The impact of characteristics of industry-university-research cooperation network on innovation performance: mediating role of social capital

Academic circles have a relatively consistent understanding of the connotation of the IUR cooperation network in some aspects. For example, Hadjimanolis believes that the IUR cooperative innovation network is a vertical or horizontal relationship generated by the exchange and interaction of participants [6]. Gai Wenqi and Wang Jici believe that the IUR cooperation network refers to a relatively stable system constructed by participants through a long-term flow of resources and information [7]. This article acknowledges that the IUR cooperation network is a sound value chain system been built by enterprises, universities, and research institutions to jointly resist the risks of social competition and realize the sustainable development.

The characteristics of IUR cooperation network are the extended study of network characteristics. Mitchell was the first to study network characteristics and believed that network characteristics should cover network scale, network structure, and network interaction relations [8]. Li Wenbo and others studied the measurement of network structure. The results showed that the division from the dimensions of network strength, network centrality, and network density has good construct validity [9]. In the study on the characteristics of IUR cooperation network, Zhu Xiangmei pointed out that network location, network intensity, network density, structural hole, centrality and hierarchy should all be taken as the structural characteristics of IUR cooperation innovation network [10]. Hui Qing and Zou Yan studied the characteristics of IUR cooperation innovation network from the network structure and network relationship [11]. Xie Xuemei proposed to analyze the characteristics of collaborative innovation network from the network scale, strength, homogeneity, and openness. Based on the current literature research results, this article describes the characteristics of IUR cooperation network in the oil industry from the three dimensions of network scale, network centrality, and network strength.

Schumpeter was the first to propose the concept of “innovation”, arguing that innovation is a recombination of factors of production whose purpose is to continuously make excess profits [12]. Innovation performance is an extension of the concept of innovation and an index to measure enterprise innovation. Hagedoorn and Cloodt pointed out that innovation performance in the narrow sense mainly refers to the speed at which an enterprise introduces new products, new processes or new equipment into the market, while innovation performance in the broad sense includes the whole process of innovation [13]. Ari believes that after the product is created, it can bring performance to the enterprise, and the whole process can be summarized as innovation performance [14]. Alegre and Chiva believe that innovation performance is the result of new product innovation and process innovation [15]. Gao Jian et al. believe that technological innovation performance represents the efficiency of an enterprise's technological innovation process, the results of its output, and its contribution to business performance [16]. To sum up, this paper believes that innovation performance is the results and benefits that an enterprise's innovation activities can bring to the enterprise [17]. There are many studies on innovation performance measurement methods. The innovation performance indicators commonly used by scholars involve the total number of technology development funds, the number of patents, and the speed of product innovation, etc [18,19,20,21].

Corporate social capital is a tangible or intangible resource that the company can use in the network structure, which can help the company achieve its strategic goals [22]. Corporate social capital is a complete network nesting relationship. In order to possess and master critical tangible or intangible information in business activities, enterprises generate social capital. Burt has studied corporate social capital at an early stage. He believes that the collection of individual social capital forms corporate social capital, which has a great influence in the academic world [23].

Nahapiet and Ghoshal believe that the three aspects of structure, cognition, and relationship can better interpret social capital. Zheng Shengli and Chen Guozhi integrated the views of many scholars and proposed to measure the social capital of enterprises from internal and external resources [24, 25]. This article agrees with the former. Structural social capital represents the number of connected objects that a company accommodates in a social network. Relational social capital represents the closeness and dependence of the company's interaction with the connected objects in the social network. Cognitive social capital usually represents the fact that in a social network, an enterprise and its connected objects share similar vision and value orientation.

The relationship between the characteristics of IUR cooperation network and innovation performance has always been the research focus of many scholars, and it is closely related to resource dependence theory and social network theory.

The theory of resource dependence holds that an organization needs various types of resources for its survival, but it cannot produce all of these resources by itself. Therefore, organizations need to continue to exchange resources with other organizations in the external environment to obtain resources to maintain survival [26]. Innovation resources are often scarce. According to the theory of resource dependence, the ability of a single subject to innovate and acquire resources is limited, and it is necessary to rely on more subjects to form a cooperative contract relationship with each other and form an IUR collaborative innovation network to achieve a higher level of innovation performance.

Network position is one of the characteristics of IUR cooperation network. According to social network theory, network position is an important index to measure the influence and reputation of individuals in the network [27]. Burt and Minor found that enterprises with favorable network positions have more advantages in obtaining scarce resources, which will have an impact on innovation performance [28]. Network centrality is a symbol of an enterprise's central network position, which can well measure the characteristics of an individual in the network position, measure the ability of an individual to obtain information and influence and control other individual economic behaviors in the network [29]. If an enterprise can maintain a good centrality, on the one hand, it can use its core position to better realize information acquisition, knowledge transfer, and resource sharing, thereby effectively developing organizational potential and improving the innovation performance of the enterprise [30]. On the other hand, the centrality of the network itself also conveys a reputation and quality signal, which can help it attract more high-quality or potential partners, so the innovation performance is continuously improved [31].

Network scale indicates the total number of relationships established between an enterprise and other organizations [32]. Network scale determines the abundance of innovation resources that can be obtained by enterprises, and then affects innovation performance [33]. Firstly, the expansion of network scale means more channels for a single entity to obtain knowledge, and also means more scope and degree of information sharing among entities participating in the interaction within the network, thus improving innovation performance [34]. Secondly, as the scale of the network expands, the number of entities that companies can interact with on the platform gradually increases. At the same time, the options for each entity also increase accordingly, which can greatly reduce transaction costs to a certain extent and thus help companies improve their innovation performance [35].

Network strength represents the closeness of network members. Granovetter proposed the tie strength theory in the social network theory and divided the tie strength in the network into strong tie and weak tie. Strong tie represents multiple social relationships with high connection frequency, closeness, mutual trust and information sharing among network members, while weak tie is the opposite [36]. Coleman and Landry et al. found that strong tie can improve innovation performance [37, 38]. This is because the increase in network strength represents the establishment of a high degree of trust and dependence through frequent communication and interaction between various actors, which will greatly help improve the success rate of product innovation. In addition, Chi Renyong studied SMEs in Zhejiang Province in China, and found that the strength of network relationships significantly promoted the increase of corporate net profit, income growth, and product technology innovation [39].

The existing literature shows that the centrality, network scale, and network strength of the IUR cooperation network will promote innovation performance. Therefore, the following hypotheses are proposed:

H₁: There is a positive correlation between the characteristics of IUR cooperation network and innovation performance;

H_1a: There is a positive correlation between the centrality of the IUR cooperation network and innovation performance;

H_1b: There is a positive correlation between the scale of the IUR cooperation network and innovation performance;

H_1c: There is a positive correlation between the strength of the IUR cooperation network and innovation performance.

First of all, there is a “location rent effect” in social capital. That is, the “location rent” is related to the company's position in the IUR cooperation network, and the “location rents” obtained by the entities of different network locations are entirely different. The location of the enterprise in the center of the network means that the network connection objects cover a wide range, the number of network nodes is large, and the “location rent” obtained is relatively large [40]. Secondly, as the scale of the IUR cooperation network continues to expand, the actors of different positions and different industries will join the network, which allows companies to obtain more heterogeneous information from it. The acquisition of such non-redundant information resources can effectively promote the accumulation of corporate social capital [41, 42]. Finally, the participants of the IUR cooperation network have frequent and close interactions and connections in cyberspace. In other words, the continuous increase of network strength can promote the generation of strong-connected social capital. The reason is that the high-strength network relationship formed between the participants can effectively transmit the invisible and complex knowledge they have, which is helpful for the development and learning of existing knowledge in the network, thus enabling the company to have substantial social capital [43]. In summary, the centrality, scale, and strength of the IUR cooperation network will all promote social capital. Therefore, this article proposes the following hypotheses:

H₃: There is a positive correlation between the characteristics of IUR cooperation network and social capital;

H_2a: There is a positive correlation between the centrality of the IUR cooperation network and social capital;

H_2b: There is a positive correlation between the scale of the IUR cooperation network and social capital;

H_2c: There is a positive correlation between the strength of the IUR cooperation network and social capital.

The positive role of social capital in promoting innovation performance has basically reached a consensus. For example, Xu Hailing used an empirical model to explore three aspects of corporate social capital and found that the structure, relationship, and cognition all have a positive role in promoting the performance of IUR collaborative innovation [44]. Bontis and Cleary analyzed data from sample companies in Ireland and verified the social capital has a positive role in promoting corporate innovation performance [45]. However, in many studies on the mechanism of social capital on innovation performance, different scholars have given explanations from various aspects due to different research directions and selected samples. Among them, the most accepted one is explained by transaction costs. They believe that corporate social capital has the economic effect of “transaction cost savings”. For example, Zheng Meiqun pointed out that in the IUR collaborative innovation network, the social capital of enterprises is relatively abundant, and each member maintains a high degree of trust. Then the transactions in this network can be reduced the cost of market exchanges, thus improving the efficiency of the organization's operation, and then improving the innovation performance [46]. Therefore, the following hypotheses are proposed:

H₃: Social capital is improving the innovation performance of petroleum companies.

The previous article analyzed the relationship between the characteristics of IUR cooperation network, social capital, and innovation performance. These relationships indirectly reveal the transformation mechanism of social capital between the characteristics of IUR cooperation network and innovation performance to a certain extent. First of all, knowledge and resources tend to shift from the center. Therefore, enterprises with a high degree of IUR cooperation network can quickly accumulate social capital, promote the inward flow of knowledge and speed up the development of new products. Secondly, with the increase of the scale of the IUR cooperation network, the company will have a more comprehensive grasp of the dynamic changes of external information, which is more conducive to the accumulation of social capital. The higher the social capital is, the more deeply the collaborative innovation network is embedded, which improves the efficiency of the organization and therefore improves the innovation performance of the enterprise [47]. Finally, with the increase in the strength of the IUR cooperation network, enterprises will have easier access to complex and implicit knowledge, which will also improve social capital, thus improving the innovation performance of enterprises [48]. Therefore, the following hypotheses are proposed:

H₄: Social capital has a mediating effect on the relationship between the characteristics of IUR cooperation network and innovation performance of petroleum companies;

H_4a: Social capital has a mediating effect on the relationship between the centrality of the IUR cooperation network and innovation performance of petroleum companies;

H_4b: Social capital has a mediating effect on the relationship between the scale of the IUR cooperation network and innovation performance of petroleum companies;

H_4c: Social capital has a mediating effect on the relationship between the strength of the IUR cooperation network and innovation performance of petroleum companies.

This article adopts structured questionnaire to obtain empirical analysis data. The questionnaire mainly consists of four parts. The first part is the basic information, including demographic variables, such as the year of establishment of the surveyed company and the educational background of the respondents. The second part is the measurement of the characteristics of the IUR cooperation network in the petroleum industry. This article refers to the relevant scales of scholars such as Li Wenbo, Nahapiet J, and Wu Aiqi, and then designs a total of six items with high discrimination. Among them, the network centrality and network strength are measured from five levels (1=completely inconsistent, 2=not very consistent, 3=basically consistent, 4=relatively consistent, 5=completely consistent), The network scale is measured from five levels (1="<5", 2="5–10", 3="11–15", 4="16–20", 5=">20"). The third part is the measurement of social capital. It mainly refers to the maturity scales of Lazega E and Burt et al. A total of three highly distinguished items are designed and divided into five levels to measure (1=completely inconsistent, 2=Not very consistent, 3=basically consistent, 4=relatively consistent, 5=completely consistent). The fourth part is the measurement of innovation performance. It mainly refers to Zhang Xiaohui's measurement method. A total of four items with high discrimination are designed and divided into five levels (1=completely inconsistent, 2=not very consistent, 3=basic Agree, 4=relatively agree, 5=fully agree) to measure. See Table 1 for specific measurement indicators.

Structured questionnaires were distributed to senior managers, technical center leaders, and product managers of A-share listed petroleum companies. A total of 500 questionnaires were distributed by mail and on-site filling, of which 351 valid questionnaires accounted for 83.8% of the 419 received questionnaires. The sample statistics show that 61.83% of the surveyed enterprises have been established for five years or more. 64.93% of the respondents had a bachelor's degree or above; 71.15% of respondents had worked for three years or more; The senior management personnel of the respondents accounted for 74.45%. The above data ensure that the respondents have a good understanding and grasp of the questionnaire, and the survey sample has good representativeness.

According to previous experience, if Cronbach's α coefficient reaches 0.7, which proves that the scale has a high reliability. The validity of the questionnaire includes structural validity and content validity. The structural validity reflects the degree of fitting between the measurement results and the theory. The content validity represents the degree to which the item fits the predicted content. In this study, the scale is revised and shaped according to the existing maturity scale and consulting relevant enterprise personnel so as to make the scale have higher content validity. The test results are shown in Table 2.

The results show that the total α value in this study is 0.901, and the α value of each variable is greater than 0.7, which indicates that the scale has relatively high reliability. The KMO value of each variable is greater than 0.6, and the total KMO value is 0.914. The significance of Bartlett sphericity is less than 0.001. This indicates that factor analysis is suitable.

As shown in Table 3, the communality of each item of social capital and innovation performance is greater than 0.6, indicating a good effect of factor analysis. In addition, the cumulative variance contribution rate of social capital is 66.917%, and that of innovation performance is 74.331%, indicating that the structure validity is good, and one factor is obtained for each.

The Component Matrix after rotation of the characteristics of IUR cooperation network is shown in Table 4.

As shown in Table 4, exploratory factor analysis extracts ten items of the characteristics of IUR cooperation network into three factors (network scale, network centrality, and network strength). These load coefficients are all greater than 0.5, and these communalities are all greater than 0.6. In addition, the cumulative variance contribution rate exceeds 60% and approaches 70%, which indicates that the structural validity of the questionnaire is appropriate.

Before regression analysis, correlation analysis should be used to prove the correlation between variables. The correlation analysis results between the two variables are given in Table 5. The results show that there is a significant positive correlation between the three characteristics of the IUR cooperation network (network scale, network centrality, and network intensity) and social capital, and there is also a significant positive correlation between the three characteristics of the IUR cooperation network and innovation performance. This indicates that regression analysis can be carried out.

This paper takes three characteristics of the IUR cooperation network as independent variables to prove their effect on the innovation performance of petroleum companies. The verification results of the correlation between the characteristics of IUR cooperation network and enterprise innovation performance are shown in Table 6. The results show that network strength has the most significant effect on innovation performance (β =0.418, P<0.001), followed by network centrality (β =0.139, P<0.001), while network scale has no significant effect on innovation performance (β =0.055, P>0.05). In addition, the F value is 115.82 (P<0.001), and it is significant. Therefore, the regression analysis results show that network strength and network centrality significantly promote the innovation performance of petroleum companies, but network scale has no correlation with innovation performance. That is, H_1a and H_1c are supported, while H_1b is not.

This paper takes three characteristics of the IUR cooperation network as independent variables to verify their impact on social capital. The verification results are shown in Table 7. The results show that network strength has the most significant effect on social capital (β =0.524, P<0.001), followed by network centrality (β =0.397, P<0.001), but network scale has no significant effect on social capital (β =0.034, P>0.05). In addition, the F value is 76.62 (P<0.001), and it is significant. Therefore, the regression analysis results show that network strength and network centrality significantly promote the innovation performance of petroleum companies, but there is no linear correlation between network scale and social capital. That is, H_2a and H_2c are supported, while H_2b is not.

This paper uses social capital as an independent variable to explore its impact on innovation performance, and the verification results are shown in Table 8. The results show that the F value in the regression equation model is 136.06 (P<0.001), and it is significant. This shows that social capital has a significant promoting effect on the innovation performance of petroleum companies. Therefore, H₃ is supported.

As it has been proved above that the network scale has no significant promoting effect on innovation performance, network scale is no longer included in the mediating effect regression model. This paper puts the two independent variables of network centrality and network strength and the intermediary variable of social capital into the regression model. The mediating effect test results of social capital are shown in Table 9. The results show that the F value in the regression equation of the mediation effect is 57.131 (P<0.001), and it is significant. Among them, the most significant influence on innovation performance is the social capital (β =0.471, P<0.001), followed by the network strength (β =0.161, P<0.05), and finally the network centrality (β =0.119, P<0.05). In addition, after the addition of the intermediary variable, the t-values of network centrality and network strength are significantly lower than before but still significant. The coefficient of social capital is also significant. The results indicate that social capital has a partial mediating effect in the influence of the centrality and strength of the IUR cooperation network on the innovation performance of petroleum companies. Therefore, H_4a and H_4c are supported, but H_4b is not.

Taking A-share listed companies in the petroleum industry as the research object, this paper empirically studies the relationship among the characteristics of industry-university-research (IUR) cooperation network, social capital, and innovation performance in the petroleum industry. The empirical research results show that the strength and centrality of the IUR cooperation network in the petroleum industry positively affect social capital and innovation performance; Network scale has no significant positive effect on social capital and innovation performance; Social capital has a partial mediating effect on the relationship between the centrality and strength of the IUR cooperation network and innovation performance.

First, petroleum companies should strive to improve their network centrality in the IUR cooperation network. High centrality means the rapid flow of knowledge and information, which can help petroleum companies have more options and more opportunities to acquire innovative knowledge, thus helping them accumulate abundant social capital, and then improve their innovation performance. At the same time, network centrality also reflects the influence and power of petroleum companies in the network, which makes it easier for companies to have strong connections and direct connections with partners, and also easier to occupy the position of structural holes, thus promoting the exchange and cooperation between partners and promoting technological innovation. Therefore, on the one hand, petroleum companies at the edge of the network should take the initiative to establish a stable cooperative relationship with the connected objects with high network centrality; on the other hand, they should gradually improve their position in the IUR cooperation network to improve their advantages in obtaining information resources.

Second, petroleum companies should have high-frequency interaction with the objects of the IUR cooperation network. Limited by their own conditions, it is usually difficult for petroleum companies to obtain accurate information in the dynamic market by themselves. However, the oil market is always unpredictable. Once a company fails to track the latest information, its response speed will lag behind, thus damaging its profits. At this time, if the company can interact with the objects of the IUR cooperation network at high frequency, it will promote the deep communication of knowledge within the network, help the network form a common cognition and vision, ensure the valuable information is exchanged across the network, thus increasing the social capital of companies, and then improving the innovation performance.

Third, while expanding the scale of the IUR cooperation network, petroleum companies should consider matching their own growth stage. Expanding the scale of the IUR cooperation network is an inevitable stage for the growth of petroleum companies. However, the network scale is not always positively correlated with the efficiency of knowledge transfer and the speed of information resource acquisition. First, as the number of objects connected to the company increases, the difficulty of resource management will also increase. If the company does not have the matching resource management ability, it will lose the advantages brought by the expansion of the network scale [49]. Secondly, if the companies in the expanded network do not have heterogeneity, the resources in the network will not be more diversified due to the expansion of the scale, which also increases the burden on the companies to some extent. Therefore, petroleum companies should pay attention to the management cost brought by the expansion of network scale, do a good job in communication and coordination, ensure the high-quality operation of collaborative innovation activities, and then improve innovation performance.