nomological validity as an issue to be addressed during the testing of the
structural (rather than measurement) model.
Because the structural model specifies theory-driven direction and causality of
relationships, Shumaker and Lomax’s approach makes greater sense with regard
to ensuring that the relationships between factors does not run counter to theory.
Nevertheless, one can see how the measurement model provides a preliminary
indication of nomological validity insofar as the measurement mode provides an
assessment of whether or not the factors do in fact covary (at a statistically
significant level) with each other; this being a prerequisite for a successful
structural model.
Face validity
Face validity is the extent to which the content of the items (the
questions/statements) is consistent with the with the construct definition (i.e. do
the items measure the concept they are intended to?) and is established prior to
the data collection stage of the modelling process (see Chapter 5 above on
developing the measures for the constructs). Face validity can also be referred to
as construct validity (see e.g. Kline 2005: 60).
7.2.6 Assessment of model fit
One of the advantages of SEM analysis is that it provides researchers with the
ability to make assessments with respect to the adequacy with which the overall
theoretical model accurately reflects the collected data (Kline 2005: 15). Step 1
of Anderson and Gerbing’s (1988) two-step procedure concentrates on
establishing a statistically and substantively robust measurement model – that is,
the focus during Step 1 is on the pattern (intra-factor) relationships. Ultimately,
however, during Step 2, SEM analysis allows researchers to test hypotheses
related to both the pattern (intra-factor) and structural (inter-factor) relations
simultaneously (Kline 2005: 209).
There is no single test of statistical significance that can be used to indicate a
correct model in structural equation modelling (Schumaker and Lomax 2004: 81).
A considerable number of fit indices have been developed for use in SEM, and