دانلود رایگان مقاله ثبات شبکه اتصال و نوآوری خروجی

abstract

With the patent co-inventing data of top 9 ICT firms with the highest patent application in China, this study establishes the co-inventing network and examines the moderate role of network connectivity, measured by classifying the individuals into two cohorts: inventors in the largest connected component and inventors in other isolated components. The network stability and innovation output demonstrate strong positive interaction, which is significant in not only the largest but also other isolated components. The clustering and centrality demonstrate significant effect on network stability and innovation output in the largest connected component, which is generally the same as that of extant studies. This impact is not significant in the other isolated components, which con- firms the moderate role of network connectivity, i.e., fully connected networks constitute the basis for the network structure to be functioning. However, the significantly positive role of the structural hole is not moderated by the network connectivity. The contributions and implications of our findings is discussed at the end of this study.

5. Discussions and limitations

Since the firms in our data file more patents in USPTO and EPO, the networks with only Chinese patents may only be partly representative of the true network structure, which may lead to a biased empirical result. However, we may preclude this case by making a comparison with other relevant literatures: 1st. There is an inverted U relationship between clustering and patent output, which not only corresponds with Zhang et al. (2014a, b), Chen and Guan (2010) who used either the country level or regional level patent co-inventing data, but also corresponds with Uzzi and Spiro (2005) who used the Broadway musical data. However, Fleming et al. (2007) found an insignificant relationship with the patent co-inventing data; 2nd. The closeness centrality, which is the inverse of the average path length, contributes positively to the patent output. This corresponds with Zhang et al. (2014a, b), Chen and Guan (2010), Fleming et al. (2007); 3rd. The structure hole contributes positively to patent output, which corresponds with Uzzi and Spiro (2005) who also used the individual networked data. However, this contradicts with Schilling and Phelps (2007) who used the inter-firm data. This may be because the negative effect of the opportunism, as is summarized by Schilling and Phelps (2007), brought by the structural holes plays the main role in the inter-firm level network, while the positive effect of structural hole led explicit inter-individual knowledgesharing plays the main role in the inter-individual level network. Similar differences of the betweenness centrality can also be found in this study and Schilling and Phelps (2007). As our results are generally similar, or at least not contradict with extant literatures, particularly the literatures with the individual networked data, we may conclude that our data is not biased by the unique characteristics of Chinese patents. Therefore, the network effect on inventor stability disclosed by this study is also robust against the single firm sampling.