Building capacity in Bayesian analysis for practicing ecologists

传统上，生态学家使用设计的实验和相关的统计程序来了解环境。虽然这种方法提供了有价值的见解，但很明显，许多紧迫的生态问题不能用经典的设计和分析方法来解决。特别是，生态学研究人员需要新的方法来诚实地表示不确定性，使用来自多个来源的数据，估计无法观测到的数量，并对环境做出预测。分层贝叶斯方法满足了这些需求，使调查人员能够解决在此之前被视为过于复杂而无法回答的问题。大学和管理机构的许多生态学家认识到这些方法的前景，但缺乏有效的方法来学习它们。研究小组将向博士后研究人员、学术人员以及机构和非政府组织的科学家提供关于分层贝叶斯方法的强化培训。将通过在里约热内卢的波多黎各大学举办一次讲习班来加强科学界任职人数不足的群体的参与。将开发一个关于贝叶斯方法在生态学中的应用的列表服务器。将维持一个可开放查阅培训材料的网站。随着新培训的科学家将贝叶斯方法应用于他们的研究、教学和决策支持，该项目的影响将扩大几倍。

Traditionally, ecologists used designed experiments and associated statistical procedures to learn about the environment. Although this approach provided valuable insight, it has become clear that many pressing problems in ecology cannot be solved using classical methods for design and analysis. In particular, ecological researchers need new methods to honestly represent uncertainty, to use data from multiple sources, to estimate quantities that cannot be observed, and to make forecasts about the environment. Hierarchical Bayesian methods meet these needs, allowing investigators to address questions that heretofore were seen as too complex to answer. Many ecologists at universities and management agencies recognize the promise of these methods but lack an efficient way to learn them. The research team will offer intensive training on hierarchical Bayesian methods to post-doctoral researchers, academic faculty, and scientists at agencies and non-governmental organizations. Participation by groups under-represented in science will be enhanced by holding one of the workshops at the University of Puerto Rico in Rio Piedras. A listserv on the application of Bayesian methods in ecology will be developed. A website with open access to training materials will be maintained. The impact of the project will be amplified several fold as newly trained scientists apply Bayesian methods in their research, teaching, and decision-support.