With the increasingly contribution of S&T to the economy development, the gender disparity in patent productivity and its influencing factors have attracted more and more attention as the low patent productivity of female researchers would enlarge the gender gap in scientific field and affect the innovation efficiency of the society. Based on a survey of researchers from institutes of CAS conducted in 2018, this study focuses on the collaboration network size and ties strength of researchers to understand the gender gap in patent productivity using variance analysis and regression analysis. The results show that:

Firstly, compared to male researchers, female researchers are at a disadvantage both of collaboration network size and ties strength. Specifically, female researchers have fewer collaborators from industries, government departments, other research institutes and interdisciplinary fields, and the communication frequency with partners of female researchers is lower than that of male researchers.

Secondly, the disadvantages at collaboration network size and ties strength are important reasons for the gender difference of patent productivity, but not all types of network deficiency have an impact on female patent productivity. The deficiency of collaboration network size and ties strength with industries and interdisciplinary fields are the key factors affecting female researcher′s patent productivity. Besides, compared with network size, it seems that ties strength has a greater influence on patent productivity. This means that the female researchers not only need more partners from industries and interdisciplinary fields, more importantly, they need to establish a relatively close communication mechanism with those partners.

Thirdly, gender does not play an obvious role in regulating the collaboration network and patent productivity as there is no significant difference in the return of collaboration network between male and female researchers in this research. In other words, it means that there is no obvious network return deficiency caused by gender bias in the famous research institutions of CAS. 

The limitation of this study is that, due to the researchers investigated mostly are from CAS, the universality of the conclusion is still worth further discussion. Previous explanations of "the puzzle of academic productivity" suggested that women are far from top research institutes might be one of the reasons of the gender difference in academic productivity, while in this study, the influence of institute′s reputation has been excluded. For the research groups in China′s well-known scientific institutes, they tend to follow the norms of "universalism" of the scientific community. The "return deficiency" of social networks caused by gender bias is not obvious, but the situation may be different considering all the researchers at the national level. 

However, the findings of this study are still of some implications for the gender policy in the S&T in China. First of all, it′s recognized that lack of social network and resources of female researchers is an important reason for their low innovation output, such as patents. It′s precisely because of their lack of social network and resources that they are more rely on formal organizational support or formal work experience. As shown in this study, the supports from institutes and early industry work experience have greater influence on female researcher′s patent productivity than that of male researchers. Therefore, the institutes should play an important role in improving the female researchers′ collaboration network through building platforms which providing learning, training and information resources, and creating opportunities for their cooperation with the industries and interdisciplinary fields.