Analyzing multimethod data

Michael Eid, Berlin, Germany

The question as to which structural equation model should be selected when multitrait-multimethod (MTMM) data are analyzed is of interest to many researchers. In the past, attempts to find a well-fitting model have often been data-driven and highly arbitrary. The purpose of this proposed workshop is to show how the measurement design (type of methods used) can guide the choice of a model. Based on the distinction between (a) interchangeable methods, (b) structurally different methods, and (c) the combination of both kinds of methods a model for each class of methods will be discussed in detail. Interchangeable methods are randomly selected from a set of possible methods. If different raters are considered as different methods, interchangeable raters are randomly selected from a group of possible raters (e.g., several interchangeable students rating their teachers). With respect to interchangeable methods, a multilevel confirmatory factor model is presented. For structurally different methods, the Correlated-Trait-Correlated-(Method–1) model is recommended. Structurally different raters are not interchangeable but differ in several important ways (e.g., parents and teachers rating a student). Finally, we demonstrate how to appropriately analyze data from MTMM designs that simultaneously used interchangeable and structurally different methods. All models are applied to empirical data to illustrate their proper use. Moreover, the M-Plus inputs and outputs will be provided to show how the models can be applied. Finally, the models will be compared to previously developed MTMM models, and some implications for modeling MTMM data are discussed.