Multimethod research programs require diverse methods of data analysis that take the multimethod character into account. Methods analyzing the convergence of different methods that are supposed to measure the same latent construct have a long tradition in psychometrics. They can be traced back to Spearman's (1904) claim of a correlational psychology that should detect common structures underlying fallible measures that are distorted by several error influences. The history of psychometrics can be considered the refinement of methodological approaches explaining multivariate associations in a more appropriate way. Because of the great importance of multimethod research strategies, a host of methodological approaches for the analysis of multimethod structures have been developed (Dumenci, 2000; Schmitt & Stults, 1986). However, not all of them can be considered in this handbook, and some reasons why we selected certain methodological approaches to be discussed in more detail in the following chapters will given in the current chapter. The aim of this chapter is threefold. First, five conceptual distinctions will be discussed that influence the choice of data-analytic approaches for analyzing multimethod data. Second, an overview of several statistical approaches for multimethod data that are dealt with in more detail in the following chapters will be given to highlight their essentials and to provide reasons why these approaches have been chosen. Third, more traditional approaches such as correlation analysis, which is not discussed in more detail in the following chap ters, will be considered here to provide the reader with in-depth knowledge of multimethod approaches.
Was this article helpful?