6.2.3.2 Sectoral and Institutional Factors
- International Cooperation and Communication
“InterEx”is the proxy variable to measure the frequencies of international
cooperation and communication for a particular university. The results in Table6.5
shows that it has trivial impact on research efficiency with no statistical significance.
This reflects that the increasingly frequent international cooperation and commu-
nication haven’t functioned well in improving research quality in Chinese
universities.
- University Reputation
“Type”,”MOE”and“ARWU”are three proxy variables to measure the repu-
tation of a university. The results in Table6.5show that“Type”and“ARWU”
variables have very small impacts on research efficiency with no statistical signif-
icance, while“MOE”has negative impact significantly at 0.1 level. It is implying
that MOE-administered universities tend to perform worse than those universities
not under MOE administration at the aspect of research efficiency. The underlying
explanation for this observation might be that MOE-administered universities are
heavily funded by the central government, resulting into their being less competi-
tive and less efficient. In contrast, those 211Us not administered by MOE have to
make efforts to improve their efficiency, so as to get better development and more
resources.
6.2.3.3 Internal Factors
- Demographic Structure of University Staff
“FePerc”is the key demographic variable, and it has negative impact on research
efficiency at 0.1 level, which is consistent with existing empirical evidences. The
underlying reasons may be that, female staff have to take care of children, senior
people, and do more housework, so they tend to under-perform male staff.
However, in order to carry out further researches we could consider the differences
on the structures of disciplines and staff.
- Human Capital Structure of University Staff
We put“DocPerc”and“TitPerc”two proxy variables into the models to measure
human capital of a university.“DocPerc”has positive impact on research efficiency,
however, its statistical significance changes in different models.“TitPerc”has
negtive impact on research efficiency, yet with no statistical significance. The
partial coefficient of“DocPerc” is larger than that of“TitPerc”, implying that
“DocPerc”is a better indicator in terms of representativeness and influence.
6.2 Econometric Analysis of Factors... 279