6. Limitations and conclusion
This study has a number of limitations. It is based on a survey sample of UI research collaborations from three industries in one country, suggesting that caution is warranted when applying its findings to different contexts (e.g., collaborations between firms or collaborations in Western countries). Furthermore, we are relying on cross-sectional data and therefore cannot capture the dynamics of trust formation in UI research collaborations in a longitudinal sense. The cross-sectoral nature of this study does not provide conclusive evidence of the direction of causality. Our inferences concerning reciprocal communication and trust in the intermediate stage and concerning decision process similarity and trust in the mature stage are based on our research design that emphasized the temporal ordering of the trust bases in our survey and took great care in their formulation and measurement to limit endogeneity. However, this limitation does not invalidate the inherent causal nature of our conceptualization (Whetten, 1989). Finally, whereas we have worked in various ways towards limiting the potential for common method variance, its existence cannot be strictly ruled out. Given the contributions and limitations of our study, further research on UI research collaborations based on dyadic data analysis can validate our findings and deepen the understanding of trust building processes in such collaborations. In particular, longitudinal studies appear to be promising avenues for future studies in order to capture the nature of trust formation as an inter-organizational relationship progresses. Additionally, contrary to prior research (Levin et al., 2006), our findings show that relationship maturity does not moderate the relationship between demographic similarity and trust. One reason for these differences could be how demographic similarity was measured in the two studies (gender versus education) and the context for the studies (intra-organizational versus inter-organizational; Anglo-Saxon countries versus an East Asian country). Thus, future research should examine specific dimensions of demographic similarity within various contexts to ascertain if demographic similarity is associated with trust and if relationship maturity moderates this relationship.