1、Omega 36 (2008) 653 analysis of university departments:An empirical studyH22817Chiang Kaoa, Hsi-Tai HungbaDepartment of Industrial and Information Management, National Cheng Kung University, Tainan, Taiwan, Republic of ChinabDepartment of Industrial Engineering and Management, Cheng Shiu University
2、, Kaohsiung, Taiwan, Republic of ChinaReceived 10 September 2005; accepted 7 February 2006Available online 30 March 2006AbstractUniversities play an important role in the development of a country in this age of the knowledge economy. As governmentsubsidies to universities have been reducing in recen
3、t years, the more efficient use of resources becomes an important issue foruniversity administrators. This paper applies data envelopment analysis (DEA) to assess the relative efficiency of the academicdepartments at National Cheng Kung University in Taiwan. The outputs considered are total credit-h
4、ours, publications, andexternal grants; and the inputs utilized by the departments are personnel, operating expenses, and floor space. An assuranceregion is constructed by the top administrators of the university to confine the flexibility in selecting the virtual multipliers inDEA. Four groups of d
5、epartments of similar characteristics are categorized via an efficiency decomposition and cluster analysis.The aggregate efficiency indicates whether the resources have been utilized efficiently by a department and the efficiencydecomposition helps identify the weak areas where more effort should be
6、 devoted so that the efficiency of the department canbe improved.H17015 2006 Elsevier Ltd. All rights reserved.Keywords: Cluster analysis; Data envelopment analysis; Efficiency; Efficiency decomposition1. IntroductionThe primary mission of a university is to explore andtransmit knowledge. The former
7、 is achieved via researchwhile the latter is achieved via teaching. In this age ofthe knowledge economy, universities play a key role inthe development of a country.Taiwan has an old tradition that well-educated peo-ple are respected by society. This tradition encouragesparents to send their childre
8、n to universities to studyfor bachelor and higher degrees. To cope with the highH22817This manuscript was processed by Area Editor Prof. B. Lev.Corresponding author. Tel.: +886 6 2753396;fax: +886 6 2362162.E-mail addresses: ckaomail.ncku.edu.tw (C. Kao),sdhungcsu.edu.tw (H.-T. Hung).0305-0483/$ - s
9、ee front matter H17015 2006 Elsevier Ltd. All rights reserved.doi:10.1016/j.omega.2006.02.003demand for education, Taiwans government allocatesmore than 15% of the nations budget to education, andmore than half of it goes to higher education. Manypeople believe that the striking economic development
10、of the last two decades, which has won Taiwan a placeas one of “the four dragons of Asia”, is attributed to thehigh education level of its people.In the past, national universities in Taiwan receivedall of their budgets from the government. At the sametime, national universities followed the rules a
11、nd regu-lations of the government completely. Little flexibilitywas granted to them. As more families wish theirchildren to receive a university education, many newuniversities have been established and many juniorcolleges promoted to universities in the last 10 years.From 1994 to 2003, the number o
12、f universities has dou-bled from 58 in 1994 to 143 in 2003, which shows a654 C. Kao, H.-T. Hung / Omega 36 (2008) 653664transition from elitism to mass higher education 1.This increment has greatly increased the financial bur-den on the government and reduced the funding that canbe allocated to each
13、 university. In 1996, the Ministryof Education (MOE) of Taiwan, Republic of China, ini-tiated the University Fund System to replace the oldOperation Budget System. Under the old system, a uni-versity had to return any unused money, including alltypes of income, to the government. The new system,in c
14、ontrast, allowed the universities to retain the un-used money for future use. Types of university incomeinclude government subsidies, tuition fees, grants andcontracts, private donations, profits of university invest-ments, etc. The subsidies from the government only ac-count for a portion of a nati
15、onal universitys budget andthe university must seek income from external sourcesto make up the deficit. For example, National ChengKung University only receives approximately 40% of itsbudget from the government. Similar to other countries,for example, Australia 2, Britain 3, and Israel 4, thesubsid
16、ies received from the government have been re-duced in recent years. Although national universities arenot completely supported by the government any more,they have gained more freedom in self-governance andmanagement of their finances. However, while morefreedom has been granted, national universit
17、ies in Tai-wan still face financial pressures. They must be morecautious in allocating their precious resources to aca-demic departments. Every professor, each dollar, andall other resources must produce satisfactory outputs.In this paper, the efficiency of the academic depart-ments at National Chen
18、g Kung University in utilizingthe scarce resources in teaching students and produc-ing research results is measured. The investigation isfocused on the efficiency of resource utilization ratherthan academic performance. This type of assessmentalleviates the problem of comparing departments withdiffe
19、rent characteristics. The results will reveal whichdepartments consume more resources in the university.Two types of performance evaluation at the univer-sity level have been extensively discussed in the liter-ature. One type compares the performance of differentuniversities. The focus is on cost ef
20、ficiency, researchproductivity, or aggregate performance. The studies ofAbbott and Doucouliagos 2, Johnes 5, Johnes andJohnes 3, Kao 6, Muiz 7, Post and Spronk 8,Ruggiero et al. 9, and Tomkins and Green 10 areexamples. Regression analysis and data envelopmentanalysis (DEA) are the major techniques u
21、sed for as-sessing the performance. Another type of evaluationcompares the teaching and research performance of thedepartments within a university. Typical examples in-clude Gander 11 and Sinuany-Stern et al. 4. The for-mer uses regression analysis and the latter applies DEA.In this paper, we use th
22、e BCC model 12 of DEA tomeasure the efficiency of the academic departments ofNational Cheng Kung University, a national universityin Taiwan, Republic of China. In addition to measuringthe efficiency of each department, the efficiency is de-composed into the efficiency contributions of individ-ual ou
23、tput factors. Based on this, the departments areclassified into groups of similar characteristics. The effi-ciency measures show the departments whose resourceshave been utilized inefficiently. The efficiency decom-position helps the departmental head identify its weakareas for future improvement.Th
24、is paper is organized as follows: in the next sec-tion, the input and output factors adopted for measuringthe efficiency of the academic departments at NationalCheng Kung University in utilizing the resources aredescribed. Section 3 applies DEA to measure the rela-tive efficiency of each department
25、and make appropriatedecompositions. Based on the efficiency measures andtheir decompositions, a cluster analysis is conducted inSection 4 to distinguish the dissimilarity of the depart-ments. Finally, a conclusion is drawn in Section 5.2. Inputoutput factorsAlthough teaching and research have been c
26、onsideredby most people as the two major tasks of the univer-sity, they are difficult to measure. We need some indica-tors which are capable of representing the achievementof these two tasks. We also need measures of the re-sources that the department has consumed in perform-ing those two tasks. The
27、 selection of input and out-put factors for evaluating the performance of univer-sity departments using DEA has been discussed in sev-eral studies 24,8,10. There are at least two difficul-ties in deciding on the indicators. One is the availabilityof data. For example, some scholars suggest using the
28、salary of the first job after graduation as a measure ofthe achievement of teaching. Unfortunately, these dataare very difficult to acquire. Besides, different profes-sions have different salary standards. It would be unfairto compare the salary of an elementary school teacherwith that of a medical
29、doctor. The other difficulty is themeasurement of quality, there lacks a common base forcomparing the quality of different research work andsubjectivity is usually involved. Since the objective ofthis paper is to investigate the work load and effort de-voted to teaching and research, only quantifiab
30、le mea-sures are considered.C. Kao, H.-T. Hung / Omega 36 (2008) 653664 655National Cheng Kung University (NCKU) is the sec-ond largest university in Taiwan in terms of the fac-ulty size, with a total number of 20 000 students. It hasbeen selected by industry as The Most Favorite Univer-sity for 11
31、consecutive years according to the surveyconducted by Commonwealth, a famous magazine inTaiwan, which means its graduates are the first choicefor industry to hire. Teaching has been traditionally avery important task of this university. As indicated byHiggins 13, the most difficult area among the in
32、ter-nal performance indicators is that of teaching quality.The output performance of students may be the resultof students initial ability which they have already ac-quired before entering the university. Student evaluationfor teachers may be biased by the nature of courses anddoes not have a common
33、 base for comparison if the stu-dents have not been taught by all teachers. Due to thesedifficulties and following the work of Sinuany-Sternet al. 4, the top administrators (including the president,vice presidents, and college deans) of NCKU decidedto use total credit-hours to represent the achievem
34、entin teaching. Academic departments are the units to beevaluated. By total credit-hours we mean the numberof credits of a course multiplied by the number of stu-dents taking that course summed over all the coursestaught by the faculty of a department. Gander 11 givesa weight of 3 to graduate course
35、s without giving a rea-son. In Taiwan, when a baccalaureate program is ap-proved, usually 50 students are enrolled each year andthe MOE gives 16 faculty positions to the university.Each student finishes approximately 144 credits beforegraduation. Hence, the teaching load allocated to eachfaculty mem
36、ber in teaching undergraduate courses eachyear is144(credits) 50(students)/16(faculty members)= 450(total credit-hours).In contrast, when a master program is approved, usu-ally 30 students are enrolled each year and the MOEgives five faculty positions to the university. Each stu-dent on average fini
37、shes 36 credits before receivingthe master degree. In this case, the load allocated toeach faculty member in teaching graduate courses eachyear is36(credits) 30(students)/5(faculty members)= 216(total credit-hours).These two figures show that the effort in teachinggraduate courses is approximately t
38、wice (450 versus216) that of teaching undergraduate courses. To makethem comparable, a weight of 2 is assigned to graduatecredits to form undergraduate-equivalent credits in theevaluation.Regarding research, there are many indicators thathave been used to measure its achievement, of whichthe number
39、of publications is probably the most widelyaccepted one, and this indicator is also adopted in thispaper. Publications include refereed journals, books,conference proceedings, monographs, etc. However,these all have different quality and impact, and differ-ent areas do not have a common base for com
40、parison.To eliminate this effect to the greatest extent, the gen-eral consensus of the top administrators at NCKU isto count only the articles published in those journalson the journal list of Science Citation Index (SCI),Social Sciences Citation Index (SSCI), and Arts andHumanities Citation Index (
41、A this factor is not important enoughto be treated as a separate variable.The second important resource is the operatingexpenses allocated to the department from the uni-versity. The expenses incurred in operating a depart-ment include the procurement and maintenance of theequipment and facilities,
42、salaries paid to the adjunctprofessors and part-time workers, telephones, travel,and all other types of expenditures. The salaries offull-time faculty and staff are paid by the Account-ing Office of the university directly, and are not in-cluded. Note that the expenses covered by the grantsare not c
43、ounted because they are not the universitysmoney.In addition to personnel and operating expenses, an-other resource considered by the top administrators atNCKU as an important input factor is the amount ofspace used by the department. NCKU has a formula forallocating space to each department. Howeve
44、r, due tosome historical reasons, indivisibility of space blocks,donated buildings from industry, etc., some departmentshave relatively more space than others. For those de-partments which have less space than that calculatedfrom the formula, the university will try to find spacefor them. Even so, d
45、epartments are always asking formore space. The best way to resolve this conflict is touse the floor area used by the department as an inputfactor. Departments with more floor area are expectedto achieve more in teaching and research. Floor area isexpressed in square meters.In sum, three input facto
46、rs are used in this paper tomeasure the efficiency of the department. They are per-sonnel expressed in number of faculty-equivalent per-sons (X1), operating expenses in 1000 NTD (X2), andfloor area in square meters (X3).NCKU has 67 departments, of which 38 offer bac-calaureate degrees, 63 offer mast
47、er degrees, and 41offer doctoral degrees. Some departments have veryclose relationship with others due to academic contentsor historical reasons. For example, the Department ofEconomics and Department of Political Science onlyoffer baccalaureate degrees; however, all of their facultymembers are join
48、tly appointed to the Department ofPolitical Economy, where only post-graduate degreesare offered. It is difficult to separate their achieve-ments in teaching and research, and resources utilized.Therefore, in the evaluation, these three departmentsare aggregated into one department, Political Econom
49、y.This type of aggregation results in 41 departments.Table 1 lists the input and output measures of thesedepartments.3. Efficiency measurementTo measure the relative efficiency of a set of decision-making units (DMUs) that utilize the same inputs toproduce the same outputs, DEA has been demonstratedto be a suitable method. Basically, DEA is a mathe-matical programming technique that produces a singleaggregate measure for each DMU in terms of its uti-lization of inputs to produce desired outputs
