Nonseparable Preferences in Public Opinion

9617679莱西一项关于大众政治行为的大量研究假设，个人有“可分离的偏好”。这一假设要求一个人在一个问题上的偏好不取决于对其他问题的偏好。这种假设是如此普遍，以至于民意调查通常会询问受访者对一个又一个问题的看法，而不是跨问题的联系。如果一个人希望联邦政府禁止堕胎，但前提是政府也增加了在性教育、公共日托和对有抚养子女的家庭的援助上的支出，那该怎么办？或者，如果一个人只希望减税，而削减的开支来自国防预算，而不是社会福利，那该怎么办？如果一个人对一个问题的看法取决于另一个问题的结果或可用选项，那么她的偏好是“不可分割的”。这个项目调查了公众舆论和投票行为中不可分离的偏好的程度和影响。小组实验帮助首席调查员制定新的调查问题，以确定不可分离的偏好对调查回答的影响。调查人员使用了新的调查问题，对俄亥俄州哥伦布市的400名居民进行了随机抽样的电话调查。这项调查的数据使社会科学家能够判断，不可分离的偏好是否能解释公众舆论中一些常见的经验性问题，包括问题顺序效应、偏好随着时间的推移不稳定，以及问题对选民选举选择的影响显然微乎其微。有了新的调查工具来捕捉不可分割的偏好，政治学家更有能力调查投票行为和民意，政策制定者对人们在重要公共政策问题上真正想要什么有了更清晰的看法。***

9617679 Lacy A substantial body of research on mass political behavior assumes that individuals have "separable preferences." This assumption requires that a person's preference on one issue does not depend on preference on other issues. The assumption is so prevalent that public opinion surveys typically ask a respondent her opinion on one issue after another without links across issues. What if a persons wants the federal government to prohibit abortions, but only if the government also increases spending on sexual education, public day care, and aid to families with dependent children? Or what if a person wants a tax cut only if spending cuts come from the defense budget, not from social welfare? If a person's opinion on one issue depends on the outcome of or the options available on another issue, her preferences are "nonseparable." This project investigates the extent and impact of nonseparable preferences in public opinion and voting behavior. Small-group experiments assist the principal investigator in formulating new survey questions to determine the impact nonseparable preferences have on survey responses. The investigator uses new survey questions in a random sample telephone survey of 400 residents of Columbus, Ohio. Data from the survey enables social scientists to judge whether nonseparable preferences explain some common empirical problems in public opinion, including question-order effects, over-time instability in preferences, and the apparently minimal impact of issues on voter choice in elections. With new survey instruments available to capture nonseparable preferences, political scientists are better equipped to investigate voting behavior and public opinion, and policymakers have a clearer view of what people actually want on important public policy issues. ***