A model for investigating the stability factors in formal science and technology collaborative networks: A case study of Iran

- A tentative model is proposed to study the stability of formal S&T networks.
- The model classifies network stability factors into two internal and external groups.
- Confirmatory Factor Analysis (CFA) confirmed the proposed model's goodness of fit.
- Fostering collaborative ecosystem will strengthen network stability.
- Alignment of network activities with national priorities contributes to its stability

In spite of vast studies performed on Science and Technology Collaborative Networks (STCNs) in recent decades, factors affecting the stability of these networks have rarely been investigated. In this study, 6 collaborative networks established between 1995 and 2005 in Iran, in the fields of nanotechnology, biotechnology, molecular medicine and herbal medicine, were evaluated. Conducting a qualitative research, and using multiple-case study method, we collected related data adequately from in-depth interviews with key informants, survey, and archived documents, and identified the stability factors of STCNs. In this paper, a tentative conceptual model is proposed for classification of stability factors in formal STCNs. Employing this model, 27 identified and screened stability factors were classified into 2 main categories; including internal factors (network management, network formation processes, collaboration mechanisms in the network and characteristics of network members) and external factors (collaboration infrastructures and network environment). Designed model was evaluated by conducting an extensive survey which provides feedbacks from the representatives of network members and other informed people on the primary model. In total, 112 individuals from 83 affiliated organizations of 6 selected networks participated in this survey. Confirmatory factor analysis (CFA) method was used to evaluate the goodness of fit of the proposed model which indicated that each structural component of the model has suitable fit, separately. In addition, the whole model offers a good fitting and it can be reliably used to achieve the research goal only by omitting two of considered factors.