RMSEA = 0.032; CFI = 0.987; and SRMR = 0.0370) excepting for the chi square
p value at 0.040. However, because the Social Support construct did not covary
at a statistically significant level with the Job Performance construct (p = 0.227)
the model was not suitable for further development as a structural model.
Because of this lack of statistically significant covariance between Social Support
and Job Performance, the hypothesis (H 15 ) that employees experiencing higher
levels of Social Support will also experience higher levels of Job Performance
cannot be tested and is effectively not confirmed.
The second hypothesis involving Social Support, H 16 (Social Support→Service
Quality) stating that employees experiencing higher levels of Social Support will
also report greater success in maintaining customer satisfaction in the presence of
service failures, is also effectively not confirmed because the Service Quality
construct does not covary at a statistically significant level with any other
construct in the model. This finding is reported below in Section 7.11 where the
introduction of the Service Quality (SQ) construct is described.
7.11 Model
The Service Quality (SQ) construct is defined as ‘Employee ability to maintain
satisfied customers in the face of adverse service conditions’ and is measured
using an exploratory seven item scale developed from earlier research findings by
Bitner et al. (1994) that compared customer and staff perceptions of satisfaction
or dissatisfaction in critical service encounters. The SQ construct is hypothesised
as an outcome of Social Support (H 16 ), Affective Organisational Commitment
(H 17 ), Job Satisfaction (H 18 ), Work Meaning (H 19 ) and Motivational Leadership
(H 2 o).
To incorporate SQ into the existing model (Model 5b), a CFA model was specified
containing all indicator variables for the constructs ML (Motivational Leadership),
EM (Empowerment), EPA (Employee Positive Attitudes), JP (Job Performance) and
DSB (Discretionary Service Behaviour). The data set used to estimate this model
contained 122 responses which was the number of cases where all of the seven
SQ items were completed - recalling that for the measurement of the SQ
construct, a non-response option was made available for individuals who never
experienced service failure in a particular critical service encounter (Section 5.4).
The CFA model was estimated with the following results:
