324700_Print.indd

(WallPaper) #1

In the evaluation on NEAM disciplines, we use the number of licensed patents to


measure this output dimension. Patent is the document granting an inventor the sole


right to an invention, demonstrating the capacity of a university to create new things


and turn them into real products. Patent is an important form of applied research


and experimental development in NEAM disciplines, and also an important aspect


of research strength and creativity of research-oriented universities. But to HSS


disciplines, such IPs are very rare, so we will not take into account this indicator in


the HSS indicator system.


4.2 Overall Status and Change of 211Us’Research


Production Efficiency


4.2.1 NEAM Disciplines


In this section, we apply DEA with constant returns to scale and variable returns to


scale technology to analyze the research efficiency of 211Us. Considering the time


lag effect of research input–output,
6
we’re trying to use the moving average with


one-year lag (MA)
7
method to analyze the efficiency status of NEAM research


production. With MA method, we can lower the estimation bias caused by irregular


data variances, and at the same time, solve the problem of time lag in research


production.


4.2.1.1 2006–2010 Results with Original Data


Applying CRS and VRS DEA models to our original data, we calculate the research


production efficiency of the full NEAM sample for thefive-year period from 2006–



  1. See Tables4.41,4.42,4.43,4.44and4.45in Appendix for detailed results.


Table4.4presents the numbers of DEA-efficient DMUs and their operation


stages in NEAM disciplines of 211Us in 2006–2010. In thefive-year period, 2006


(^6) Pakes and Griliches ( 1984 ) and Hausman et al. ( 1984 ) proposed that there is time lag between
research input and output. Research output depends on the research inputs in the current period,
and possibly the research inputs in the past periods.
(^7) A detailed explanation seems to be needed here. For example, to every input indicator, we
transform input data of 2006, 2007, 2008, 2009, 2010 into input data of 2006–2007, 2007–2008,
2008 – 2009, 2009–2010, by calculating their two-year averages. To every output indicator, we do
the similar calculations, and get output data of 2006–2007, 2007–2008, 2008–2009, 2009–2010.
One-year lag is to consider the time lag between input and output. We assume that the inputs in
one year will create products in the next year. Namely, the output in 2007 are created by inputs in
2006, and etc. Combing two-year average and one-year lag together, our MA method will asso-
ciate inputs data of 2006–2007, 2007–2008, 2008–2009 to outputs data 2007–2008, 2008–2009,
2009 – 2010 respectively in the DEA model. Therefore, there are only three periods of data (2007–
2009) in our analysis.
122 4 Evaluation on Research Efficiency of 211Us: The DEA Approach

Free download pdf