QuantDev Brown Bag Seminar Series (Online)

Model equivalences: How contrasts keep showing up in different repeated measures approaches.

Abstract: 

Contrasts represent an important analytic tool in the completion of many repeated measures analytic approaches.  They can represent planned comparisons or follow-up comparisons.  They can be used to create dependent variables used to model change, or they can be used to create predictor variables that model the same characteristics of change.  Linear combinations of simple contrasts form more complex polynomials.  As such, they implicitly show up in many different approaches including repeated measures ANOVA, linear mixed model (LMM) curve analysis, and factor analytic approaches to growth curve analysis.  When the same contrasts appear in different models, they create ways of considering the equivalence of different analytic approaches. Some of these equivalences will be demonstrated and ways in which they fit into a more general approach to model equivalences will be discussed.