ABSTRACT
More and more diverse applications of structural equation modeling (SEM) have been used with increasing frequency in the organizational and social sciences since the early 1980s (Austin, Scherbaum, & Mahlman, 2002; Stone-Romero, Weaver, & Glenar, 1995). In part, this is due to the power and flexibility of SEM and due to the fact that SEM allows simultaneous estimation of measurement and structural relations in a single, integrated application. Several chapters in this volume have referred to two particular research design challenges that face work relationship researchers and that we address in this chapter on two particular SEM applications. First, relationships are intrinsically developmental and change over time as they commence, mature, and eventually dissolve. As such, it is critical in relationship research to be able to model data longitudinally, to track changes in behavior over time, and to model determinants and consequences of behavior change. A second challenge arises from the fact that individuals are naturally nested within dyadic relationships, which are nested within departments, business units, or families, which are nested within organizations, extended families, communities, and societies. That is, relationships are routinely hierarchically structured or multilevel.
