One of the biggest barriers to learning statistics is learning the lingo. Unfortunately, terms, abbreviations, and mathematical notation vary between disciplines (eg biometrics and econometrics) and even authors within the same discipline. For example, "nested data" is also called "hierarchical" and "multilevel" (Zuur et al 2009). Similarly, "Split plots, strip plots, repeated measures, multi-site clinical trials, hierarchical linear models, random coefficients, analysis of covariance are all special cases of the mixed model" (COMSYS Information Technology Services, paper 198-30).
An article that looks like it might provide some tools for understanding how different ANOVA models relate to each other, and why they are given different names, is Schuster and Von Eye (2001): "The relationship of ANOVA models with random effects and repeated measures". Its available online in the Journal of Adolescent Research. The authors assert that despite the multitude of different ANOVA designs that are typically presented in statistics books, there are unifying conditions -- block vs treatment factors, fixed vs random effects, and crossed vs nested factors -- that allow many designs with different names to shown to be equivalent conceptually, if not mathematically.
No comments:
Post a Comment