correlation estimates between these constructs are: ME AOC = 0.880; ME JS
= 0.892; and AOC JS = 0.823).
These three constructs, ME, JS and AOC, with the highest squared correlation
estimates (and therefore having the strongest correlations with each other) are
the three attitudinal variables, Work Meaning, Job Satisfaction and Affective
Organisational Commitment. On the one hand, the finding that these constructs
are not empirically distinct is somewhat disappointing – on the other hand, the
fact that these constructs are all located within the Employee Attitudes domain
does provide a theory-based rationale for their convergence.
Following the empirical findings and their theoretical underpinnings, the research
goes on to respecify Model 3 with all of the Employee Attitude indicators loading
on one construct and to develop a new model (Model 3b) from this starting point.
7.6 Model 3b
The structural part of Model 3b is specified as illustrated in Figure 7 - 11 and the
measurement model is specified with the indicator variables for Work Meaning (5
items), Job Satisfaction (6 items) and Affective Organisational Commitment (3
items) all loading on a single Employee Attitudes (EA) construct (see Table 7 - 18 ).
Two new hypotheses H 21 and H 22 are generated to replace the original hypotheses
H 3 to H 8.
Figure 7-11 Structural specification for Model 3b
The exact description of H 21 and H 22 cannot be formulated until such a time as the
new EA construct has been successfully measured as a distinct construct and its
identity determined through the interpretation of the indicators that load on the