Figure 7.3 presents the final measurement model of collaborative freight distribution construct.
The standardised loading, composite reliability, Cronbach alpha and AVE results are presented in the Table 7.9-7.12 and the value of the figure is rounded up by AMOS version 21. Partner selection dimension consists of observed variables PS_7.5, PS_7.6, PS_7.9, PS_7.10. These observed variables are shown to exhibit convergent validity criterion since the standardised loadings are greater than the threshold value of 0.5 (0.58 < β < 0.71) (p<0.01), and construct validity with the value of CR (.73) greater than the value of AVE (.45) (see Table 7.9 and 7.11). Moreover, they demonstrate discriminant validity, since they are strongly correlated with the partner selection (PS) dimension while having weaker correlation with other dimensions (see Table 7.10), with covariance ranging between 0.61 and 0.67. These observed variables are reliable, since their SMC is greater than the minimum threshold of 0.3 (0.34 < SMC < 0.5) (see Table 7.9). Moreover, they are reliable because the Cronbach’s alpha is .74, composite reliability is .73, and AVE is .45 (see Table 7.11).
The benefits and risks sharing dimension consists of two observed variables BR_8.5 and BR_8.7, which ideally cannot explain the factor. However, Kline (2005) posits that if a standard CFA model with a single factor has at least three indicators, the model is identified. If a standard model with two or more factors has at least two indicators per factor, the model is identified. Therefore, two-item factor is not an issue. These observed variables are shown to exhibit convergent validity, since the standardised loadings exceed the threshold value of 0.5 (0.69 < β < 0.71) (p<0.01), and construct validity with the value of CR (.66) greater than the value of AVE (.51) (see Table 7.9 and 7.11). Moreover, they meet the discriminant validity criterion, since they are clearly clustered into their respective dimensions (see Table 7.10) with covariance varies between 0.61 and 0.85. These observed variables are reliable, since their SMC is greater than the minimum threshold of 0.3 (0.47 < SMC < 0.51), as well as because the Cronbach’s alpha is .66, composite reliability is .66, and AVE is .51 (see Table 7.9 and 7.11).
The advanced information technology dimension consists of four observed variables IT_9.3, IT_9.4, IT_9.7 and IT_9.8. These observed variables exhibit convergent validity as their respective standardised loadings are greater than the threshold value of 0.5 (0.60 < β < 0.83) (p<0.01), and construct validity with the value of CR (.81) greater than the value of AVE (.52) (see Table 7.9 and 7.11). Moreover, they meet the discriminant validity criterion, since they are clearly clustered into their respective dimensions with covariance varies between 0.67 and 0.85
179 (see Table 7.10). These observed variables are reliable, as their SMC is greater than the minimum threshold of 0.3 (0.36 < SMC < 0.70), as well as because the Cronbach’s alpha is .83, composite reliability is .81, and AVE is .52 (see Table 7.9 and 7.11).
Note: PS= partner selection, BR = benefits and risks sharing, IT = advanced information technology
Figure 7.3: Standardized estimates for collaborative freight distribution construct
180 Table 7.9: Standardized factor loading, squared multiple correlation and p value of
collaborative freight distribution construct Collaborative Freight Distribution
Partner selection
Question items Item descriptions Standardised
Loading**
Squared Multiple Correlation
P-value
PS_7.5 You are willing to assess and evaluate your partner’s goals/objectives before choosing the partner.
0.65 0.42 0.001
PS_7.6 You consider complementary skills of your partner, e.g., partner’s experience, capabilities, and potential for making real contribution, when choosing an alliance partner.
0.58 0.34 0.001
PS_7.9 Peer relationship between the top executives of you and your partner’s firm must be established.
0.58 0.34 0.001
PS_7.10 You are willing to learn a new working environment.
0.71 0.50 0.001
Benefits and risks sharing
Question items Item descriptions Standardised
Loading**
Squared Multiple Correlation
P-value
BR_8.5 You will implement collaborative freight distribution, if it is going to improve sales of you and your partner’s firm.
0.71 0.51 0.001
BR_8.7 You will implement collaborative freight distribution, if it is going to improve on-time delivery of you and your partner’s firm.
0.69 0.47 0.001
Advanced Information Technology
Question items Item descriptions Standardised Squared P-value
181 Loading ** Multiple
Correlation IT_9.3 You are going to implement market-based
system (i.e., hubs, portals)
0.64 0.41 0.001
IT_9.4 You are going to implement collaborative planning and forecasting-based systems (i.e., CPFR)
0.60 0.36 0.001
IT_9.7 You will implement information technology, if it is going to improve service levels, e.g., higher on-time performance, of you and your partner’s firm.
0.83 0.70 0.001
IT_9.8 You will implement information technology, if it is going to increase visibility, e.g., identifying location of freight in the supply chain, of you and your partner’s firm.
0.76 0.58 0.001
Achieved Fit Indices
Chi-square =43.50, Degrees of Freedom = 30, P = 0.05, Bollen-Stine p value = 0.16,
CMIN/DF = 1.45, GFI = 0.96, AGFI = 0.93, NFI = 0.95, TLI = 0.97, CFI = 0.98, RMSEA = 0.04 Note: ** Statistically significant at p < 0.01 (two-tailed)
182 Table 7.10: Correlations of measurement items and sub-constructs under collaborative freight
distribution construct
IT BR PS
IT 1.000
BR 0.85 1.000
PS 0.67 0.61 1.000 c9.3 0.64 0.55 0.43 c9.4 0.60 0.51 0.41 c9.7 0.83 0.61 0.56 c9.8 0.76 0.65 0.51 c8.5 0.60 0.71 0.44 c8.7 0.59 0.69 0.42 c7.5 0.44 0.40 0.65 c7.6 0.39 0.36 0.58 c7.9 0.39 0.36 0.58 c7.10 0.48 0.43 0.71
Table 7.11: Validity and reliability test of collaborative freight distribution construct Cronbach’s alpha
(α)
Composite reliability (CR)
Average variance extracted (AVE) Collaborative
Freight Distribution
0.85 0.89 0.50
PS 0.74 0.73 0.45
BR 0.66 0.66 0.51
IT 0.83 0.81 0.52
183 Based on the above findings, it is evident that partner selection, benefits and risks sharing, and advanced information technology dimensions are reliable and valid for collaborative freight distribution construct, since the composite reliability is .89, Cronbach’s alpha is 0.85, and AVE is 0.50 (see Table 7.11). Moreover, the Pearson’s correlations between dimensions are less than 0.9 (0.61 < r < 0.85) which indicates discriminant validity and unidimensionality (Table 7.10).
Referring to Table 7.12, all measurement dimensions also demonstrate discriminant validity, since their chi-square differences are significant. The measurement model fits the data very well, since the Chi-square =43.50, degrees of freedom = 30, p value = 0.05 (Bollen-Stine p value = 0.16, which is not significant at the 0.05 level). Other fit measures also indicate the goodness of fit of the model (CMIN/DF = 1.45, GFI = 0.96, AGFI = 0.93, NFI = 0.95, TLI = 0.97, CFI = 0.98, RMSEA = 0.04) (Table 7.9).
Table 7.12: Chi-square difference test of collaborative freight distribution 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
PS & BR 7.89 7 839.74 8 831.85 1 0.000 Significant
PS & IT 21.00 17 667.47 18 646.50 1 0.000 Significant
BR & IT 15.44 7 544.86 8 529.42 1 0.000 Significant