ABSTRACT
There are several definitions for structural equation modeling (SEM) and most of these definitions are field-specific (e.g., Byrne, 2001; Hoyle, 1995; Kaplan, 2000; Kline, 2005; Schumacker & Lomax, 2004). However, the most concise definition of SEM can be found in Klem (2000), who states, “In its most general form, SEM can be viewed as a combination of path analysis and factor analysis” (p. 227). The path analysis side focuses more specifically on the testing of simultaneous equations and the origins of this portion of SEM can be found in the fields of econometrics and genetics. The factor analysis side is comprised of restricted (i.e., confirmatory) factor analysis and stems from work completed in psychology and, more specifically, psychometrics (Kaplan, 2000). The origins of the technique are more in measurement than in the testing of structural relationships between variables (see Haggland, 2001), but Joreskog was instrumental in linking the two seemingly disparate techniques into a single analytical procedure we now call SEM (Kline, 2005).
