中山大学管理学院2012年度前沿方法系列研讨班：潜变量混合模型
一、潜变量混合模型简介About Mixture Latent Modeling
潜变量混合模型（Mixture Latent Modeling）作为近几年新兴的前沿统计方法包括：潜类别分析模型（Latent Class Analysis ， LCA），潜剖面分析模型（Latent Profile Analysis ， LPA），增长混合模型 （Growth Mixture Modeling，GMM），以及混合模型与传统模型（例如验证性因子分析、多层线性模型分析、纵向数据分析等）的各种结合分析。这一新兴方法运用复杂成熟的概率运算原理，弥补了传统统计方法的基于观测变量和单一分析的局限，打破了传统方法关于总体同质性的基本假设，将潜变量分析与各种统计方法融为一体，提供了更准确和深入的分析思路。潜变量混合模型能够基于数据分析发现潜类别，识别潜类别的比例、特征、及其各自动态增长趋势，同时能够检验潜类别及动态增长趋势的相关预测变量。近年来，国际管理学期刊上关于研究潜变量混合模型的应用取得了快速的发展，国际管理学会年会上也开始有相应的专题教学和研讨。潜变量混合模型能帮助管理界学者更加科学地准确地分析数据，能够同时验证对总体、潜类别及个体的多种假设，因此能够更好地解释和预测企业组织的异质性和动态性。潜变量混合模型方法将成为各类高层次学术刊物、高层次管理研究的必备方法。
Mixture latent modeling, an important new development in latent variable analysis and change modeling, embraces major sub-topics such as latent class analysis (LCA), latent profile analysis (LPA), growth mixture modeling (GMM), and other mixture models that combine latent variable techniques with traditional statistical methods (e.g., confirmatory factor analysis (CFA), hierarchical linear modeling (HLM), longitudinal analysis, etc). Based on sophisticated probability hypotheses and algorithm, these mixture models are able to identify latent subgroups of distinct properties, estimate the proportion of the population in each latent subgroup, specify different subgroup-specific growth trajectories allowing individual variation around subgroup-mean trajectories, and test predictors of the trajectories and predictors of latent class membership concurrently. Until recently, management scholars have started to apply mixture latent modeling in scholarly works. Related workshops have also been increasingly offered during major international management conferences. Mixture latent modeling allows the management scholars to conduct more in-depth analysis, to model a variety of substantive hypotheses, and hence better explain and predict the heterogeneous and dynamic nature of organizations in a comprehensive fashion.
二、关于研讨班及目标学员About The Workshop and Who Should Attend
中山大学管理学院一向致力于向华人管理学者介绍最前沿的应用统计方法。本次讲座作为中山大学管理学院方法论系列讲座第二期，是首次全国范围内的关于潜变量混合模型的中文讲座。本次讲座致力于向华人学者推广潜变量混合模型统计方法，介绍相关基本概念，解释其运算原理和模型假设，并示范如何使用相关统计软件（MPLUS）建立及验证合适的模型和报告相关结果。讲授者将结合自己多年的实践经验，通过实例对主要潜变量混合模型分析方法进行了演示和说明，并阐述其及在管理领域中的应用，从而协助学员更深入和全面地进行组织管理相关研究。
This Mixture Latent Modeling Workshop is the 2nd annual workshop sponsored by SYSBS, featuring applications-oriented workshops focusing on the latest trends in statistical analysis. Led by a group of experts in the field, as the first national Mixture Latent Modeling Workshop in Chinese, the workshop aims to tutor Chinese scholars how to understand the basic concepts and statistical rationale, learn and apply major mixture models in the management field. In addition, the instructors will exemplify how to build and test major mixture models with software MPLUS, as well as how to report and interpret related results.
温馨提示：学员应提前熟知相关的传统统计方法（包括聚类分析、结构方程、验证性因子分析、多层线形模型和纵向数据分析等）。熟悉MPLUS统计软件将有助于更好的理解研讨班中的实例演示。
Reminder: Participants should be familiar with related conventional statistical methods, inducing cluster analysis, structure equation modeling, CFA, HLM, and longitudinal studies. Familiarity with MPLUS would be helpful but is not required.
三、培训日程（暂定）Tentative Schedule
日期：2012年7月21日
Date：July 21th, 2012
|
8:30-9:00 |
Check in |
|
9:00-9:30 |
Welcome Introduction Yadong Luo, University of Miami and Sun Yat-sen Business School |
|
9:30-11:00 |
Section 1
1. Mixture model overview
2. Latent class analysis (LCA) a. Categorical latent variable modeling with continuous indicators (LPA). b. Categorical latent variable modeling with categorical indicators (LCA). |
|
11:00-11:15 |
Tea Break |
|
11:15-12:30 |
Section 2 3. Mixture in structural equation modeling and multilevel modeling a. Mixture in CFA and Structural equation modeling b. Mixture in multilevel modeling (random intercepts and slopes) |
|
12:30-2:00 |
Lunch Break |
|
2:00-3:30 |
Section 3 4. Growth mixture modeling (GMM) a. Latent growth modeling and GMM b. Advanced GMM i. GMM with time-varying covariates ii. GMM with zero-inflated observations |
|
3:30-3:45 |
Tea Break |
|
3:45-5:30 |
Section 4 5. Missing value modeling in latent mixture analysis 6. In-class examples and software demo |
|
5:30-6:00 |
Questions and Answers |
四、讲授者简介Instructor Biography
Leading Instructor
Mo Wang, a tenured Associated Professor at University of Florida, specializes in research and applications in the areas of retirement and older worker employment, occupational health psychology, cross-cultural HR management, leadership, and advanced quantitative methodologies. He has received Academy of Management HR Division Scholarly Achievement Award (2008), Careers Division Best Paper Award (2009) and Erasmus Mundus Scholarship for Work, Organizational, and Personnel Psychology (2009) for his research in these areas. He also received Early Career Achievement Awards from Society for Industrial-Organizational Psychology (2012), Academy of Management’s HR Division (2011) and Research Methods Division (2011), and Society for Occupational Health Psychology (co-sponsored by the APA and NIOSH, 2009). He currently serves as an Associate Editor for Journal of Applied Psychology. He also serves on the Editorial Boards of Personnel Psychology, Journal of Management, Organizational Research Methods, Journal of Occupational Health Psychology, and Journal of Business and Psychology. He is the Editor for the Oxford Handbook of Retirement. He has been contracted by several Fortune 500 companies and government agencies to provide consulting services in both English and Chinese.
Coordinating Instructor
Stephanie L. Wang is a research fellow at the University of Miami, School of Business Administration where she teaches strategic management and international business. Her main research interests include global strategy, multinational management, internationalization by emerging market firms, business/knowledge process outsourcing, and entrepreneurship. She has publications in Journal of International Business Studies, Academy of Management Perspective, Journal of International Management, Organizational Dynamics and others. Her latest research projects deal with the relationships between home operations and overseas operations of emerging market multinationals. She uses numerous methodologies in her research, including structural equation modeling, multi-level modeling, meta-analysis, latent growth modeling, among others.
五、注册信息Registration
该讲座将对全国师生免费（食宿差旅费学员自行承担）。由于资源有限，请有意参加者尽早报名，超出人数将会在回复邮件中通知，请注意查收。学员可选择通过管理学院预定当日午餐（自助餐每人50元人民币）。
Due to limited space, we encourage early application and pre-registration. The registration will be on a first-come-first-serve basis. Please apply as soon as possible to book your place. The workshop is free for all faculty members and doctoral students. The participants can choose to attend the buffet lunch prepared by Sun Yat-sen Business school at a cost of 50 RMB.
1) 请联系魏菲女士（weifei@mail.sysu.edu.cn）通过电子邮件报名注册，于7月1日前填写并回寄下列表格。
|
姓名 |
|
专业 |
|
学校 |
|
||
|
学生 |
是 否 |
电话 |
|
Email |
|
||
|
通讯地址 |
|
||||||
|
是否具备聚类分析的知识 |
是 否 |
是否熟悉验证性因子分析 |
是 否 |
||||
|
是否具备结构方程的知识 |
是 否 |
是否熟悉多层数据分析 |
是 否 |
||||
|
是否熟悉纵向数据分析 |
是 否 |
是否用过MPLUS软件 |
是 否 |
||||
Please contact Ms. Fei Wei at weifei@mail.sysu.edu.cn to register for the workshop. Please send out the registration request with the following form attached before July 1, 2012.
|
Name |
|
Major |
|
University |
|
||
|
Students |
Yes No |
Phone No. |
|
Email |
|
||
|
Mailing address |
|
|
|
||||
|
Are you familiar with cluster analysis |
Yes No |
Are you familiar with CFA |
Yes No |
||||
|
Are you familiar with SEM |
Yes No |
Are you familiar with HLM |
Yes No |
||||
|
Are you familiar with longitudinal data analysis |
Yes No |
Do you know how to use Mplus |
Yes No |
||||
2) 具体培训材料将稍后提供。
Readings: A set of readings to complement the lectures will be available on a secured website for registered participants to download. Please read them before the workshop to get most out of lectures. The link of the website will be provided to you later.
3) 如要取消注册，请于一周前通知主办方。
Cancellation: Cancellation made one week before the workshop commencement will not incur cancellation charges.
Sun Yat-sen Business School
Sun Yat-sen Univeristy
Guangzhou, Guangdong, China
中山大学管理学院
2012年3月