Figure 7.5 presents the final measurement model of all research constructs including co- opetition (COOP), freight consolidation (FREIGHT), collaborative freight distribution (COFREIGHT), and sustainable distribution (SUS). The model was produced to test validity and reliability of measurement dimensions and research constructs prior to use for structural modeling. In the model, the observe variables from previous analysis were aggregated to the measurement dimension level (Byrne, 2009). For example, management commitment (MC) dimension is derived from an aggregation of observed variables MC_1.5 and MC_1.6. The standardised loading, composite reliability, Cronbach alpha and AVE results are presented in the Table 7.17-7.20 and the value of the figure is rounded up by AMOS version 21.
Co-opetition construct consists of management commitment (MC), relationship management (RM), and communication management (CM). These measurement dimensions are shown to exhibit convergent validity since their standardised loadings are greater than the threshold value of 0.5 (0.61 < β < 0.78) (p<0.01), and construct validity with the value of CR (.75) greater than the value of AVE (.50) (see Table 7.17 and 7.19). Moreover, they meet the discriminant validity criterion, as clustered into their respective construct (see Table 7.18) with covariance varies from .41 to .48 and within the threshold values 0.80. These measurement dimensions are reliable, since their SMC exceeds the minimum threshold of 0.3 (0.38 < SMC <
0.61) (see Table 7.17). Moreover, they are reliable, as the Cronbach’s alpha is 0.74, composite reliability is 0.75, and AVE is 0.50 (see Table 7.19).
Freight consolidation construct consists of location of freight consolidation centre (LO), geographical coverage (GC), and utilization of transport modes (UT). These measurement dimensions are shown to exhibit convergent validity since their standardised loadings are greater than the threshold value of 0.5 (0.79 < β < 0.84) (p<0.01), and construct validity with the value of CR (.85) greater than the value of AVE (.66) (see Table 7.17 and 7.19). Moreover, they meet the discriminant validity criterion, as clustered into their respective construct (see Table 7.18) with covariance varies between 0.40 and 0.48. These measurement dimensions are reliable, since their SMC exceeds the minimum threshold of 0.3 (0.63 < SMC < 0.70) (see Table 7.17). Moreover, they are reliable, as the Cronbach’s alpha is 0.85, composite reliability is 0.85, and AVE is 0.66 (see Table 7.19).
Collaborative freight distribution construct consists of partner selection (PS), benefits and risks sharing (BR), and advanced information technology (IT). These measurement dimensions are
190 shown to exhibit convergent validity, since their standardised loadings are greater than the threshold value of 0.5 (0.56 < β < 0.87) (p<0.01), and construct validity with the value of CR (.78) greater than the value of AVE (.54) (see Table 7.17 and 7.19). Moreover, they meet the discriminant validity criterion, as clustered into their respective construct (see Table 7.18) with covariance varies between 0.44 and 0.54. These measurement dimensions are reliable, since their SMC exceeds the minimum threshold of 0.3 (0.31 < SMC < 0.76) (see Table 7.17).
Moreover, they are reliable, as the Cronbach’s alpha is 0.77, composite reliability is 0.78, and AVE is 0.54 (see Table 7.19).
Sustainable distribution construct consists of economics (EC), social (SO), and environmental (EV) factor. These measurement dimensions are shown to exhibit convergent validity, since their standardised loadings are greater than the threshold value of 0.5 (0.60 < β < 0.87) (p<0.01), and construct validity with the value of CR (.76) greater than the value of AVE (.52) (see Table 7.17 and 7.19). Moreover, they meet the discriminant validity criterion, as clustered into their respective construct (see Table 7.18) with covariance varies between 0.40 and 0.54.
These measurement dimensions are reliable, since their SMC exceeds the minimum threshold of 0.3 (0.36 < SMC < 0.76) (see Table 7.17). Moreover, they are reliable, as the Cronbach’s alpha is 0.76, composite reliability is 0.76, and AVE is 0.52 (see Table 7.19).
191 Note: COOP = co-opetition, FREIGHT = freight consolidation, COFREIGHT = collaborative freight consolidation, SUS = sustainable distribution
Figure 7.5: Standardized estimates for research constructs
192 Table 7.17: Standardized factor loading, squared multiple correlation and p value of research
constructs Co-opetition
Measurement dimension
Item descriptions Standardised
Loading **
Squared Multiple Correlation
P-value
MC Management Commitment 0.62 0.38 0.001
RM Relationship management 0.78 0.61 0.001
CM Communication management 0.71 0.51 0.001
Freight consolidation Measurement
dimension
Item descriptions Standardised
Loading **
Squared Multiple Correlation
P-value
LO Location of freight consolidation centre 0.81 0.65 0.001
GC Geographical coverage 0.84 0.70 0.001
UT Utilization of transport modes 0.79 0.63 0.001
Collaborative freight consolidation Measurement
dimension
Item descriptions Standardised
Loading **
Squared Multiple Correlation
P-value
PS Partner selection 0.56 0.31 0.001
BR Benefits and risks sharing 0.74 0.55 0.001
IT Advanced information technology 0.87 0.76 0.001
Sustainable distribution Measurement
dimension
Item descriptions Standardised
Loading **
Squared Multiple Correlation
P-value
EV Environmental factor 0.60 0.36 0.001
EC Economic factor 0.87 0.76 0.001
SO Social factor 0.67 0.44 0.001
193 Achieved Fit Indices
Chi-square = 58.62, Degrees of Freedom = 48, P = 0.14, CMIN/DF = 1.22, GFI = 0.96, AGFI = 0.93, NFI = 0.94, TLI = 0.98, CFI = 0.99, RMSEA = 0.03
Note: ** Statistically significant at p < 0.01
Table 7.18: Correlations of sub-constructs of research constructs
COOP FREIGHT COFREIGHT SUS
COOP 1.000
FREIGHT 0.48 1.000
COFREIGHT 0.48 0.44 1.000
SUS 0.41 0.40 0.54 1.000
MC 0.61 0.29 0.30 0.21
RM 0.78 0.37 0.37 0.27
CM 0.71 0.34 0.34 0.24
LO 0.39 0.81 0.35 0.33
GC 0.40 0.84 0.37 0.34
UT 0.38 0.79 0.35 0.32
PS 0.27 0.25 0.56 0.30
BR 0.36 0.33 0.74 0.40
IT 0.42 0.38 0.87 0.47
EV 0.20 0.24 0.32 0.60
EC 0.30 0.35 0.47 0.87
SO 0.23 0.27 0.36 0.67
Table 7.19: Validity and reliability test of research construct Cronbach’s alpha
(α)
Composite reliability (CR)
Average variance extracted (AVE)
COOP 0.74 0.75 0.50
FREIGHT 0.85 0.85 0.66
COFREIGHT 0.77 0.78 0.54
SUS 0.76 0.76 0.52
194 Based on the above result, COOP, FREIGHT, COFREIGHT, and SUS are reliable and valid to be used in structural equation modeling. The Pearson’s correlations between constructs are less than 0.9 (0.40 < r < 0.54), which indicates discriminant validity and unidimensionality (Table 7.18). Referring to Table 7.20, all constructs also demonstrate discriminant validity, since their chi-square differences are significant. The measurement model fits the data very well, since the Chi-square =58.62, degrees of freedom = 48, p value = 0.14. Other fit measures also indicate the goodness of fit of the model (CMIN/DF = 1.22, GFI = 0.96, AGFI = 0.93, NFI = 0.94, TLI
= 0.98, CFI = 0.99, RMSEA = 0.03) (Table 7.17).
Table 7.20: Chi-square difference test of research construct Pairs of
Constructs
χ2 of model 1 (correlation is unconstrained)
df of model 1
χ2 of model 2 (correlation is constrained to 1)
df of model 2
∆ χ2 ∆df p-value Chi-Square Critical Values;
p =0.05 COOP
&FREIGHT
4.73 8 237.71 9 232.98 1 0.000 Significant
COOP &
COFREIGHT
5.28 8 392.64 9 387.36 1 0.000 Significant
COOP &
SUS
9.92 8 307.03 9 297.38 1 0.000 Significant
FREIGHT &
COFREIGHT
8.23 8 506.78 9 498.55 1 0.000 Significant
FREGITH &
SUS
13.44 8 611.80 9 598.36 1 0.000 Significant
COFREIGHT
& SUS
13.91 8 572.03 9 558.12 1 0.000 Significant
In summary, CFA confirmed that all measurement dimensions of each latent factor (constructs) are valid and reliable, and are thus ready to be employed for path analysis using AMOS 21.