We tackle two problems. One is understanding parts of the operation of the US Supreme Court. The other is a fundamental problem for network analysis. It is delineating the fundamental structures of networks. Even more important, within this second problem, is delineating changes of this network structure over time. We present a method for doing both for signed networks. The two problems are coupled closely as the data come from the completed years of the Roberts US Supreme Court, named after its Chief Justice, for the 2005 through 2018 terms. For the issues selected by the court for consideration in each term, the justices vote on the decision that will be issued by the court. These votes are either to support a decision or to dissent from it. These votes can be recorded into a signed 2-mode network for each term. While we examine these networks, our primary focus is on the 1-mode projection from the 2-mode network having the justices as units. Using signed relaxed structural balance blockmodeling, we establish the fundamental structure of the relations between justices for each of the terms. For 13 of the 14 terms considered, the criterion function for the blockmodels is zero. This structure changes in clear ways and shows that the conventional divide between conservative and liberal justices is overly simplistic. Of greater interest is identifying the structural roles of the court's justices.

我们解决两个问题。一种是了解美国最高法院运作的一部分。另一个是网络分析的基本问题。它描述了网络的基本结构。在第二个问题内，更重要的是描绘该网络结构随时间的变化。我们提出了一种用于签名网络的方法。这两个问题紧密相关，因为数据来自以首席大法官命名的罗伯茨美国最高法院关于2005年至2018年任期的数据。对于法院在每个任期中选择要考虑的问题，法官将对法院将发布的裁决进行表决。这些投票要么是支持一项决定，要么是反对该决定。可以将这些选票记录到每个学期的签署的2模式网络中。当我们检查这些网络时，我们主要关注以正义为单位的2模网络的1模投影。使用带符号的宽松结构平衡模块，我们为每个术语建立了正义之间关系的基本结构。对于所考虑的14个项中的13个，块模型的标准函数为零。这种结构以明显的方式变化，表明保守和自由司法之间的传统划分过于简单。更加有趣的是确定法院法官的结构性角色。块模型的标准函数为零。这种结构以明显的方式发生了变化，表明保守派和自由派司法之间的传统划分过于简单。更加有趣的是确定法院法官的结构性角色。块模型的标准函数为零。这种结构以明显的方式发生了变化，表明保守派和自由派司法之间的传统划分过于简单。更加有趣的是确定法院法官的结构性角色。