practical quantitative results that are helpful to the improvement of university
research efficiency.
Specifically, the Tobit model in this section is as below:
DEAscoreit¼a 0 þb 1 Capitalitþb 2 Easternitþb 3 Centralitþb 4 JUitþb 5 SHZUit
þb 6 PerGDPitþ
X
bnXnitþaiþlitwhereDEAscoreitis the DEA efficiency score of ith university in t year;a 0 is
constant term;b 1 ;b 2 ;b 3 ;...;bn;are slope coefficients; Xnitare explanatory vari-
ables like SciTech, HighTech, and so on;litis error term.
To deal with panel data, two common methods are usually estimated, one is
fixed effect model by including a universityfixed effect invariant with time; the
other is random effect model. According to the results of Hausman test,fixed effect
model is much better for our data.
Table 6.5 Estimation results of influencing factors of NEAM research efficiency
(1) (2) (3) (4)
Coef. P > z Coef. P > z Coef. P > z Coef. P > z
Captial 0.0315 0.54 −0.0115 0.818 0.0324 0.528 −0.0112 0.823
East 0.0572 0.198 ––0.0591 0.178 ––
Central −0.0662 0.18 ––−0.0661 0.181 ––
JU ––0.1921 0.016 ––0.1710** 0.021
HSZU ––0.1138 0.022 ––0.1008** 0.031
PerGDP −0.0026** 0.02 −0.0029*** 0.01 −0.0026** 0.02 −0.0028** 0.012
SciTech 0.0017* 0.100 0.0017* 0.100 0.0017 0.143 0.0018 0.128
HTech 9.0E−06 0.787 8.0E−06 0.812 9.0E−06 0.788 9.0E−06 0.805
211Us 0.0010 0.727 −0.0037 0.329 ––––
985Us ––––0.0024 0.83 −0.0097 0.484
IntEX −0.0007 0.836 −0.0007 0.83 −0.0007 0.84 −0.0008 0.819
Type −0.0119 0.765 -0.0076 0.847 −0.0140 0.721 −0.0016 0.967
MOE −0.0668* 0.066 −0.0652* 0.072 −0.0664* 0.07 −0.0650* 0.08
ARWU −0.0175 0.55 −0.0187 0.523 −0.0178 0.545 −0.0184 0.531
FePerc −0.0049* 0.06 −0.0048* 0.063 −0.0049* 0.06 −0.0048* 0.066
TitPerc −0.0003 0.87 −0.0011 0.619 −0.0003 0.878 −0.0010 0.65
DocPerc 0.0021 0.117 0.0023* 0.084 0.0021 0.11 0.0021 0.102
N 500 500 500 500
Log
likelihood260.188 261.043 260.149 260.818
Wald chi^2
(P > chi^2 )22.29 24.37 22.23 23.75
0.073 0.041 0.074 0.049
Notesp< 0.1; p< 0.05; p< 0.01
6.2 Econometric Analysis of Factors... 277