Abstract:     After 30 years of development, especially since Chinese government proposed the "Mass Entrepreneurship and Innovation" initiative in 2014, the number of China′s Technology Business Incubators (TBIs) has gained a dramatic growth. In 2017, the total number reached 4,069, making China the largest country in the world with the number of TBIs. Also, China′s TBIs have helped 40,000 technology companies to obtain 194 billion yuan of venture capital. The number of listed incubatees after graduation reached 2,777 which accounts for 1/7 of the listed companies on the GEM, and 1/10 of the listed companies in the New Third Board in 2007. However, the incubator research theory on China′s and global TBIs is still in its youth stage, and academic research lags far behind the practice due to the shortage of data. 
    The performance of TBIs depends on three main factors: the city, the incubator and the incubator. Existing research has the most adequate understanding of the role of incubators in incubation performance, and the importance of physical assets, service capabilities, knowledge reserves, financial support, and social networks has been confirmed. In recent years, studies have begun to focus on the impact of cities on incubation performance, especially the impact of factors such as urbanization, urban wealth, human capital, and technological development on technology business incubators. However, as far as we know, the literature on the effect of incubatees′ quality on hatching performance is quite limited. The comparative analysis of the three factors is even rarer.
     Therefore, this paper hopes to use the latest large-scale statistical data of China′s technology incubator to enrich the incubator research theory and bridge the gap between theory and practice by answering three questions: First, whether cities, incubators and incubators have a positive impact on incubation performance? Second, the impact of the three main factors on the incubation performance is weak and weak? Third, which specific variables have the greatest impact on incubation performance? Through the research in this paper, in addition to supplementing existing theoretical research, it can also provide some reference for the work of policy makers, incubator managers and start-up entrepreneurs.
     Based on the resource-based theory, this article compares the effects of resources of cities, TBIs and incubatees on incubation performance with the application of the data of 857 TBIs from 33 biggest cities in China from 2015 to 2017. We use Fisher discriminant analysis approach to conduct our research. First, we adopt research variables, both dependent and independent ones, on the basis of the literature research. Second, we test the effective predictive ability of the independent variables and construct the general discriminant equation. Third, we compare the discriminant equation models of the three types of high, medium and low results. Eventually we can obtain the final intensity comparison of the impact of the resources of the city, TBI and incubatees on incubation performance. 
    The study found that the resources of cities, TBIs and incubated enterprises all have positive impacts on the performance of incubation. Among them, the impact of incubated enterprises′ resources is the strongest, the TBIs′ is second, and the impact of cities′ resources on incubation performance is the weakest. The impact of incubated enterprises′ resources on incubation performance is the strongest, consistent with the conclusions of the relevant research literature on incubatee selection, which confirms the practical significance and value of the incubator setting in the selection indexes. Meanwhile, this also gives us another inspiration, that is, simply looking at the performance of the TBIs may not accurately reflect the level of their operations. We should consider to introduce some new evaluation methods which can eliminate the impact of incubatees as much as possible. 
     Second, TBIs′ resources also have an important influence on incubation performance, and the impact is only slightly lower than the incubatees′. This is consistent with the conclusions of previous large-scale incubator studies which proves the value of TBIs and provides a theoretical basis for incubator managers to improve the TBIs′ qualities of physical and consulting services. In addition, the combination of the above two conclusions confirms that the incubator and the incubatee can only achieve better results if they complement and work with each other.
     Interestingly, the impact of urban innovation resources on incubation performance is the weakest which indicates the impact of urban macro factors on incubators and incubatees is not so strong as people intuitively imagined. Taking into account the data set used in this paper, maybe it also shows that the gap of the innovation environment among China′s largest cities has been converging. Furthermore, from the perspective of specific variables, human resources are the most important. Human resources variables are the most powerful indicators of prediction to incubation performances. Incubator service personnel including the number and quality of entrepreneurial instructors, and the reserves of technicians in the incubatees are very effective indicators of incubation performance.
     This paper has some policy implications. First TBIs must establish screening and selecting indicators on incubatees. Establishing indicators for incubating enterprises is conducive to optimizing the allocation of resources, avoiding the phenomenon of "bad money drives out good", and promoting the positive interaction and development between enterprises within the incubator. Second, TBIs managers must upgrade the incubators′ resources and capabilities. While paying attention to physical assets and hardware configuration, they should pay more attention to the configuration and upgrade of "software" and strengthen the service capabilities of management and service personnel training, entrepreneurial tutors, and external networks. The third is to accumulate the human capital of TBIs and incubatees. It is necessary to establish effective talent introduction, incentive, training and service mechanisms to continuously improve human capital. TBIs should strengthen the training of employees and continuously improve the professional ability and professional level of the employees. 

Key words: technology business incubator, incubation performance, comparative research, discriminant analysis