Inno-Qual efficiency of higher education: Empirical testing using data
envelopment analysis
Jui-Kuei Chen
a,1, I.-Shuo Chen
b,* aGraduate Institute of Futures Studies, Tamkang University, 4F, No. 20, Lane 22, WenZhou Street, Taipei City 10616, Taiwan b
Institute of Business & Management, National Chiao Tung University, 4F, No. 20, Lane 22, WenZhou Street, Taipei City 10616, Taiwan
a r t i c l e
i n f o
Keywords: Inno-Qual efficiency Higher education
The Inno-Qual performance system Data envelopment analysis
a b s t r a c t
Since the overall quality of Taiwanese university education has been decreasing in recent years, and uni-versities are losing their competitive advantage while facing foreign countries’ education systems and the threat of closing, upgrading innovation performance and improving total quality performance (Inno-Qual performance) so as to enhance overall operational performance have become an urgent issue. Although relative measurement models are increasingly being used for conquering the above-mentioned difficul-ties, such as the Inno-Qual performance system (IQPS), which integrates the features of innovation and TQM, currently, no studies empirically evaluate the efficiency of such improvement, meaning that the costs of using the Inno-Qual performance system are increasing, particularly the human administrative cost for providing intellectual products, which is the nature of higher education. To overcome this problem, in this study, the IQPS is adopted by using data envelopment analysis (DEA) to evaluate the Inno-Qual efficiency of 99 Taiwanese universities divided into five types (research-intensive, teaching-intensive, profession-teaching-intensive, research & teaching-teaching-intensive, and education-in-practice-intensive). On the basis of the empirical results, we found that over half (73%) of the universities are highly inefficient in improving the Inno-Qual performance, and thus we conclude that improving the Inno-Qual efficiency based on our results will be helpful for reducing the majority of cost expenditures.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
In today’s knowledge-based economic system, it is well known that higher education is the foundation of fostering high-tech tal-ent, the key factor in raising national quality, and the main way to upgrade national competitive ability (Fairweather, 2000; Meek, 2000). The importance of higher education is especially empha-sized in Taiwan; however, with the birth rate dropping, the num-ber of universities increasing, and Taiwan joining the WTO, as compared to foreign countries, the overall quality of Taiwanese universities is decreasing, and universities are losing their compet-itive advantages (Chen and Chen, 2009a, 2009b, 2009c, submitted for publication; Taiwan Assessment and Evaluation Association, 2006). In this regard, regaining and increasing their competitive advantages has now become a critical problem for the Taiwanese government and for the universities (Department of Higher Educa-tion, 2004).
To overcome these problems, evaluating and improving the innovation performance and total quality management perfor-mance are needed (Chen & Chen, 2009a, 2009b, 2009c, submitted for publication), and, therefore, a growing body of research has proposed several related measurement models (Chen & Chen, 2008a, 2008b; Chin & Pu, 2006; Lin, Wang, Wang, & Yen, 2006; Tang, 2006); in addition, officially, visiting and standard procedure evaluation are some of the main performance evaluation methods for the Taiwanese Ministry of Education. Such evaluation methods, nevertheless, have numerous drawbacks and biases, such as mea-surement criteria proposed under the assumption of independence among each other, which is not the real-world situation; the func-tion of each model can only measure the performance of TQM or innovation separately; and ignore the features of each type of university (e.g., research-intensive, teaching-intensive, and profes-sion-intensive, proposed byLi (2007), and education-in-practice-intensive, proposed by Chen & Chen (2008a, 2008b)); and the methods largely focus on internal organizational improvement, which makes performance evaluation incomprehensive. Recently, a measurement model focusing on higher education proposed by
Chen and Chen (2009a, 2009b, 2009c, submitted for publication)
overcame the above-mentioned problems. It was called the Inno-Qual performance system (IQPS), and it is believed to be able to 0957-4174/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2010.07.111
* Corresponding author. Tel.: +886 911393602.
E-mail addresses: [email protected] (J.-K. Chen), ch655244@yahoo. com.tw(I.-Shuo Chen).
1 Tel.: +886 912272961.
Contents lists available atScienceDirect
Expert Systems with Applications
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / e s w aprovide accurate results while evaluating and improving the over-all quality of a university, as we argue.
Although the Inno-Qual performance system (IQPS) that can overcome the above-mentioned difficulties, currently, no studies empirically evaluate the Inno-Qual efficiency when universities using it, resulting in the Inno-Qual performance improving along with the costs increasing, particularly the human administrative costs due to one of nature of higher education, providing intellec-tual products. In this paper, we aim to overcome the above prob-lems by adopting the IPQS and using data envelopment analysis (DEA). We believe that this study can not only indicate how to im-prove the Inno-Qual’s efficiency so as to aid universities for improving future performance, but it can also enhance the validity of the IPQS for future use.
The rest of this paper is organized as follows. The overview of Inno-Qual in higher education is discussed in Section2. Data envel-opment analysis is introduced in Section3. Empirical testing and discussions are conducted and presented in Section4. Conclusions are in the last section.
2. The overview of Inno-Qual efficiency in higher education The term innovation has received more attention than ever be-fore for its ability to sustain competitive advantages (Daft, 2004; Krause, 2004). Therefore, its definition is not clear and will con-stantly change. Innovation performance has various criteria and indices along with different points of emphasis and concepts (Chen & Chen, 2009a, 2009b, 2009c, submitted for publication), ranging from an invisible concept or phenomenon to a visible product (Acs, Anselin, & Varga, 2001; Bosworth & Rogers, 2001; Chen & Chen, 2008a, 2008b; Dzinkowski, 2000; Gambardella & Torrisi, 2000; Guthrie & Petty, 2000; Hall & Bagchi-Sen, 2002; O’Sullivan, 2000; Ordaz, Lara, & Cabrera, 2005; Schoenecker & Swanson, 2002; Subramaniam & Youndt, 2005; Toivanen, Stoneman, & Bosworth, 2002; Van Buren, 2000).
In the field of higher education, innovation indices for universi-ties to conduct innovation performance evaluation and improve-ment are numerous (Mei & Lee, 2006). Nevertheless, although the number of related measurement models is increasing, they face several biases, such as measurement criteria, which are proposed under the assumption of independence among each other, which is unsuitable for real-world situations, or ignorance of the features of the different types of universities. In this regard, among current measurement models, the innovation support system (ISS) and the pro-performance appraisal system (PPAS) are the latest systems developed not merely to enhance and measure innovation perfor-mance (Chen & Chen, 2008a, 2008b) for higher education but also to overcome the above-mentioned biases, as shown inTable 1and
Fig. 1.
The criteria include the extent of international academic inter-action, the amount of financial support from the National Science Council (NSC), the number of journal articles accepted and pub-lished, the number of conferences and chaired professors, and the extent of a results-oriented organizational culture. Together,
these criteria compose the ISS. Moreover, an innovation accelera-tion force involving transformaaccelera-tional leadership is the latest ISS development to promote effective innovation performance across the three main types of universities (e.g., research-intensive, teach-ing-intensive, and profession-intensive) (Chen & Chen, 2008a, 2008b). Similarly, the criteria include employee turnover, the num-ber of promotions, the numnum-ber of articles published in interna-tional journals, the number of patents, winning student thesis, plans given by the NSC, and the level of satisfaction in industries. Together, these criteria compose the PPAS. Additionally, a support appraisal system to address financial support and budget planning is the newest PPAS development that ensures successful innovation.
Similar to innovation, the importance of the total quality man-agement concept and related measurement models to assess its performance is increasing. Thus, the TQM criteria vary from one re-searcher to another (Chen& Chen, 2008a, 2008b, 2009a, 2009b, 2009c, submitted for publication; Dinh, Barbara, & Tritos, 2006; Escrig-Tena, 2004; Han, Chen, & Ebrahimpour, 2007; Ismail, 2006; Keng, Nooh, Veeri, Lorraine, & Loke, 2007; Kenneth & Cyn-thia, 2004; Nusrah, Ramayah, & Norizan, 2006; Ozden & Birsen, 2006; Wanger & Schaltegger, 2004).
Without the concept of total quality management for the ISS and the PPAS, Chen and Chen (2009a, 2009b, 2009c, submitted for publication) proposed an advanced innovation performance measurement system called the network hierarchical feedback sys-tem (NHFS) (Fig. 2). They integrated the characteristics and revised the drawbacks of the ISS and the PPAS and combined the concepts of the TQM. InTables 2 and 3, detailed definitions of measurement criteria and indices for the NHFS and the types of universities are given.
Although the NHFS overcomes the majority of drawbacks and biases of the previous models and integrates the characteristics of innovation and total quality management, this system nonethe-less has been criticized for its strong focus on internal organiza-tional improvement and the lack of attention to external features, which does not fit the current overall trend for most Tai-wanese universities. In view of this shortcoming, Chen and Chen improved the NHFS by providing the concept of external organiza-tional improvement after a series of quantitative and qualitative studies to today’s Inno-Qual performance system (IPQS) (Fig. 3). In Tables 2 and 3, detailed definitions of measurement criteria, indices and the types of universities evaluated under the IPQS are provided.
Due to overcoming drawbacks and biases that current measure-ment models face, we claim that the IPQS is the most appropriate model for precisely evaluating innovation, total quality manage-ment performance and efficiency (Inno-Qual efficiency). Currently, no studies that evaluate the Inno-Qual efficiency for universities in Taiwan exist, which makes today’s universities that use the IPQS face high expenditures, particularly on human administrative costs owing to one of nature of higher education, providing intellectual products. In this regard, we aim to overcome the aforementioned dilemmas by adopting the IPQS with data envelopment analysis (DEA). We believe that this study can not only further indicate Table 1
An innovation support system (ISS).
IS system IS dimension IS criteria Optimal IS type
A novel innovation support system (ISS)
Academic research International academic interaction Research-intensive university (RU) Financial support of NSC
Journals accepted and published External academic support Number of conferences
Number of chaired professors Organizational culture Results-oriented
how to improve Inno-Qual’s efficiency for improving the perfor-mance of universities, but it can also enhance the validity of the IPQS for future utilization.
3. Data envelopment analysis
The concept of data envelopment analysis (DEA) comes from the non-parametric frontier approach proposed by Farrel (1957).
Charnes, Cooper, and Rhodes (1981)further advanced Farrel’s ap-proach by extending it from single-input–single-output technical efficiency measurements to multiple-input and multiple-output measurements. They also proposed the CCR model, making the scores model of the DEA transform into the linear programming model, and, they introduced duality theory in order to make calcu-lation with greater ease. However, the CCR model does not con-sider the restrictions of convexity; that is, the CCR model works
under the assumption of a constant return to scale (CRS), which is not fit for real practice. Hence, Banker, Charnes, and Cooper (1984)revised the CCR model and finally proposed the BCC model, adopting the assumption of variable return to scale (VRS).
The DEA aims to understand when a corporate, under a specific output, if an organization inputs too many resources or when, un-der a specific input level, if an organization outputs too little. Therefore, the application models of the DEA can be categorized into the input orientation model and the output orientation model. Under a certain quantity of outputs, using the minimization input level to compare the efficiency of the decision-making unit (DMU) is called input orientation. Similarly, under a certain input level, using the maximum quantity of output to compare the efficiency of decision-making unit (DMU) is called output orientation. Since higher education mainly provides invisible products (e.g., knowl-edge and skills) and the variable costs are relatively low, in this study, the output orientation model is adopted to evaluate
Subsystem Appraisal Dimension Appraisal Criteria
Core appraisal
system
Organizational development
The employee turnover The number of promotions
Academy performance
Number of articles published in international journals
Number of patents Number of winning student thesis
External behavior
Number of plans given by the NSC The level of satisfaction of industries
Financial support and budget planning
Support appraisal system
Fig. 1. A pro-performance appraisal system (PPAS).
Strategic method Strategic Goal
Transformational Leadership Financial Support and
Budge Planning
Total Quality Management Dimension
Innovation Dimension RU I3 I4 I1 I5 I9 T11 T9 T10 T12 T13 T19 T28 T32 High Operation Performance TU PU EIPU Type Translation Characteristics Integration PPAS ISS Organizational Goal Optimal Type Conduct Conduct Improve Improve
Inno-Qual efficiency. The CCR model is used to explore slack vari-ables of input and output under a constant return to scale (CRS).
The BCC model is used to evaluate pure technical efficiency (PTE) and scale efficiency (SE) of a single period.
3.1. The CCR model
The DEA evaluates samples that are going to be evaluated for their efficiency as decision-making units (DMU). Assuming that there are n DMUs, each piece of DMU (DMUi) utilizes m kinds of in-put xij(j = 1, 2, 3, . . . , m), xij=0; and produces s kinds of output yir (r = 1, 2, 3, . . . , s), yir=0. The CCR model transforms these multiple inputs and outputs into a single input and output by using virtual Table 2
Definition of measurement criteria and indices of the NHFS and IQPS.
Number Definition of measurement criteria and indices Weights Number Definition of measurement criteria and indices Weights
Internal External
T9 Process redefinition of R&D and innovation (C) 0.136 T4 TQM Culture Construction (C) 0.087
T10 Input of R&D and innovation (C) 0.111 T10 Input of R&D and innovation (C) 0.109
T11 Evaluation of R&D and innovation results (C) 0.173 T11 Evaluation of R&D and innovation results (C) 0.115
T12 Market operation strategy development (C) 0.111 T13 Business relations management (C) 0.112
T13 Business relations management (C) 0.198 T14 Customer relationship management (C) 0.278
T19 Knowledge management (C) 0.086 T24 Supportive activity planning (C) 0.299
T28 Financial performance (C) 0.086 I6 The responsibility of the instructor (I) 0.187
T32 Unique competitive ability gaining performance (C) 0.099 I8 Promotion and job acquisition for previous students (I) 0.206
I1 Research patents (I) 0.197 I9 Appropriate use of multimedia (I) 0.228
I3 Financial support from national science council (I) 0.168 I10 The number of cooperating international universities (I) 0.170 I4 Journals accepted and published (I) 0.270 I15 Teaching that combines practice, attending courses and
learning theory (I)
0.209
I5 Government tender planning (I) 0.195
I9 Number of chaired professors (I) 0.170
(C): measurement criteria and (I): measurement indices.
Table 3
Definition of the types of universities for the NHFS and IQPS.
Number Definition of types of universities
RU Research-intensive university
TU Teaching-intensive university
PU Professional-intensive university
EIPU Education-in-practice-intensive university
Transformational Leadership Financial Support and
Budge Planning
High Operation Performance
Type Translation Characteristics Integration Innovation Dimension I3 I4 I1 I5 I9 I6 I8 I9 I10 I15
Total Quality Management Dimension T11
T9 T10 T12 T13 T19 T28 T32 T4 T10 T11 T13 T14 T24
RU Conduct
Improve External-organization oriented improvement Internal-organization oriented improvement
Strategic Method Strategic Goal Organization Goal
Financial Support and Budge Planning Transformational
Leadership
TU PU EIPU
Table 4
The input and output data of sample universities. Source:Higher Education Evaluation and Accreditation Council of Taiwan (2008, 2009), Department of Statistics (2008) and National Science Council (2009).
University type Sample university Output Input No. of graduates (2008) Journal articles accepted and published (ESI) (2009) Quantity of financial support from theNSC (2009) Research Patents (N > 20) (2004–07) No. of cooperating foreign countries (2008) No. of domestic students (2008) No. of International Members (2008) No. of domestic full-time faculty (2008) R DMUR1 7719 28384 722 205 66 33416 2516 1937 DMUR2 3565 10876 223 258 50 14184 608 698 DMUR3 2785 10963 333 158 41 11775 440 604 DMUR4 5608 16237 438 248 58 21972 1551 1207 DMUR5 3062 6741 172 143 35 11954 593 579 DMUR6 2617 6760 158 152 29 9348 739 475 DMUR7 892 7170 89 0 22 4296 176 383 T DMUT1 3278 2621 142 0 40 15514 3802 876 DMUT2 1603 0 19 0 6 7492 175 301 DMUT3 2068 0 49 0 3 8135 81 390 DMUT4 833 0 9 0 1 3704 29 178 DMUT5 1007 0 22 0 3 4632 36 179 DMUT6 896 0 12 0 4 4534 59 200 DMUT7 1027 0 11 0 2 4347 48 210 DMUT8 1429 0 14 0 1 5964 145 214 P DMUP1 1968 4224 91 93 27 8737 303 382 DMUP2 2063 1318 48 24 11 8487 264 335 DMUP3 2416 0 12 34 30 10609 117 365 DMUP4 2261 1478 49 46 21 8973 145 426 DMUP5 1437 0 23 23 7 6367 52 258 DMUP6 3522 0 11 47 6 9880 136 356 DMUP7 1338 0 9 0 4 6082 23 228 DMUP8 1173 0 5 36 2 7774 14 277 DMUP9 2839 704 18 0 9 13937 143 393 DMUP10 1562 4061 69 0 9 7049 320 510 DMUP11 3496 0 18 32 17 18184 346 587 DMUP12 2334 0 3 32 15 11730 64 452 DMUP13 3507 0 16 55 4 15687 14 509 DMUP14 2343 0 5 0 7 10955 89 284 DMUP15 1067 3068 51 0 5 5411 249 412 DMUP16 1614 2189 31 0 9 7584 229 465 DMUP17 1708 0 1 0 4 8949 9 261 DMUP18 2555 0 12 51 4 13360 55 456 DMUP19 1566 3091 34 0 9 7667 208 535 DMUP20 1896 0 9 0 8 10506 28 365 DMUP21 1634 0 6 0 7 10423 14 429 DMUP22 1919 0 2 0 3 10441 18 296 DMUP23 1216 0 4 0 3 6204 14 299 DMUP24 1470 0 0 54 3 6918 19 325 DMUP25 1033 0 1 0 6 6010 23 357 DMUP26 1681 0 3 0 4 7957 7 284 DMUP27 1587 0 2 0 4 7540 17 343 DMUP28 1007 0 3 577 3 5863 6 289 DMUP29 958 0 9 0 2 6028 6 270 DMUP30 1197 0 1 0 5 6732 26 286 DMUP31 855 0 2 0 3 5900 11 297 DMUP32 482 0 1 77 2 4109 6 244 DMUP33 1951 0 13 126 2 9525 10 331 DMUP34 630 0 2 0 1 3036 1 113 DMUP35 852 0 9 0 2 3698 9 442 DMUP36 1693 0 12 0 1 8692 6 336 DMUP37 1120 0 3 0 2 6841 5 255 DMUP38 1829 0 2 0 2 9143 7 364 DMUP39 1639 0 7 0 3 7186 6 330 DMUP40 799 0 6 0 1 6336 6 253 DMUP41 1143 0 6 0 0 5607 0 330 DMUP42 380 0 2 0 0 2488 0 172
R&T DMUR&T1 3617 0 156 0 60 15588 1322 677
DMUR&T2 3900 5861 148 136 37 17204 671 752 DMUR&T3 2073 2654 43 0 21 8496 196 379 DMUR&T4 2992 3358 126 27 12 12044 369 498 DMUR&T5 2335 0 30 0 13 9928 413 326 DMUR&T6 2592 0 31 0 11 12239 174 501 DMUR&T7 1112 0 35 0 14 5234 162 196 DMUR&T8 2328 0 65 0 8 10502 206 515 DMUR&T9 1178 791 45 0 9 5260 348 240 DMUR&T10 1232 0 11 0 3 4325 40 180
multipliers urand
vj
and makes such a virtual output and input ra-tio DMU’s efficiency value (hi). The DEA adopts the maximum value for each efficiency of DMU and from the feasible solution sets of each DMU’s virtual multipliers to explore the best weighted valuefor the DMU, making hia maximum. After considering its original unlimited scope and solutions, the CCR model (function one) there-by revises and provides the following:
Max u;y hi¼ Xs r¼1 uryir s:t: X j¼1
v
jxij¼ 1 Xs r¼1 uryir Xm j¼1v
jxij50 ð1Þ where ur,vj
=0; i = 1, 2, 3, . . . , n; j = 1, 2, 3, . . . , m; r = 1, 2, 3, . . . , s. In 1984, Boyd and Fare found that when either urorvj
is zero in the CCR model, solutions will become degenerate, making the effi-Table 4 (continued) University type Sample university Output Input No. of graduates (2008) Journal articles accepted and published (ESI) (2009) Quantity of financial support from theNSC (2009) Research Patents (N > 20) (2004–07) No. of cooperating foreign countries (2008) No. of domestic students (2008) No. of International Members (2008) No. of domestic full-time faculty (2008) DMUR&T11 1346 193 16 0 4 7404 45 292 DMUR&T12 1176 0 18 0 1 5127 20 234 DMUR&T13 1309 0 26 0 12 6128 79 228 DMUR&T14 3762 0 47 0 14 17604 591 524 DMUR&T15 5891 1267 41 0 26 27534 1288 680 DMUR&T16 3324 0 26 0 13 15934 344 443 DMUR&T17 3617 2426 66 117 13 16476 218 486 DMUR&T18 6502 2648 64 0 42 27729 1818 791 DMUR&T19 5651 0 26 0 18 26924 1334 742 DMUR&T20 4896 2171 47 52 31 20783 596 654 DMUR&T21 2690 0 22 0 19 12357 569 371 DMUR&T22 1614 5238 52 70 7 7380 108 594 DMUR&T23 2185 1666 62 36 24 9398 147 320 DMUR&T24 2162 0 10 0 23 8961 103 303 DMUR&T25 2585 0 22 0 7 10583 41 352 DMUR&T26 889 0 6 0 7 4383 29 140 DMUR&T27 3279 1424 28 0 13 14695 107 549 DMUR&T28 2364 0 11 0 11 10895 315 316 DMUR&T29 4098 0 16 0 64 17969 1282 660 DMUR&T30 3160 0 12 0 13 14733 271 387 DMUR&T31 1546 0 4 0 15 6352 83 207 DMUR&T32 2430 0 9 0 5 10732 62 370 DMUR&T33 1152 992 8 27 2 5115 19 217 DMUR&T34 679 700 14 0 8 3112 269 300 DMUR&T35 1504 0 3 0 8 7004 27 176 DMUR&T36 2153 0 8 0 10 11324 78 325 DMUR&T37 1749 0 11 0 18 8870 280 294 DMUR&T38 444 0 5 0 11 2578 113 133 DMUR&T39 936 0 9 0 4 6668 29 233 DMUR&T40 2220 0 13 0 7 10080 50 355 DMUR&T41 1160 0 5 0 3 4922 13 178 EP DMUEP1 806 0 0 0 3 3071 5 120R: research-intensive; T: teaching-intensive; P: professional-intensive; R&T: research & teaching-intensive; and EP: education-in-practice-intensive.
Table 5
Efficiency value and return to scale for research-intensive universities.
Sample university CRSTE VRSTE SE RS
DMUR1 1.000 1.000 1.000 CRS DMUR2 1.000 1.000 1.000 CRS DMUR3 1.000 1.000 1.000 CRS DMUR4 0.865 0.966 0.895 DRS DMUR5 0.678 0.679 0.998 DRS DMUR6 1.000 1.000 1.000 CRS DMUR7 1.000 1.000 1.000 CRS Table 6
The input and output slack for research-intensive universities.
University type Sample university Output Input No. of graduates (2008) Journals accepted and published (ESI) (2009) Quantity of financial support from theNSC (2009) Research patents (N > 20) (2004–07) No. of cooperating foreign countries (2008) No. of domestic students (2008) No. of international members (2008) No. of domestic full-time faculty (2008) R DMUR1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR2 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR3 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR4 1.520 0.000 259.752 0.000 0.000 12.401 0.000 350.444 DMUR5 0.000 575.076 0.000 36.131 0.000 0.000 0.000 208.168 DMUR6 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR7 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
ciency value incorrect. Thus, they improved this problem by intro-ducing the Archimedean quantity (
e
), making ur, andvj
into =e
. To make the above model easily used, the dual problem of linear pro-gramming should be used to as to minimize the number of con-straints. Thus, the relative efficiency value of the DMU can be acquired. Additionally, under constant return to scale (CRS), there is still room for the DMU to improve its input and output, and thus slack variables of input and output Sij;S þ ir
can be introduced. The above model therefore can be revised as follows (function two). In the revised model, kiare the weights of each DMU, and hiis the relative efficiency of DMUi. When hi=1, this means that the DMU contains operational efficiency so that S
ij¼S þ
ir¼0. Otherwise, the DMU does not contain operational efficiency, and therefore slack variables of input and output, namely, S
ij and S þ ir, can be calculated Min hi¼ hi
e
Xm j¼1 S ijþ Xs r¼1 Sþ ir " # s:t: X n i¼1 kixij hixijþ Sij ¼ 0 Xn i¼1 kiyir Sþir ¼ yir ð2Þ where ki=0; i = 1, 2, 3, . . . , n; j = 1, 2, 3, . . . , m; r = 1, 2, 3, . . . , s. 3.2. The BCC modelSince the CCR model contains the assumption of a constant re-turn to scale (CRS), which does not fit well with the real-world sit-uation,Banker et al. (1984)proposed the assumption of variable return to scale (VRS), using four axioms of the produce possible set (PPS) (e.g., convexity, inefficiency, ray unboundness, and mini-mum extrapolation) and the distance function ofShephard (1970)
to derive pure technical efficiency (PTE) and scale efficiency (SE). After considering variable return to scale (VRS), the following lin-ear programming model (function(3)) can be adopted:
Min h;k hi s:t: X n i¼1 kjxij5hixij Xn i¼1 kjyir=yir Xn i¼1 kj¼ 1 ð3Þ where ki=0; i = 1, 2, 3, . . . , n; j = 1, 2, 3, . . . , m; r = 1, 2, 3, . . . , s. Table 7
Efficiency value and return to scale for teaching-intensive universities.
Sample university CRSTE VRSTE SE RS
DMUT1 0.396 0.402 0.987 IRS DMUT2 0.794 0.877 0.905 IRS DMUT3 1.000 1.000 1.000 CRS DMUT4 0.978 0.978 1.000 CRS DMUT5 0.565 1.000 0.565 IRS DMUT6 0.891 0.931 0.957 DRS DMUT7 0.386 1.000 0.386 IRS DMUT8 1.000 1.000 1.000 CRS Table 9
Efficiency value and return to scale for profession-intensive universities.
Sample university CRSTE VRSTE SE RS
DMUP1 1.000 1.000 1.000 CRS DMUP2 0.389 0.415 0.937 IRS DMUP3 0.385 0.416 0.924 DRS DMUP4 0.362 0.375 0.966 IRS DMUP5 0.689 0.825 0.835 IRS DMUP6 0.546 0.549 0.994 DRS DMUP7 0.626 0.964 0.649 IRS DMUP8 0.299 0.299 0.998 DRS DMUP9 0.602 0.800 0.753 DRS DMUP10 1.000 1.000 1.000 CRS DMUP11 0.338 0.453 0.746 DRS DMUP12 0.280 0.329 0.850 DRS DMUP13 0.405 0.474 0.855 DRS DMUP14 0.322 0.350 0.920 DRS DMUP15 1.000 1.000 1.000 CRS DMUP16 0.844 0.855 0.953 IRS DMUP17 0.828 0.850 0.974 IRS DMUP18 0.518 0.561 0.924 DRS DMUP19 0.824 0.842 0.979 IRS DMUP20 0.861 0.917 0.939 DRS DMUP21 0.627 0.653 0.960 DRS DMUP22 0.908 0.958 0.947 DRS DMUP23 0.370 0.463 0.801 IRS DMUP24 0.734 0.862 0.852 IRS DMUP25 0.078 1.000 0.078 IRS DMUP26 0.907 0.967 0.939 IRS DMUP27 0.783 0.861 0.909 IRS DMUP28 1.000 1.000 1.000 CRS DMUP29 1.000 1.000 1.000 CRS DMUP30 0.809 1.000 0.809 DRS DMUP31 0.678 0.753 0.901 DRS DMUP32 1.000 1.000 1.000 CRS DMUP33 1.000 1.000 1.000 CRS DMUP34 1.000 1.000 1.000 CRS DMUP35 1.000 1.000 1.000 CRS DMUP36 1.000 1.000 1.000 CRS DMUP37 0.198 0.219 0.904 IRS DMUP38 1.000 1.000 1.000 CRS DMUP39 0.954 1.000 0.954 IRS DMUP40 0.593 0.755 0.785 DRS DMUP41 1.000 1.000 1.000 CRS DMUP42 0.833 1.000 0.833 IRS Table 8
The input and output slack for teaching-intensive universities.
University type Sample university Output Input No. of graduates (2008) Journals accepted and published (ESI) (2009) Quantity of financial support from the NSC (2009) Research patents (N > 20) (2004–07) No. of cooperating foreign countries (2008) No. of domestic students (2008) No. of international members (2008) No. of domestic full-time faculty (2008) T DMUT1 0.000 19.689 6580.519 0.000 129.137 0.000 0.000 0.000 DMUT2 0.000 0.000 618.242 2.060 73.937 0.000 0.000 0.000 DMUT3 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUT4 0.000 0.027 0.000 0.002 0.981 0.000 6.077 19.982 DMUT5 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUT6 0.000 951.442 5.043 0.000 0.000 0.000 0.000 0.323 DMUT7 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUT8 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
When the efficiency value of the constant return to scale (CRS) calculated from Eq.(2)divided by Eq.(3), it provides the efficiency value of the variable return to scale (VRS), also called scale effi-ciency (SE). If scale effieffi-ciency (SE) equals one, it means that the DMU is achieving constant return to scale (CRS). On the contrary, if the scale efficiency (SE) is less than one, it means that the DMU is scale inefficient. However, in order to see whether the scale inefficiency is in increasing or decreasing scale, non-increasing re-turn to scale conditions must be introduced so as to compare the scale efficiency under the variable return to scale (VRS). When sidering the non-increasing return to scale conditions, the con-straint Pni¼1kj¼ 1 in Eq. (3) needs to be revised to Pni¼1kj51. The model can then be revised as follows (function(4))
Min h;k hi s:t: X n i¼1 kjxij5hixij Xn i¼1 kjyir=yir Xn i¼1 kj51 ð4Þ where ki=0; i = 1, 2, 3, . . . , n; j = 1, 2, 3, . . . , m; r = 1, 2, 3, . . . , s. In the comparison with efficiency values calculated from Eqs.
(4) and (3), if two efficiency values equal each other, it means that the DMU is an increasing return to scale (IRS); otherwise, the DMU is a decreasing return to scale (DRS).
4. Empirical testing and discussion
Since there are no studies on Inno-Qual efficiency for industries and because evaluating and improving innovation and total quality management performance are becoming critical issues for higher education in Taiwan, the aim of this study is to overcome the above-mentioned problems by adopting the IPQS using the DEA.
To provide comprehensive and precise results, sample universi-ties were chosen originally based on four types of universiuniversi-ties (i.e., research-intensive, teaching-intensive, professional-intensive, and education-in-practice-intensive). However, this categorization is overbroad for precise measurements because some universities to-day focus equally on teaching and research. Thus, we decided to create a new dimension, the research & teaching-intensive univer-sity. After discarding the universities that did not belong to the
Table 10
The input and output slack for profession-intensive universities.
University type Sample university Output Input No. of graduates (2008) Journals accepted and published (ESI) (2009) Quantity of financial support from theNSC (2009) Research patents (N > 20) (2004–07) No. of cooperating foreign countries (2008) No. of domestic students (2008) No. of international members (2008) No. of domestic full-time faculty (2008) P DMUP1 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP2 0.000 0.000 2144.407 0.000 0.000 0.000 0.000 19.545 DMUP3 0.329 0.000 3977.663 66.510 0.000 6.185 0.000 0.000 DMUP4 0.000 0.000 585.487 0.000 0.000 0.000 0.000 169.157 DMUP5 0.000 0.000 1000.667 0.000 2.845 0.000 0.000 0.000 DMUP6 0.000 0.000 672.825 5.319 34.530 0.000 0.000 323.621 DMUP7 0.000 0.000 0.000 1.598 0.000 0.000 0.000 0.000 DMUP8 0.000 0.000 97.057 0.000 0.000 0.000 0.000 57.638 DMUP9 1.784 0.000 299.255 12.059 115.249 0.000 0.000 358.096 DMUP10 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP11 16.846 0.000 2419.464 53.925 0.000 7.712 0.000 0.000 DMUP12 2.747 0.000 2194.295 44.280 10.760 0.000 0.000 97.972 DMUP13 11.740 0.000 716.986 0.000 0.000 0.000 0.000 212.421 DMUP14 0.000 0.000 843.047 14.224 118.854 0.000 0.000 389.709 DMUP15 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP16 0.000 0.000 0.000 6.280 55.799 0.000 0.000 0.000 DMUP17 0.000 0.000 376.421 19.001 94.233 0.000 0.000 304.849 DMUP18 8.505 0.000 656.959 11.141 0.000 0.000 0.000 0.000 DMUP19 0.000 0.000 0.000 8.268 66.516 0.000 0.000 108.441 DMUP20 2.596 0.000 1215.995 31.800 107.351 0.000 0.000 0.000 DMUP21 0.763 0.000 1269.050 40.209 97.449 0.000 0.000 0.000 DMUP22 2.234 0.000 334.376 21.110 104.257 0.000 0.000 0.000 DMUP23 0.000 0.000 0.011 0.597 0.002 0.000 0.000 0.000 DMUP24 0.000 0.000 274.568 18.204 0.000 0.000 0.000 62.337 DMUP25 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP26 0.000 0.000 418.223 18.094 62.825 0.000 0.000 202.320 DMUP27 0.000 0.000 311.515 15.989 71.700 0.000 0.000 0.000 DMUP28 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP29 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP30 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP31 0.000 0.048 3.234 0.020 0.000 0.000 295.390 4.787 DMUP32 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP33 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP34 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP35 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP36 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP37 0.000 0.000 114.968 2.021 29.682 0.000 0.000 0.000 DMUP38 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP39 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP40 0.000 0.027 0.000 0.027 0.252 0.388 0.000 0.000 DMUP41 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUP42 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
university but in higher education, such as colleges and institutes, all five types of universities (99 universities in total) were studied. As for output data, according to the IPQS, the Inno-Qual indices include research patents, financial support from the National Sci-ence Council (NSC), journal articles accepted and published, gov-ernment tender planning, the number of chaired professors, promotion and job acquisition for all previous students, appropri-ate use of multimedia, the number of cooperating international universities, and teaching that combines practice, attending courses and learning theory (Table 2). To precisely measure the Inno-Qual efficiency, after discussing with senior experts (11 from research-intensive universes, four from teaching-intensive universities, six from professional-intensive universities, 15 from research and teaching universities, and three from education-in-practice-intensive universities), five critical Inno-Qual indices were created. They were: Journal Articles Accepted and Published, Research Patents, Financial Support from the National Science Council, the Number of Cooperating International Universities, and Promotion and Job Acquisition for all Previous Students. These were extracted from external and internal organizational-oriented improvement dimensions (Table 4).
To fit the usage of the DEA and increase Inno-Qual efficiency visibility, we revised the indices while maintaining their character-istics. First, the journal articles accepted and published are re-stricted to those cited by essential science indicators (ESI) that have accumulated until 2009 (Higher Education Evaluation & Accreditation Council of Taiwan, 2009). Second, research patents include domestic and international ones. In this study, the total number of patents includes both types. Based on the latest calcula-tion by theHigher Education Evaluation and Accreditation Council of Taiwan (2008), universities with more than 20 research patents for 2004–2007 were used in this study. Third, financial support from the National Science Council is calculated by the number of cases which accepted to be supported and is limited to 2009. Fourth, we calculated the number of cooperating international uni-versities in accordance with their countries (Department of Statis-tics, 2008) and based on more than half of the senior experts’ opinions that replacing the number of cooperating international universities by their countries will better identify the degree of diversification of a target university. Lastly, obtaining employment, advancing to graduate or professional school, and joining the mil-itary, under Taiwanese law, are the three main paths graduates can take. Hence, to acquire precise results on the last index, we revised it as the number of graduates and restricted it to 2008, according to the latest statistics available from the Department of Statistics in 2009.
As for the input data, since most of the output data are mainly from domestic students, foreign members, including faculty, stu-dents and domestic full-time faculty, and domestic and interna-tional human administrative costs for students and faculty. These do not currently have a specific standard. Thus, such costs are al-ways a major part of the overall costs for a university. In this re-gard, the numbers of domestic students, foreign members, and domestic full-time faculty members calculated in 2009 for 2008 by the Department of Statistics were used as the input data. We be-lieve that finding the efficiency input will help reduce unnecessary costs for Taiwanese universities.
According to the results for research-intensive universities pre-sented in Table 5, the CRSTE, the VRSTE, and the SE of DMUR1, DMUR2, DMUR3, DMUR6, and DMUR7are equal to one; five of them are already at the efficient frontier. Therefore, we suggest that uni-versities should look to provide more profitable external develop-ment opportunities for students or internal development opportunities for high-prestige foreign members and to motivate faculty members to conduct more valuable R&D or submit papers to higher level journals listed by the ESI so as to acquire a higher
level of Innovation and TQM performance for the same amount of input.
In addition, DMUR4and DMUR5are in the decreasing return to scale stage, meaning that both universities should especially de-crease the scale of domestic full-time faculty members or inde-crease the scale of research or student development support so as to in-crease Inno-Qual scale efficiency. More detailed indications regard-ing improvregard-ing Inno-Qual efficiency for the each DMU are presented inTable 6.
Based on the results regarding teaching-intensive universities presented inTable 7, the CRSTE, VRSTE, and SE of DMUT3, DMUT4, and DMUT8are equal to one. This implies that these three univer-sities are at the efficient frontier and thus cannot increase scale efficiency by their original input scale. We suggest that these three universities emphasize the output more on quality than on quan-tity. Since the nature of the teaching-intensive university is to ex-plore its area of research (Chen & Chen, 2008a, 2008b), the value of its field will grow higher if researchers provide more abstruse and detailed insight. Five of the universities are also encouraged to integrate the characteristics of the research-intensive university, such as adding more practical courses and making DMUR1, DMUR2, DMUR3, DMUR6, and DMUR7their benchmarks for future innovation and TQM performance upgrades.
In addition, DMUT1, DMUT2, DMUT5, and DMUT7 are in the increasing return to scale stage; that is, four of them need to in-crease the amount of research or inin-crease the external and internal
Table 11
Efficiency value and return to scale for research & teaching-intensive universities.
Sample university CRSTE VRSTE SE RS
DMUR&T1 0.522 0.645 0.809 DRS
DMUR&T2 0.781 0.985 0.793 DRS
DMUR&T3 0.369 0.399 0.923 IRS
DMUR&T4 1.000 1.000 1.000 CRS
DMUR&T5 0.374 0.375 0.996 IRS
DMUR&T6 0.538 0.589 0.914 DRS
DMUR&T7 0.302 0.881 0.343 IRS
DMUR&T8 0.689 0.693 0.995 DRS
DMUR&T9 0.527 1.000 0.527 IRS
DMUR&T10 0.568 1.000 0.568 IRS
DMUR&T11 0.495 0.548 0.902 IRS
DMUR&T12 1.000 1.000 1.000 CRS
DMUR&T13 0.563 0.712 0.791 IRS
DMUR&T14 0.467 0.672 0.695 DRS DMUR&T15 0.361 0.654 0.552 DRS DMUR&T16 0.233 0.300 0.775 DRS DMUR&T17 0.543 0.670 0.811 DRS DMUR&T18 0.216 0.391 0.552 DRS DMUR&T19 0.260 0.484 0.538 DRS DMUR&T20 0.455 0.731 0.623 DRS DMUR&T21 0.593 0.676 0.877 DRS DMUR&T22 1.000 1.000 1.000 CRS
DMUR&T23 0.315 0.339 0.929 IRS
DMUR&T24 0.177 0.178 0.991 IRS
DMUR&T25 0.563 0.599 0.939 DRS DMUR&T26 0.938 0.983 0.954 DRS DMUR&T27 0.236 0.296 0.796 DRS DMUR&T28 0.336 0.369 0.909 DRS DMUR&T29 0.065 0.093 0.698 DRS DMUR&T30 0.122 0.152 0.805 DRS
DMUR&T31 0.840 1.000 0.840 IRS
DMUR&T32 0.406 0.440 0.923 DRS
DMUR&T33 0.766 1.000 0.766 IRS
DMUR&T34 1.000 1.000 1.000 CRS
DMUR&T35 1.000 1.000 1.000 CRS
DMUR&T36 0.150 0.157 0.954 DRS
DMUR&T37 0.811 0.819 0.991 IRS
DMUR&T38 0.808 1.000 0.808 IRS
DMUR&T39 0.642 0.797 0.806 DRS
DMUR&T40 0.261 0.263 0.992 IRS
development opportunities for domestic students and potential foreign members in order to increase the Inno-Qual scale effi-ciency. Regarding DMUT6, it is in the decreasing return to scale stage, and thus it should especially decrease the scale of domestic full-time faculty members or increase the number of publications it produces listed by the ESI and supported by the NSC. More de-tailed indications on improving Inno-Qual efficiency for teaching-intensive universities that are not at the efficient frontier are given inTable 8.
According to the results on profession-intensive universities presented inTable 9, the CRSTE, the VRSTE, and the SE of DMUP1, DMUP10, DMUP15, DMUP28, DMUP29, DMUP32, DMUP33, DMUP34, DMUP35, DMUP36, DMUP38, and DMUP41are at the efficient frontier and in a constant return to scale stage. Hence, they cannot increase Inno-Qual efficiency by adjusting their input. We suggest that these universities should increase interactions with government-owned or privately run corporations so as to increase advanced R&D opportunities. We also suggest that they upgrade to the re-search-intensive university or integrate characteristics from this type of university in order to enhance its ability to create invisible creation such as new thoughts or theories. By doing so, their com-petitive abilities will catch up to those of research-intensive uni-versities; additionally, future enhancements to the Inno-Qual
performance will be as smooth as those of research-intensive universities.
DMUP2, DMUP4, DMUP5, DMUP7, DMUP16, DMUP17, DMUP19, DMUP23, DMUP24, DMUP25, DMUP26, DMUP27, DMUP37, DMUP39, and DMUP42are in the increasing return to scale stage. Thus, they need to increase the scale of their output so as to increase the Inno-Qual scale efficiency, especially the quantity of financial support from the NSC and research patents. Also, we suggest that they de-crease the scale of their domestic full-time faculty.
Additionally, DMUP3, DMUP6, DMUP8, DMUP9, DMUP11, DMUP12, DMUP13, DMUP14, DMUP18, DMUP20, DMUP21, DMUP22, DMUP30, DMUP31, and DMUP40are in the decreasing return to scale stage. Therefore, they need to decrease the amount of input in order to increase Inno-Qual scale efficiency, such as decreasing the number of domestic full-time faculty. More detailed indications on improv-ing Inno-Qual efficiency for profession-intensive universities that are not at the efficient frontier are given inTable 10.
According to the results on research & teaching-intensive uni-versities presented inTable 11, the CRSTE, the VRSTE, and the SE of DMUR&T4, DMUR&T12, DMUR&T22, DMUR&T34, and DMUR&T35are at the efficient frontier and in a constant return to scale stage. There-fore, these universities can either transform into research-inten-sive universities, improving Inno-Qual performance, or focus
Table 12
The input and output slack for research & teaching-intensive universities.
University type Sample university Output Input No. of graduates (2008) Journals accepted and published (ESI) (2009) Quantity of financial support from theNSC (2009) Research patents (N > 20) (2004–07) No. of cooperating foreign countries (2008) No. of domestic students (2008) No. of international members (2008) No. of domestic full-time faculty (2008)
R&T DMUR&T1 10.396 0.000 7986.772 0.000 115.075 28.545 0.000 0.000
DMUR&T2 14.034 0.000 3.335 0.000 0.000 10.900 0.000 0.000 DMUR&T3 0.000 420.932 0.000 25.233 87.827 0.000 0.000 0.000 DMUR&T4 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T5 0.000 0.000 2583.163 0.000 117.485 0.000 0.000 300.138 DMUR&T6 1.942 0.000 2583.516 40.549 58.986 0.000 0.000 0.000 DMUR&T7 0.000 50.479 636.375 0.000 0.001 6.050 0.000 0.000 DMUR&T8 0.000 0.000 2047.789 0.000 15.865 0.000 0.000 332.265 DMUR&T9 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T10 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T11 0.000 0.000 821.192 0.000 48.850 0.000 0.000 0.000 DMUR&T12 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T13 0.000 0.000 1873.968 0.000 0.002 3.815 0.000 0.000 DMUR&T14 20.526 0.000 2428.154 0.000 86.515 0.000 0.000 125.107 DMUR&T15 49.382 0.000 1818.419 31.103 72.567 4.074 0.000 0.000 DMUR&T16 6.376 0.000 2661.637 0.000 68.274 0.000 0.000 595.588 DMUR&T17 12.075 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T18 37.113 0.000 0.000 0.000 104.501 11.571 0.000 51.055 DMUR&T19 45.427 0.000 2688.065 8.063 98.652 0.000 0.000 287.807 DMUR&T20 32.650 0.000 1080.112 22.876 18.866 4.958 0.000 69.375 DMUR&T21 4.238 0.000 3727.351 76.901 56.698 0.294 0.000 0.000 DMUR&T22 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T23 0.000 79.700 2131.845 0.000 0.000 0.000 0.000 0.000 DMUR&T24 0.000 0.000 3556.286 22.566 85.528 0.000 0.000 223.454 DMUR&T25 0.431 0.000 1010.868 0.000 110.518 0.000 0.000 73.267 DMUR&T26 0.000 897.979 9.211 0.020 0.000 0.000 0.000 0.000 DMUR&T27 0.000 0.000 0.000 37.838 44.335 0.000 0.000 556.211 DMUR&T28 0.000 0.000 1519.317 11.179 113.615 0.000 0.000 295.530 DMUR&T29 0.000 0.000 6396.000 0.000 112.615 0.000 0.000 247.201 DMUR&T30 0.000 0.000 2504.261 0.000 75.735 0.000 0.000 344.804 DMUR&T31 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T32 0.000 0.000 505.859 1.830 121.636 0.000 0.000 183.028 DMUR&T33 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T34 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T35 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T36 0.000 0.000 2118.905 24.670 74.235 0.000 0.000 0.000 DMUR&T37 0.000 0.000 2718.448 49.853 93.368 0.000 0.000 169.083 DMUR&T38 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 DMUR&T39 0.000 0.244 0.000 0.243 0.000 0.000 0.000 4.641 DMUR&T40 0.000 0.000 1345.441 2.385 10.819 0.000 0.000 0.000 DMUR&T41 0.000 72.000 0.000 6.000 0.000 0.000 0.000 597.000
more on the quality of their research by publishing papers in high-level journals listed by the ESI.
DMUR&T3, DMUR&T5, DMUR&T7, DMUR&T9, DMUR&T10, DMUR&T11, DMUR&T13, DMUR&T23, DMUR&T24, DMUR&T31, DMUR&T33, DMUR&T37, DMUR&T38, DMUR&T40, and DMUR&T41are in the increasing return to scale stage. They should increase Inno-Qual efficiency by increasing their output, the amount of financial support from the NSC and the internal development opportunities for foreign mem-bers in particular. In addition, they should decrease the number of domestic full-time faculty members.
Except above, DMUR&T1, DMUR&T2, DMUR&T6, DMUR&T8, DMUR&T14, DMUR&T15, DMUR&T16, DMUR&T17, DMUR&T18, DMUR&T19, DMUR&T20, DMUR&T21, DMUR&T25, DMUR&T26, DMUR&T27, DMUR&T28, DMUR&T29, DMUR&T30, DMUR&T32, DMUR&T36, and DMUR&T39are in the decreasing return to scale stage. Hence, they should decrease their amount of input, number of domestic full-time faculty, and number of domestic students in order to improve Inno-Qual scale efficiency. Also, they need to increase the quantity of financial sup-port from the NSC and the number of student development oppor-tunities in order to increase the Inno-Qual scale efficiency. More detailed indications on improving the Inno-Qual efficiency for re-search & teaching-intensive universities that are not at the efficient frontier are given inTable 12.
According to the results on education-in-practice-intensive uni-versities presented inTable 13, the CRSTE, the VRSTE, and the SE of DMUEP1are equal to one and thus at the efficient frontier as well as in the constant return to scale stage. However, since education-in-practice-intensive is the newest type of university, and based on the data inTable 4, we found that the amount of input and output of DMUEP1is relatively low as compared with other types of uni-versities. We assume that the Inno-Qual performance of DMUEP1 is not very good. We suggest that this university adopt translation or characteristics integration in order to become like a profes-sional-intensive university on account of its similar characteristics and improve its Inno-Qual performance and competitive advan-tage in the Taiwanese academic community.
5. Conclusions
Since the birth rate in Taiwan is decreasing, the number of uni-versities is increasing and Taiwan has joined the WTO, an increas-ing number of Taiwanese universities have recently tried to upgrade their innovation performance and improve their total quality performance (Inno-Qual performance) so as to enhance their overall operational performance. Although there are many relative measurement models, such as the Inno-Qual performance system (IQPS), which integrate the features of innovation and TQM, currently, no studies empirically evaluate the efficiency of such improvement, resulting in the Inno-Qual performance improving along with the cost increasing, particularly human administrative cost due to the nature of higher education, providing intellectual product. To overcome this problem, we adopted the IQPS using data envelopment analysis (DEA) to evaluate the Inno-Qual effi-ciency of five types of universities for a total of ninety-nine univer-sities in Taiwan. Based on the empirical results, we found that over half (73%) of the universities are highly inefficient in improving the Qual performance, and we conclude that improving Inno-Qual efficiency in accordance with our results will help to reduce the majority of cost expenditures.
References
Acs, Z. J., Anselin, L., & Varga, A. (2001). Patents and innovation counts as measures of regional production of new knowledge. Research Policy, 31, 1069–1085.
Banker, R. D., Charnes, A., & Cooper, W. W. (1984). Some models for estimating technical and scale efficiencies in data envelopment analysis. Management Science, 30(9), 1078–1092.
Bosworth, D., & Rogers, M. (2001). Market value, R&D and intellectual property: An empirical analysis of large Australian firms. The Economic Record, 77(239), 323–337.
Charnes, A., Cooper, W. W., & Rhodes, E. (1981). Evaluating program and managerial efficiency: An application of DEA to program follow through. Management Science, 27(6), 668–697.
Chen, J. K., & Chen, I. S. (2008a). A pro-performance appraisal system for the university. In V. Kieber (Chair). Germany, France, & Switzerland conference, Gottenheim, Germany.
Chen, J. K., & Chen, I. S. (2009a). An innovation support system for Taiwanese higher education using a novel conjunctive MCDM approach. In D. Ort (Chair). Conference of the international journal of arts & sciences proceeding (Vol. 1(15), CD Version). Toronto, Ontario, Canada: Ryerson University.
Chen, J. K., & Chen, I. S. (2009b). The construction of a network hierarchical feedback system for Taiwanese universities: The perspective from total quality management and innovation. In W. Tafoya (Chair). WorldFuture 2009: Innovation and creativity in a complex world, Chicago Hilton, Chicago, Illinois, USA. Chen, J. K., & Chen, I. S. (submitted for publication). A solid Inno-Qual performance
system for higher education – A novel hybrid fuzzy model application. Chen, J. K., & Chen, I. S. (2008b). VIKOR method for selecting universities for future
development based on innovation. Journal of Global Business Issues, 2(1), 53–59. Chen, J. K., & Chen, I. S. (2009c). TQM measurement model for the biotechnology
industry in Taiwan. Expert Systems with Applications, 35(5), 8789–8798. Chin, J. M., & Pu, S. W. (2006). The concepts and implementations of school
innovation management. Journal of Educational Research and Development, 2(3), 123–150.
Daft, R. L. (2004). Organization theory and design. NY: West.
Department of Higher Education (2004). Operation book. Taipei: Taiwan Assessment and Evaluation Association.
Department of Statistics. (2008). Basic file database for each school. <http:// www.edu.tw/statistics/content.aspx>Retrieved 16.06.09.
Dinh, T. H., Barbara, I., & Tritos, L. (2006). The impact of total quality management on innovation Findings from a developing country. International Journal of Quality and Reliability, 23(9), 1092–1117.
Dzinkowski, R. (2000). The measurement and management of intellectual capital: An introduction. Management Accounting, 78(2), 32–36.
Escrig-Tena, A. B. (2004). TQM as a competitive factor: A theoretical and empirical analysis. International Journal of Quality and Reliability Management, 21(6), 612–637.
Fairweather, J. S. (2000). Diversification or homogenization: How markets and governments combine to shape American higher education. Higher Education Policy, 13, 79–98.
Farrell, M. J. (1957). The measurement of productive efficiency. Journal of the Royal Statistical Society, 120(3), 253–290.
Gambardella, A., & Torrisi, S. (2000). The economic value of knowledge and inter-firm technological linkages: An investigation of science-based firms. Dynamo TSER Project (Contract No. SOE1-CT97-1078).
Guthrie, J., & Petty, R. (2000). Intellectual capital: Australian annual reporting practices. Journal of Intellectual Capital, 1(3), 241–251.
Hall, L. A., & Bagchi-Sen, S. (2002). A study of R&D, innovation, and business performance in the Canadian biotechnology industry. Technovation, 22, 231–244.
Han, S. B., Chen, S. K., & Ebrahimpour, B. (2007). The impact of ISO 9000 on TQM and business performance. Journal of Business and Economic Studies, 13(2), 1–21. Higher Education Evaluation and Accreditation Council of Taiwan. (2008). Study on
weigh of patents. Evaluation Bimonthly, 14, 29–34.
Higher Education Evaluation and Accreditation Council of Taiwan. (2009). Performance statistics.<http://www.heeact.edu.tw/lp.asp>Retrieved 16.06.09. Ismail, S. (2006). Examining the effects of contextual factors on TQM and
performance through the lens of organizational theories: An empirical study. Journal of Operations Management, 25(2007), 83–109.
Keng, B. O., Nooh, A. B., Veeri, A., Lorraine, V., & Loke, A. K. Y. (2007). Does TQM influence employees’ job satisfaction: An empirical case analysis? International Journal of Quality and Reliability Management, 24(1), 62–77.
Kenneth, M. Y., & Cynthia, E. M. (2004). Causation or covariation: An empirical re-examination of the link between TQM and financial performance. Journal of Operations Management, 22(2004), 291–311.
Krause, D. E. (2004). Influence-based leadership as a determinant of the inclination to innovative and of innovate-related behavior: An empirical investigation. The Leadership Quarterly, 15, 79–102.
Li, U. C. (2007). Enter university by 2.8 point each subject.<http://news.msn.com.tw/ print.aspx?id=210245>Retrieved 18.08.07.
Lin, H. F., Wang, H. L., Wang, C. M., & Yen, J. F. (2006). The relationship between elementary principle leadership, continuous knowledge management and innovative operation-take three cities in northern Taiwan as example. In Education development in Chinese society conference, University of Macau, China.
Table 13
Efficiency value and return to scale for education-in-practice-intensive universities.
Sample university CRSTE VRSTE SE RS
Meek, V. L. (2000). Diversity and marketisation of higher education: Incompatible concepts. Higher Education Policy, 13, 23–39.
Mei, Y. F., & Lee, L. S. (2006). A study to develop performance indicators for colleges of technology in Taiwan. Journal of Education Research, 1(1), 25–67.
National Science Council. (2009). Academy research.<https://nscnt12.nsc.gov.tw/ OPEN_QUERY/OPENQUERY00.aspx>Retrieved 16.06.09.
Nusrah, S., Ramayah, T., & Norizan, M. S. (2006). TQM practices, service quality, and market orientation: Some empirical evidence from a developing country: Some empirical evidence from a developing country. Management Research News, 29(11), 713–728.
Ordaz, C. C., Lara, A. B. H., & Cabrera, R. V. (2005). The relationship between top management teams and innovative capacity in companies. The Journal of Management Development, 24(8), 683–704.
O’Sullivan, M. A. (2000). Contests for corporate control: Corporate governance and economic performance in the United States and Germany. NY: Oxford University Press.
Ozden, B., & Birsen, K. (2006). An analytical network process-based framework for successful total quality management (TQM): An assessment of Turkish manufacturing industry readiness. International Journal of Production Economics, 105(2007), 79–96.
Schoenecker, T., & Swanson, L. (2002). Indicators of firm technological capability: Validity and performance implications. IEEE Transactions on Engineering Management, 49(1), 36–44.
Shephard, R. W. (1970). Theory of cost and production functions. NJ: Princeton University Press.
Subramaniam, M., & Youndt, M. A. (2005). The influence of intellectual capital on the types of innovative capabilities. Academy of Management Journal, 48(3), 450–463.
Taiwan Assessment and Evaluation Association. (2006). Evaluation Bimonthly, 1, 48– 49.
Tang, C. M. (2006). School building and campus planning. Taipei: Wu-Nan Press. Toivanen, O., Stoneman, P., & Bosworth, D. (2002). Innovation and the market value
of UK. Oxford Bulletin of Economics and Statistics, 64(39), 39–61.
Van Buren, M. E. (2000). Making knowledge count: Knowledge management systems and the human element. Unpublished manuscript, University of Missouri, Columbia.
Wanger, M., & Schaltegger, S. (2004). The effect of corporate environment strategy choice and environmental performance on competitiveness and economic performance: An empirical study of EU manufacturing. European Management Journal, 22(5), 557–572.
