ABSTRACT
This chapter introduces basic tenets and uses of structural equation modeling (SEM). The term SEM denotes classes of models that include confirmatory factor analysis, structural equation models for observed variables, and latent variable structural equation models (Loehlin, 1992). Confirmatory factor analysis models specify measurement relationships among observed and latent variables; observed variable structural equation models specify theoretical relationships among variables without a measurement structure imposed, and full structural equation models specify relationships among latent variables. Use of SEM is widespread in nursing research because it is suited to measurement problems (e.g., Folse, 2007; Lynn, Morgan, & Moore, 2009; Sousa, Ryu, Kwok, Cook, & West, 2007) and theory building (e.g., Melnyk, Crean, Feinstein, & Fairbanks, 2008; Phillips & Stuifbergen, 2009; Spence Laschinger & Leiter, 2006) and software is readily available. We present SEM models suitable for continuously varying data, discuss estimation and model assessment, introduce several special topics, provide an overview of common software packages, and list additional resources.
