Parties influence public opinion far beyond their partisans: evidence from two quasi-experiments

Are parties able to influence public opinion beyond their supporters? Much of the literature on parties and public opinion finds that parties exercise considerable sway over their own partisans. However, I argue that parties in multi-party systems have broader influence – namely, they can persuade the partisans of their ally parties (“ally partisans”). Thus, I extend a key study in the field by performing difference-in-differences analyses on two rare instances in which some Danish parties shifted their policies on salient issues (unemployment benefits and early retirement), while other parties did not. However, unlike previous research, I test the influence on ally partisans. In addition, I analyse the difference-in-differences for the original partisans while omitting ally partisans from the control groups. My study also examines two potential control groups – partisans of opposing parties and nonpartisans – and concludes that the former group is more appropriate as a control. The findings demonstrate that parties influence (1) their ally partisans considerably, (2) their own partisans to a greater extent than found in past estimates, and (3) possibly even nonpartisans. This indicates that parties exercise broader and more powerful influence over public opinion than previously expected.

政党能否在自身支持者之外影响公众舆论？现有关于政党与公众舆论的研究文献普遍发现，政党对其自身党籍支持者拥有相当程度的影响力。然而，本文提出，多党制中的政党具备更广泛的影响力——具体而言，它们能够说服其盟友政党的支持者（ally partisans，以下简称"盟友支持者"）。据此，本文对该领域的一项核心研究进行拓展：针对丹麦部分政党就两大显著性议题（失业救济金与提前退休）调整政策、其余政党未作出调整的两起罕见案例，采用双重差分法（difference-in-differences）展开分析。与以往研究不同的是，本文重点检验政党对盟友支持者的影响力。此外，本文在将盟友支持者排除出对照组的前提下，针对原政党支持者开展双重差分分析。本研究同时设置了两类潜在对照组——反对党支持者与无党派人士，并得出结论：前者更适合作为对照组。研究结果表明，政党的影响力体现在三个方面：（1）对盟友支持者具备显著影响；（2）对自身支持者的影响力超出过往研究的估算值；（3）甚至可能对无党派人士产生影响。这意味着，政党对公众舆论的影响力范围与强度均远超此前的预期。