leadership and motivation in hospitality

Construct Item Standardised factor loading estimates
ML ME JS AOC JP DSB

Motivational
Leadership

ML1 .905

ML2 (^) .951
ML3 .880
Work Meaning
ME1 .876
ME3 (^) .867
ME6 .800
Job Satisfaction
JS1 .778
JS5 .740
JS6 .891
Affective
Organisational
Commitment
AOC1 (^) .828
AOC2 (^) .851
AOC4 .882
Job Performance
JP1 .882
JP2 .844
JP4 .456
Discretionary
Service Behaviour
DSB1 .950
DSB2 .683
Average
variance
extracted (AVE)
0.768^ 0.745^ 0.684^ 0.764^ 0.618^ 0.734^
Construct
reliability (CR) 0.909 0.898 0.866 0.907 0.819 0.843
Model fit
statistics
χ^2 = 144.838; d.f. = 104; sig = 0.005
RMSEA = 0.043 (0.059; 0.024; pclose = 0.748)
CFI = 0.9 84
SRMR = 0.0378
CN (0.05) = 189

Table 7-17 Estimates for CFA 3:4

Before moving on to estimate the structural model, it is necessary to confirm that
the measurement model complies with the requirements for convergent and
discriminant validity. As in previous models, this is done by checking the values
for construct AVEs (Average Variance Extracted), CR (Construct Reliability) and
the squared correlation estimates for the model.

Examining these values in Table 7 - 17 it can be seen that all of the CR values are
satisfactory at > 0.7. The lowest AVE values are 0.618 for JP and 0.684 for JS
and these are lower than the highest squared correlation estimates (ME  AOC =
0.774; ME  JS = 0.796; AOC  JS = 0.677). As described in Section 7.2.5,
where any squared correlation estimates are greater than the lowest AVE values
in a model, then the constructs are failing to measure truly distinct factors.
Kline’s (2005: 60, 73) advice that correlations between latent factors in the region
of >0.85 and >0.90 are indicative of constructs that are not sufficiently distinct
from one and other also flags these constructs as problematic (the [non-squared]