Introductory Guide to HLM with HLM 7 Software
Introductory Guide to HLM with HLM 7 Software
复制标题
使用 HLM 7 软件的 HLM 入门指南
DOI：
10.4135/9781483384450.n3
复制
发表时间：
2013
影响因子：
10.5
通讯作者：
G. Garson
中科院分区：
文献类型：
--
作者：
G. Garson
由于 Stephen Raudenbush 和 Anthony Bryk 的开创性工作，LM 软件已成为分层线性建模的领先统计软件包之一，他们创建了该软件并撰写了分层线性和非线性建模的主要文本（Bryk & Raudenbush，1992；Raudenbush & Bryk，2002）。尽管随着时间的推移，软件包功能之间的差异已经减少，但 HLM 7 提供了许多吸引人的优势和功能。其中包括许多人认为更直观的模型规范环境、更容易创建三级和四级模型、广泛的估计选项选择、集成似然比假设检验、图形选项以及轻松处理异构分层线性模型的能力（其中依赖项被认为对于某些分组变量（例如代理机构）的不同级别具有不同的误差方差）。 1 学生版功能齐全，包括示例，但模型的大小和复杂性受到限制（尽管它适用于软件提供的所有示例文件）。 HLM 7 软件通过多个模块运行，每个模块针对不同类型的 HLM 模型而设计，由于篇幅限制，此处只能说明其中的一些模块：HLM2。适用于具有一个因变量的两级线性和非线性模型。 HLM3 和 HLM4。适用于具有一个因变量的三水平和四水平模型。 HGLM。对于除正态分布和除恒等之外的链接函数的广义线性模型，处理伯努利、二项式、泊松、多项式和序数模型中的二元、计数、多项式和序数结果变量。
LM software has been one of the leading statistical packages for hierarchical linear modeling due to the pioneering work of Stephen Raudenbush and Anthony Bryk, who created the software and authored the leading text on hierarchical linear and nonlinear modeling (Bryk & Raudenbush, 1992; Raudenbush & Bryk, 2002). Though differences among software packages' capabilities have diminished over time, HLM 7 offers a number of appealing advantages and capabilities. Among these are what many consider to be a more intuitive model specification environment, greater ease in creating three-and four-level models, its wide choice of estimation options, integrated likelihood ratio hypothesis testing, graphics options, and the ability easily to handle heterogeneous hierarchical linear models (where the dependent is thought to have different error variances for different levels of some grouping variable such as Agency). 1 The student edition is full-featured, including examples, but is limited in the size and complexity of models (though it will work with all example files provided with the software). HLM 7 software operates through several modules, each designed for a different type of HLM model, only some of which can be illustrated here due to space constraints: HLM2. For two-level linear and nonlinear models with one dependent variable. HLM3 and HLM4. For three-level and four-level models with one dependent variable. HGLM. For generalized linear models for distributions other than normal and link functions other than identity, handling binary, count, multinomial, and ordinal outcome variables in Bernoulli, binomial, Poisson, multinomial, and ordinal models.