MCS_陈凌隽Lyn_新浪博客

http://blog.sina.com.cn/u/1895434237

首页博文目录关于我

个人资料

微博

加好友发纸条

写留言加关注

博客等级：
博客积分：

博客访问：
关注人气：
获赠金笔：0支
赠出金笔：0支
荣誉徽章：

正文字体大小：大中小

MCS

(2019-02-27 09:13:12)

标签：

杂谈

分类：知无涯

Flow of Monte Carlo Simulation (MCS): repeat (generate dataset+analyses) multiple times-->get meta-statistics by summarizing the results

[How to do Monte Carlo Study with SEM?]

sim() function in the R package "simsem" can use "lavaan" to complete the above flow.

Please search "R Codes for a Simulation Using R Package simsem and lavaan"

in paper

Lee, S. (2015). Implementing a Simulation Study Using Multiple Software Packages for Structural Equation Modeling. SAGE Open. https://doi.org/10.1177/2158244015591823

The codes are not complicated.

How can it be useful?

Example:

Hsu, H. Y., Kwok, O. M., Lin, J. H., & Acosta, S. (2015). Detecting misspecified multilevel structural equation models with common fit indices: A Monte Carlo study. Multivariate Behavioral Research, 50, 197-215.

It may not be so useful for applied research right now. However, some kind of simple simulation in below may be required by reviewer(s).

examples of simulate the data just once using lavaan:

http://lavaan.ugent.be/tutorial/mediation.html

https://rdrr.io/cran/lavaan/man/simulateData.html

search "# specify population model"

[other tests using MCS]

There are many other packages. I just listed 2 here.

http://philchalmers.github.io/SimDesign/pres.pdf introduced some concepts related to MCS, such as RE, RMSE

RE: relative efficiency, treating one estimator as the reference-->compare and evaluate estimator

https://cran.r-project.org/web/packages/MonteCarlo/vignettes/MonteCarlo-Vignette.html

阅读┊ 收藏 ┊ 喜欢 ▼ ┊打印┊举报/Report

前一篇：“美妙人生的关键在于你能迷上什么东西”

后一篇：最重要的不是时间管理方法，而是

新浪BLOG意见反馈留言板　欢迎批评指正