**3. Findings and discussion**

**Table 2** reveals the descriptive analysis. From the analysis, it was revealed what the respondents thought of enablers of collaboration. The respondents agreed that relationship building and management for implementing collaboration with a mean score of 4.05, resource investment and development in the reverse supply chain with a mean of 3.98, which reveals that the respondents agreed with the extent to which this construct enable collaboration. The respondents revealed with a mean of 3.97 that free information flow in the organisation enables collaboration, this means that the respondents agreed with the variables. The respondents revealed with a mean of 3.92 that Internal alignment of the organisation and partners to the implementation of collaboration, which means that the respondents agreed with the variable. Lastly, top management support for collaboration revealed that the respondents agreed with the variable with a mean of 3.79, this means that the respondents agreed with it.

Furthermore, a descriptive analysis for collaboration practices revealed that quick response on returned goods revealed that the statement was agreed with a mean of 4.06, rapid processing of order returns revealed that a mean of 4.01 which means that the respondents agreed with the statement. Information sharing with suppliers on the returned products revealed that that the statement was agreed by the respondents with a mean of 3.99. Joint knowledge creation among the stakeholders on reverse supply chain revealed that the respondents agreed with the statement with a mean of 3.98. lastly, the respondents agreed with the statement close relationship with customers who purchase the products revealed that the respondent agrees with the statement with a mean of 3.71.

**Table 3** reveals the convergent validity and internal consistency of the collaboration construct. The threshold of average variance extracted (AVE) is above 0.5, as recommended by Fornell and Larcker, Hair et al. [81, 82]. The composite reliability (CR) threshold as recommended by Litwin [83] is 0.5, but Fornell and Larcker [81]


#### **Table 2.**

*Descriptive statistics for collaboration.*


#### **Table 3.**

*Convergent validity and internal consistency.*

recommended 0.7. The AVE finding of this study is above 0.5, meeting the cut-off criteria. The CR for this construct is 0.96 because they show that all the indicator variables measure the same phenomenon [77].

Furthermore, the regression weights of the variables measuring collaboration showed that they were all significant, which means that they accurately measured collaboration.

Although several measures for deciding the fitness of a model exit. Hu and Bentler [84] suggested that use of the ML-based standardised root mean squared (SRMR)

