Bayesian Analytical Methods: A Methodological Prescription for Public Administration

Abstract

In this article we describe in detail the Bayesian perspective on statistical inference and demonstrate that it provides a more principled approach to modeling public administration data. Because many datasets in public administration are population-level, one-time unique collections, or descriptive of fluid events, the Bayesian reliance on probability as a description of unknown quantities is a superior paradigm than that borrowed from Frequentist methods in the natural sciences where experimentation is routine. Here we provide a thorough, but accessible, introduction to Bayesian methods and then demonstrate our points with data on interest group influence in US state administrative agencies. 《贝叶斯分析模型:对公共行政管理学研究方法的探讨》 作者:杰夫·吉尔,克利斯朵夫·维特科 我们在本文中具体讨论了贝叶斯理论 (Bayesian perspective) 在统计推断中的应用及其在分析公共行政管理数据的优越性。绝大多数公共行政管理的数据是宏观的统计数据。此类数据的样本采集往往是一次性的而并非重复随机性的。此外,公共行政管理的很多数据都用来纪录多变的(而非恒定的)事件。所有这些数据的特点决定了贝叶斯概率理论比源自与自然科学试验研究的频率学派的方法 (Frequentist methods) 更适用与公共行政管理的研究。在此,我们以利益集团 (interest group) 影响美国州政府部门的数据为例,用通俗易懂的方式全面介绍贝叶斯分析方法及其应用。