Fitting Dynamic Regression Models to Seshat Data - Supplemental Material

This article presents a general statistical approach suitable for the
analysis of time-resolved (time-series) cross-cultural data. The goal is
to test theories about the evolutionary processes that generate cultural
change. This approach allows us to investigate the effects of predictor
variables (proxying for theory-suggested mechanisms), while controlling
for spatial diffusion and autocorrelations due to shared cultural history
(known as Galton’s Problem). It also fits autoregressive terms to account
for serial correlations in the data and tests for nonlinear effects. I
illustrate these ideas and methods with an analysis of processes that may
influence the evolution of one component of social complexity, information
systems, using the Seshat: Global History Databank.

本文提出一种适用于分析时间分辨（time-series）跨文化数据的通用统计方法，旨在检验关于驱动文化变迁的演化过程的理论。该方法能够在控制空间扩散以及由共享文化历史导致的自相关（即高尔顿问题，Galton’s Problem）的同时，探究预测变量（代表理论提出的机制）的效应；此外，还通过拟合自回归项解释数据中的序列相关，并检验非线性效应。笔者利用Seshat：全球历史数据库，通过分析可能影响社会复杂性组成部分之一——信息系统演化的过程，对这些理念与方法进行了阐释。