A Comparative Analysis of Political Factions Using Multinomial Probit Techniques

政治理论中一个长期的讨论是关于派系之间的关系和实现共同利益的可能性。例如，麦迪逊认为，派系的影响在某些情况下是可以减轻的。这项研究的前提是，派系涉及到选择高度相关的个人群体。这一假设与传统的派系概念一致，允许使用多项概率（MNP）模型中的最新统计技术来间接计算群体中的行为相关性。为了测试这种建模策略的能力，MNP被用来生成六个欧洲政治和美国国会的几何表示。这些模型阐明了向心性或离心性政治力量将派系模式与政党联系在一起的程度。选择进行分析的国家允许测试几种不同的选举制度，以及这些投票规则如何产生一个占主导地位的中间政党的众多政党。

A longstanding discussion in political theory concerns the relationship between factions and the possibility of attaining the common good. Madison, for example, argued that the effects of faction could be mitigated under certain circumstances. This research begins with the premise that factions involve groups of individuals whose choices are highly correlated. This assumption, which is in line with traditional concepts of factions, allows recent statistical techniques in multinomial probit (MNP) modelling to be used to indirectly compute behavioral correlations in a population. To test the capacity of this modelling strategy, MNP is used to generate geometric representations of six European polities and the U.S. Congress. Such models clarify the extent to which centripetal or centrifugal political forces connect factional patterns to political parties. The nations chosen for analysis allow for tests of several different electoral systems and how those voting rules give rise to a multitude of parties with a dominant center party.