Multilevel and longitudinal modeling with IBM SPSS
Multilevel and longitudinal modeling with IBM SPSS
复制标题
使用 IBM SPSS 进行多层次和纵向建模
DOI：
10.1080/1743727x.2011.573269
复制
发表时间：
2011
影响因子：
2
通讯作者：
Xing Liu
中科院分区：
文献类型：
--
作者：
Xing Liu
赫克、托马斯和Tabata(2010)介绍了如何使用SPSS开发和分析多水平和纵向模型，并提供了清晰和循序渐进的解释。这本书的目的是帮助读者建立和分析多水平和纵向数据，并用SPSS混合程序解释结果。这是作者关于多级别建模技术的第二本书；然而，这是市场上第一本使用SPSS如此深入和全面地介绍这些模型的书。这本书的特色章节致力于介绍多水平和纵向建模；准备和检查多水平数据；建立两个和三个水平的多水平模型，使用重复测量ANOVA和多水平建模检查纵向数据的个人变化；检查组织水平的变化，并开发更高级的模型，如多变量和交叉分类的多水平模型。本书以使用SPSS18.0进行多层次和纵向建模的介绍性章节开始。介绍了目前已有的多水平数据分析软件包、使用SPSS的合理性、SPSS混合过程的特点及其局限性。给出了开发多层模型的一般策略，并简要讨论了SPSS混合语法和菜单命令的制定。本章为希望深入研究与多层建模相关的方法学问题的读者提供了有用的参考。第二章利用SPSS软件编制了几种类型的多水平和纵向数据。它解释了如何设置1级和2级数据集，以及如何将它们合并为一个数据集。对SPSS中的基本数据管理功能，如重新编码、计算、文件匹配、数据聚合等进行了清晰的回顾和介绍，我认为这对不熟悉这些功能的初学者非常有帮助。对于高级学习者来说，这一部分可以作为很好的复习材料。本章还演示了如何将纵向水平数据(每个人具有多变量的人级别数据结构)重构为垂直数据(人-期间数据结构，每个人中嵌套了多条记录)，以便为多级别分析做好准备。数据管理技能对于多级建模至关重要，因为数据最初需要准备好具有正确的结构。这也是使用专门的程序包(如HLM、MLwiN和Mplus)进行多层建模的研究人员的先决条件。因此，第二章可能会对所有潜在的读者有所帮助。第三章和第四章分别向读者介绍了建立两级和三级模型的步骤。作者的写作风格非常有效。从第三章开始，每章都有一个研究实例，介绍研究问题、数据、如何使用SPSS建立多水平模型的清晰、循序渐进的说明，
Heck, Thomas, and Tabata (2010) introduce how to develop and analyse multilevel and longitudinal models using SPSS with clear and step-by-step explanations. This book aims to help readers to build and analyse multilevel and longitudinal data, and interpret results with the SPSS Mixed procedure. This is the authors’ second book on multilevel modeling techniques; however, it is the first book on the market to cover these models to such a deep and comprehensive extent with SPSS. The book features chapters devoted to introducing multilevel and longitudinal modeling; preparing and examining multilevel data; building two and three-level multilevel models, examining individual change with longitudinal data using repeated measures ANOVA, and multilevel modeling; examining organizational-level change, and developing more advanced models, such as, multivariate and cross-classified multilevel models. The book begins with an introductory chapter on multilevel and longitudinal modeling using SPSS 18.0. It provides an introduction of currently available software packages for multilevel data analyses, a rational of using SPSS, features of SPSS Mixed procedure and its limitations. It presents a general strategy for developing multilevel models, and it briefly discusses SPSS Mixed syntax and menu command formulation. This chapter provides useful references for readers who wish to delve into methodological issues related to multilevel modeling. Chapter 2 focuses on preparing several types of multilevel and longitudinal data using SPSS. It explains how to set up level 1 and level 2 data sets, and how to merge them into one data set. The basic data management functions in SPSS, such as recoding, computing, matching files, and aggregating data, are reviewed and introduced clearly, which I think is extremely helpful to beginner learners who are not familiar with these functions. For advanced learners, this section can serve as a good review. The chapter also demonstrates how to restructure longitudinal horizontal data (person-level data structure with multivariate variables for each person) to vertical data (person-period data structure, with multiple records nested within each person) to prepare them for multilevel analysis. Data management skills are essential for multilevel modeling, since the data need to be initially prepared with the right structure. This is also a prerequisite for researchers who use specialized packages for multilevel modeling, such as HLM, MLwiN, and Mplus. Thus, Chapter 2 may be helpful for all potential readers. Chapters 3 and 4 take readers through steps of building a two-level and a three-level model, respectively. The authors’ writing style is very effective. Starting from Chapter 3, each chapter features an example of a study, presenting the research questions, the data, clear, step-by-step instructions of how to build a multilevel model using SPSS,