First, a confirmatory measurement model that specifies the
posited relations of the observed variables to the underlying
constructs, with the constructs allowed to intercorrelate freely,
was tested. The adequacy of the individual items and the
composites were assessed by measures of reliability and validity.
The composite reliability as calculated with LISREL estimates, is
analogous to coefficient alpha and is calculated by the formula
provided by Fornell and Larcker (1981). Further, convergent
validity and discriminant validity were assessed. Convergent
validity was assessed from the measurement model by determining
whether each indicator’s estimated pattern coefficient on its
posited underlying construct factor is significant (greater than
twice its standard error). Discriminant validity was assessed for
every possible pair of constructs by constraining the estimated
correlation parameter between them to 1.0 and then performing a
x2 difference test on the values obtained for the constrained and
unconstrained models (Anderson and Gerbing, 1988; Joreskog,
1993). A significantly lower x2 value in an unconstrained model
indicates that discriminant validity is achieved.