leadership and motivation in hospitality

Although all factor loadings and covariances are statistically significant, the model
diagnostics indicate a model that does not fit adequately well: χ^2 is significant;
RMSEA is not satisfactory at 0. 084 (also the upper limit is 0.096) and the pclose
value of 0.000 indicates that the RMSEA estimate is not robust; CFI is not
satisfactory at <0.96; SRMR is satisfactory at <0.09; and CN is not satisfactory at
<200. The AVE and CR values will be estimated when the CFA has been modified
and has satisfactory fit.

Following the same procedure described above for Model 1, the factor loadings
and standardised residual covariances are checked for CFA 2:1. Table 7 - 10
summarises the modifications made. Once again, DSB4 shares a problematic
standardised residual covariance with JP4 and, once again, DSB4 is removed and
JP4 retained.

Item Loading Standardised
residual covariances
(SRCs)

Action taken / justification

1.96 >2.58 >4

ML4 0.613 1 3 0 Removedloading + four excessive^ from model / moderate SRCs

ML5 0.615 1 0 0 Removedloading + one^ from modelexcessive^ / moderate SRC

ME5 0.610 2 2 0 Removedloading + four excessive^ from model / moderate SRCs

ME7 0.303 4 0 0 Removedloading from model / very low

JP3 0.556 1 1 0 Removedloading +^ two excessive from model / moderate SRCs

JP4 0. 493 2 2 0 Retained owing to substantive interest

DSB4 0.696 0 1 0

Removed from model / one SRCs

>2.58 is shared with JP4; because JP4

is substantively more important than

DSB4, JP4 is retained

Table 7-10 Modifications to CFA 2:1

The measurement model is respecified without ML4, ML5, ME5, ME7, JP3 and
DSB4. The new set of estimates and fit diagnostics are described below in Table
7 - 11.