- 最后登录
- 2014-10-14
- 注册时间
- 2014-8-2
- 威望
- 0
- 金钱
- 105
- 贡献
- 35
- 阅读权限
- 20
- 积分
- 140
- 日志
- 0
- 记录
- 0
- 帖子
- 15
- 主题
- 5
- 精华
- 0
- 好友
- 0
![Rank: 3](static/image/common/star_level1.gif) ![Rank: 3](static/image/common/star_level1.gif) ![Rank: 3](static/image/common/star_level1.gif)
- 注册时间
- 2014-8-2
- 最后登录
- 2014-10-14
- 积分
- 140
- 精华
- 0
- 主题
- 5
- 帖子
- 15
|
谢谢老师的及时回复。陈晓萍徐淑英老师主编的《组织与管理研究的实证方法》第十七章多层次理论模型的建立及研究方法P451页里面有公式组内相关(1)即ICC(1)=组间方差/(组间方差+组内方差)这里的组间、组内方差正是由HLM分析直接得到的。同时,Paul D. Bliese(2002)的Multilevel Random Coefficient Modeling in Organizational Research Examples Using SAS and S-PLUS这一文章里也有提到,P420是这样说的The unconditional means model provides between-group and within-group variance estimates in the form of A and B respectively. As with the ANOVA model, it is often valuable to determine how much of the total variance is between-group variance. This can be accomplished by calculating the Intraclass Correlation Coefficient (ICC) using the formula: ICC=A/(A+B) (see Bryk & Raudenbush, 1992; Kreft & de Leeuw, 1998). The ICC is equivalent to Bartko's (1976) ICC(l) formula and to Shrout and Fleiss's ICC(1,1) formula (see Bliese, 2000b). Both of these latter two ICC measures are estimated from one-way random-effects ANOVA models. Note that the estimates of within-group and between group variance from the random coefficient model will be nearly identical to those from the AN OVA model as long as restricted maximum likelihood estimation (REML) is used in the random coefficient modeling (this is the default in most programs). If full maximum likelihood is used, the variance estimates may differ somewhat from the AN OVA estimates, particularly if sample sizes are small. All programs allow a choice between the two options.(注:限于此平台的不足,组间、组内方差分别用A、B代替)。另外,ICC(1)和ICC(2)一般来说比较小就可以了,可我们的分析结果普遍偏大,在0.7-0.9范围内,不知道这是不是个好的结果呢?还是其他什么原因呢?请老师解惑!再次感谢老师! |
|