You are on page 1of 1

a poor fit to the data and needed modification as the cut-off ranges of fit indices are

below the recommended levels with a high χ2 value of 17.368 (df=5 and p = .004) and
with unreasonable CMIN/DF and RMSEA. Expected change statistics of error
covariances suggested misspecification associated with ‘item 4’ and ‘item 5’ and
measurement model fit statistics showed relatively poor fit.

Table 5.7
Summary of Initial Findings (CFA): Communication
Quest. Item wording Initial Final
Items Standardised Standardised C.R.
Loadings Loadings (t)
Comm1 We keep our supplier informed about
changes in our business
.58 .58 8.68
Comm2 Our major supplier and our company
exchange those information that may
.68 .71 11.08
benefit both
Comm3 This supplier and our company keep
each other informed about events and
.79 .82 12.97
changes in the market
Comm4 Our major supplier frequently
discusses and informs us about new
.71 .65 9.89
possibilities for business development
Comm5 Our major supplier informs us
immediatelyMd._Abu_Saleh_Thesis 2006
if any problem arises
.68

Achieved Fit Indices


CMIN/DF
RMSEA IFI TLI CFI
(χ2/df)
3.474
Initial .103 .967 .933 .966
(17.37/5)
Final 1.795 .059 .994 .981 .994
(3.59/2)
Composite Construct Reliability .69

In terms of ‘item 5’ (‘our major supplier informs us immediately if any problem arises’), it
seemed not to be an adequately perceived aspect/factor of communication in the
importer supplier commitment relationship in the present context. However, while
‘item 5’ exhibited an acceptable standardised loading of .68, this item also showed
relatively weak error association with ‘item 4’ of this measure. As a result, upon
deleting of communication ‘item 5’, the better fitted model was then identified with
reduced χ2 value from 17.368 to 3.589 (df=2 and p =.166) and all other fit indices
showed substantial enhancement to the overall fit to the model (Table 5.7). Although

111

You might also like