like any data statistical test, there are few/many assumptions those should check before that particular statistical test. For example, in case of regression, there are many assumptions including, multi-collinearity, normality of data bla bla bla.
I want to ask the assumptions those should be checked before applying multi-level analysis / HLM.
and if these assumptions very from software to software like for HLM and SPSS, there are different assumptions or same?
Please guide
And also recommend any basic book that address these basic questions related to HLM/ multi level modeling
Thanks
There are several assumptions for multilevel analyses:
- Proper model specification
- Homogeneity of variance of residuals
- Random sampling
- Adequate sample size
- Similar group sizes (use full information maximum likelihood FIML estimators when group sizes differ significantly)
- Continuous outcome variables
- Linearity
- No multicollinearity
- Normal distribution of variables
- Normal distribution of residuals
- Absence of statistical outliers
- Properly specified covariance structures
- Independent level-two groupings (while observations at level one are not assumed to be independent, the groups formed by the subject variable(s) are assumed to be independent and to have the same covariance structures)
For more information on all of these assumptions, please refer to chapter 2 in Hierarchical Linear Modeling: Guide and Applications (Garson, 2013).
https://methods.sagepub.com/book/hierarchical-linear-modeling
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Software