来衡量由于投票结构的变化而影响代表权的各种结果。提供选举产品的这种变化可能对某些政党有利，抵消妇女当选的可能性，和/或影响在正式选票上的候选人的照片中观察到的个人属性是否影响当选的可能性。研究选票的结果及其对选举结果的影响有助于解释人们如何投票以及投票给谁，从而揭示代表性的不同方面，例如少数群体获得立法席位的机会以及选票提供的信息线索的有效性和相关性。该项目的成果将为考虑或参与选票设计改革的决策者和公民提供宝贵的信息，了解改革对代表性和问责制的潜在影响，特别是对特定政治制度中选举民主质量的潜在影响。本项目通过解决选票设计和投票程序对代表性的作用，有助于加深对选举过程的理解。本文采用了两种不同的分析策略来评估向选民提供选举产品的方式如何影响选民行为和选举结果。首先，该项目使用观察数据来评估不同的选票设计是否会导致不同的选举结果。其次，考虑到因果机制在个人层面上运作，选民的行为可能受到无数因素的影响，我们采用实验研究设计来评估选民在实际选举中如何判断候选人的特征。对于第一个策略，PI使用了来自阿根廷各省的数据，这些省份已经用不同版本的澳大利亚选票取代了传统的党派纸质选票。她使用准实验研究设计分析了选举提供的不同展示如何使某些政党受益。根据1992年至2008年的智利市政数据，PI研究了代表人数不足群体的候选人在公开名单多成员比例代表制中按顺序当选的可能性。对于第二种策略，她在三个开放名单制度国家（萨尔瓦多、厄瓜多尔和洪都拉斯）进行了实验，采用了panachage（选民有机会将他们的多张选票分配给来自多个成员区的不同政党的候选人）。实验评估候选人的个人属性是否会影响当选的概率，以及投票站提供的上下文信息如何通过提供额外的工具来解码信息线索，改变选民对候选人的评价和偏好。

General Summary This project focuses on the role played by ballot design and vote casting procedures in shaping the decisions made by voters in the polling booth and, therein, electoral outcomes. The PI argues that the different modes through which electoral alternatives are presented to voters are critical to understanding how people vote, and the extent to which certain candidates or parties are disproportionately advantaged or disadvantaged by different types of ballots. The countries selected for this project (Argentina, Chile, Ecuador, El Salvador and Honduras) offer the opportunity to measure a variety of outcomes affecting representation as a result of variations in the ballot structure. Such variations in presenting the electoral offering may benefit certain parties, offset the probability of women being elected, and/or affect whether candidates' personal attributes observed in the picture of them included on the official ballot affect the probability of being elected. Studying the consequences of ballots and their features on electoral outcomes can help to explain how and who people vote for, shedding light on different aspects of representation, such as minority groups' access to legislative seats and the effectiveness and relevance of informational cues provided by the ballot. The results of this project will provide policymakers and citizens considering or engaged in ballot design reform with invaluable information regarding the potential impact of reforms on representation and accountability in particular and the quality of electoral democracy in a given political system more generally.Technical SummaryThis project contributes to a deeper understanding of the electoral process by addressing the role of ballot design and vote casting procedures on representation. Two different analytic strategies are applied to evaluate how the way the electoral offering is presented to the voter affects voter behavior and electoral outcomes. First, the project uses observational data to assess whether different ballot designs lead to different electoral outcomes. Second, given that the causal mechanism operates at an individual level, and that voter behavior can be influenced by a myriad of factors, an experimental research design is used to assess how voters may judge candidates' traits in actual elections. For the first strategy, the PI uses data from Argentina provinces that have replaced the traditional partisan paper ballot with different versions of the Australian ballot. She analyzes how different presentations of the electoral offering benefit certain parties using a quasi-experimental research design. Using Chilean municipal data from 1992 to 2008, the PI examines the probabilities of candidates from underrepresented groups being elected given their order placement on the ballot in open list multi-member proportional representation systems. For the second strategy, she conducts experiments in three open list system countries (El Salvador, Ecuador and Honduras) that employ panachage (where voters have the opportunity to distribute their multiple votes among candidates from different parties in multimember districts). The experiments assess whether candidates' personal attributes affect the probability of being elected and how contextual information provided in the voting booth alters voters' evaluation of and preference for candidates by supplying additional tools to decode informational cues.