Following this comparison, Model 5b is accepted as the optimal model based on
the following criteria:
there is little difference in adequacy of model fit measures between the two
models;
Model 5b it is more representative, making use of all 213 valid cases (in
comparison with 209 cases in Model 5); and
Model 5b has a smaller ECVI values, indicating that it has a greater potential
for cross-validation in an independent sample.
7.9.1 Employee attitudes and discretionary service behaviour
Section 4.4 above described (i) how Simons and Roberson (2003) measured the
Affective Commitment → DSB path in their structural equation model and (ii)
noted that this relationship may not be theoretically robust. Specifically, it not
clear how an individual’s affective organisational commitment (an individual’s
attitude) can influence the behaviour of colleagues (DSB was measured by
Simons and Roberson, in the same as in this research, as a respondent’s
assessment of their colleagues’ service-orientated extra effort).
As a matter of interest, the Model 5b structural model was respecified (as SEM
5b:3, see Figure 7 - 23 ) to include the EPA→DSB path (which reflects Simon’s and
Roberson’s Affective Commitment → DSB path insofar as individual attitudes are
being hypothesised as predictors of colleagues’ behaviours). The path ( =
0.255) was found to be significant at the 0.001 level and a χ^2 difference test found
that, in comparison with Model 5b:2, model fit was improved by a statistically
significant amount (Δχ^2 p = 0.005).
One theoretical rationale that can be employed to underpin the practical (rather
than statistical) existence of the EPA→DSB path is that, employees who
experience motivational leadership and are concomitantly higher on positive work
attitudes (EPA) have colleagues who experience the same leader behaviour and,
as a consequence exhibit extra effort. This argument falls down, however, when
we consider that this effect is already being measured (more directly) by the
existing JP→DSB path that essentially posits the same premise – i.e. that
respondents’ colleagues experience the same type of leader behaviour and that
this has a positive effect on their performance. Indeed, the JP→DSB path in the
original SEM 5b:2 drops from = 0.403 to = 0.245 in SEM 5b:3. In effect,