variance) where models are fitted to raw data based on the assumption that there
is no error in the measurement of the independent variable/s (Raykov and
Marcolides 2006: 2-3).
Other major advantages of SEM analyses are the ability to test complex
multivariable models and to test hypotheses about both direct and indirect
effects.
SEM analyses can be used in both confirmatory mode (for the purposes of theory-
testing) and in exploratory mode (for theory-building). In theory-building, the
operation is exploratory insofar as models are tested, modified and tested again
in the search for an optimal model (Kline 2005: 10-11).
6.2 Modelling approach
This research employs Jöreskog’s (1993: 295) Model Generating (MG) approach in
which the model is modified and tested again using the same data. Raykov and
Marcolides (2006: 7) write that:
In contrast to the confirmatory mode of SEM applications, theory
development assumes that no prior theory exists—or that one is
available only in a rudimentary form—about a phenomenon under
investigation.This describes very well the context for this research. While a number of
hospitality studies focussing on leadership, and on motivation, and several linking
leadership with motivation/job performance have been published, the overall
state of knowledge remains rudimentary. This situation is largely due to the wide
ranges of (a) theoretical approaches adopted and (b) specific foci chosen by
individual researchers/research teams (as described in Chapter 3 for hospitality
leadership studies and section 4.3 for hospitality motivation studies).
While the broad approach to the SEM analyses is model generating, the specific
type of SE (structural equation) model to be used is the structural regression
model (Kline 2005: 209). The strategy for developing and testing the structural
regression models is based on Anderson and Gerbing’s (1988) two-step procedure
with each step following the five-stage development procedure described by
Schumaker and Lomax (2004).