Doctoral Dissertation Research in Political Science: The Effect of Polls on Political Behavior

主要研究者：Jon Krosnick和Neil malhotra题目：博士论文研究：民意调查对政治行为的影响摘要选举是美国民主的核心特征。它们为公民提供了让现任者对其表现负责的能力，也为确保政府代表公众对政策问题的看法提供了一种手段。新闻媒体在美国的选举中发挥着重要作用，经常向公民传达有关候选人的信息，帮助选民做出明智、知情的决定。然而，媒体提供的信息有时可能妨碍人们的决策，妨碍他们作出最符合国家利益和最能反映其个人偏好的最佳投票选择的能力。此外，新闻媒体的报道可能会激励人们在选举中投票并参与民主进程，但媒体也可能使公民远离政治并阻碍投票率。本项目将探讨美国新闻媒体如何影响选民，并确定这种影响是建设性的还是有害的。本文的重点是媒体提供给公民的一种特殊类型的信息：记录候选人和公共政策受欢迎程度的民意调查结果。这种民意调查经常由研究机构进行，并由新闻媒介在重大的全国竞选期间定期报道。他们不断向公民更新有关竞选状态的信息，可能会影响人们对竞争候选人的评价。通常，这些民意调查甚至在政党的提名人被选出之前就进行了，让政治精英和公众在竞选正式开始之前就对候选人的潜在当选能力有了一种感觉。事实上，民意调查正日益成为专家们所称的“永久性竞选”的核心特征。从学术的角度来看，这些民意调查对公民政治态度和行为的潜在影响是有趣的，但从规范的角度来看，也可能是令人痛苦的。个人可能不会对自己的政治选择进行内省和仔细思考，而可能会遵从民意调查中其他人表达的意见。人们这样做可能是因为他们觉得自己可以从别人的集体意见中学习，从而做出明智的决定。或者，人们可能会感到遵从他人意见的压力，以避免被或感觉被排斥。此外，得知一场竞选可能不会势均力敌，可能会使人们不愿投票，或者尤其可能使“处于劣势”的候选人的支持者退缩。这些影响可能在某些人口中特别普遍，例如受教育程度最低和政治意识最低的人。这个项目将测试一系列这样的相关假设，这些假设概述在一个精心设计的理论模型中，这些理论模型将通过嵌入在互联网上进行的美国成年人口调查中的实验来评估。以前的研究已经探讨了上面讨论的一些问题，但过去的研究没有像本项目那样将这些想法整合到一个统一的、全面的理论模型中，过去的研究也没有以方法论上严格的方式测试过如此广泛的假设。这项拟议中的研究将向受访者提出一系列关于希拉里·克林顿和约翰·麦凯恩参议员之间假想的2008年总统竞选的问题。调查对象将获得有关竞争激烈程度的不同信息，以便研究这些信息如何影响候选人的选择和参与决定。这些数据可以检验公民在形成自己的政治态度和行为意图时，是否受到媒体对民意调查结果报道的影响，因为公民要么寻求集体智慧，要么有从众的愿望。无论哪种情况，调查结果都将对选举的学术研究以及人们对他们在民主政体中作为行动者的角色的理解产生重要影响。这个项目的发现为社会提供了几个潜在的好处。最重要的是，美国政府内外的许多人都有兴趣了解民主是如何运作的，以及如何最好地使民主公民具备履行职责的能力。了解美国民主是如何运作的，将有助于在其他国家设计和实施民主政府。洞察媒体如何影响美国的政治结果将有助于这个社会更好地了解自己，以便未来的媒体报道计划可以认识到它对选民和选举结果的影响。该项目还将加强学术社区的教学、培训和学习，允许博士生在执行项目时受到美国选举、公众舆论和政治心理学领域专家的密切指导。此外，本科生将协助进行研究，从而了解研究过程。这些调查实验将吸引来自不同种族和地理背景的受访者参与政治进程，有可能激励他们更多地了解美国政府。这个项目的研究结果将有助于深入了解代表性不足的群体是否特别受到新闻媒体的影响。研究结果还将改进政府、企业和学术界的民意研究界所使用的工具和方法，因为该项目将帮助调查方法学家更好地了解其工作的后果。

Primary Investigators: Jon Krosnick and Neil MalhotraTitle: Doctoral Dissertation Research: The Effect of Polls on Political Behavior 0720444Abstract Elections are a central feature of American democracy. They provide citizens with the ability to hold incumbents accountable for their performance, as well as a means to ensure that the public's views on policy issues are represented by the government. The news media play an important role in the conduct of elections in the United States, often conveying information to citizens about candidates in ways that can assist the electorate in making intelligent, informed decisions. However, information provided by the media may sometimes handicap people's decision making, hindering their ability to make optimal vote choices that are in the best interest of the country and are most reflective of their personal preferences. Additionally, news media reports may inspire people to vote in elections and participate in the democratic process, but the media can also turn the citizenry off from politics and discourage turnout. This project will explore how the American news media influence voters and ascertain whether that influence is constructive or harmful. The focus of this dissertation is on one particular type of information that the media provide to citizens: the results of polls documenting the popularity of candidates and public policies. Such polls are frequently conducted by research organizations and reported regularly by the news media during major national campaigns. They constantly update citizens about the state of a race, potentially influencing people's evaluations of competing candidates. Often, these polls are conducted even before the parties' nominees are selected, providing both political elites and the public a sense of the potential electability of contenders even before the campaign has officially commenced. Indeed, public opinion polling is becoming an increasingly central feature of what pundits have referred to as "the permanent campaign." The potential effects of these polls on citizens' political attitudes and behaviors are interesting from a scholarly perspective but may also be distressing in a normative sense. Individuals may not engage in introspection and careful thought about their political choices and may instead conform to the opinions expressed by others in polls. People may do so because they feel they can learn from the collective opinion of others in order to make wise decisions. Or people may feel pressure to conform to the opinions of others to avoid being or feeling ostracized. Additionally, learning that a race is likely not to be close may discourage people from voting, or may especially demobilize supporters of the "underdog" candidate. These effects may be particularly prevalent among some members of the population, such as the least educated and the least politically aware. This project will test a series of such related hypotheses outlined in an elaborate theoretical model, which will be evaluated via experiments embedded in surveys of the U.S. adult population conducted over the Internet. Previous research has explored some of the issues discussed above, but no past study has integrated these ideas within a unified, comprehensive theoretical model as this project will, nor has past research tested such a broad array of hypotheses in as methodologically rigorous a fashion. The proposed study will ask respondents a series of questions about a hypothetical 2008 presidential race between Senators Hillary Clinton and John McCain. Respondents will be given varying information about the closeness of the race, allowing examination of how such information affects candidate choice and the decision to participate. The data may test whether citizens are influenced by media reports of poll results when formulating their own political attitudes and behavioral intentions, either because citizens seek collective wisdom or have a desire to conform. In either case, the findings will have important implications for the academic study of elections and also for people's understandings of their roles as actors within the democratic polity. The findings from this project offer several potential benefits to society. Most importantly, many people in the U.S. government and outside of it are interested in understanding how democracies operate and how best to equip democratic citizens to carry out their responsibilities. Understanding how American democracy operates will help in the design and implementation of democratic governments in other nations. And insight into how the media influence American political outcomes will help this society to better understand itself, so that future planning of media coverage can be done cognizant of its impact on voters and election outcomes. This project will also enhance teaching, training, and learning within the scholarly community by allowing a PhD student to be closely mentored by a faculty member expert in the fields of American elections, public opinion, and political psychology while carrying out the project. Also, undergraduates will assist in the conduct of the research and will thereby learn about the research process. The survey experiments will engage respondents who are drawn from diverse ethnic and geographic backgrounds in the political process, potentially motivating them to learn more about American government. And findings from this project will yield insight into whether under-represented groups are especially influenced by the news media. The findings will also improve the instruments and methods used by the public opinion research community in government, business, and academia, because this project will help survey methodologists to better understand the consequences of their work.