- 1、本文档共17页,可阅读全部内容。
- 2、有哪些信誉好的足球投注网站(book118)网站文档一经付费(服务费),不意味着购买了该文档的版权,仅供个人/单位学习、研究之用,不得用于商业用途,未经授权,严禁复制、发行、汇编、翻译或者网络传播等,侵权必究。
- 3、本站所有内容均由合作方或网友上传,本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺!文档内容仅供研究参考,付费前请自行鉴别。如您付费,意味着您自己接受本站规则且自行承担风险,本站不退款、不进行额外附加服务;查看《如何避免下载的几个坑》。如果您已付费下载过本站文档,您可以点击 这里二次下载。
- 4、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等,请点击“版权申诉”(推荐),也可以打举报电话:400-050-0827(电话支持时间:9:00-18:30)。
查看更多
Motivating example Random Effects Other topics References HLM Short Course Introduction.pdf
Motivating example Random Effects Other topics References
HLM Short Course: Introduction
Lane Burgette
Department of Statistics
University of Wisconsin-Madison
/~burgette
March 5, 2008
Motivating example Random Effects Other topics References
Note
These slides will ignore pretty much all the math. See, e.g.,
Pinheiro and Bates (2000) for the details.
Motivating example Random Effects Other topics References
Hierarchical structures
Hierarchical linear models (HLM) (or mixed models for
statisticians) are a tool for analyzing data when we have a
nested structure
For example:
• Students within schools
• Residents within neighborhoods
• Multiple observations per person
Motivating example Random Effects Other topics References
Example
• Interested in modeling academic performance for students
• If we ignore nesting of students within classrooms, we
have two options
• Either model with a school-level error, or student-level error
• Both methods are problematic
Motivating example Random Effects Other topics References
School-level analysis
• Possible to model average performance, with school-level
error
• How do we weight observations if within-school sample
sizes vary?
• Cannot include student-level predictors
• Lose information about student-level variation – may be of
intere
文档评论(0)