INTRODUCTION
Data envelopment analysis (DEA) was introduced by Charnes
et al. (1978). One of the scores of DEA method is lack of necessity
of definition of relationship depends between inputs and outputs. It was originally
intended for use as a performance measurement tool for organization that lacked
a profit motivation. However since its introduction, it has been developed and
expanded for a variety of uses in forprofit as well as notforprofit situation.
Measuring performance can be simultaneously done on the basis of quantitative
and qualitative concept of system.
The history of DEA application in educational assessment: In recent years, numerous studies were carried out for universities performance assessment using DEA method. Universities are institutes with many inputs and outputs. At least two different outputs of education and research are products of each university. A number of most important studies done in this field are as follow:
Rodhes and Southwick (1986); carried out an analysis
on the efficiency of private universities in comparison to state universities
in USA through the DEA model which all the country's universities were viewed
as DMU. Tomkins and Green (1988) used DEA for 20 accountancy
departments test in UK. DMUs were universities accountancy departments in this
case which studies from two grounds of education and research were. Harris
(1990), carried out a study in relation to research performance in Australian
Economics University. DMUs were the Australian Economics departments which were
analyzed from the research programme perspective.
The economic efficiency of university department also studied (Hashimoto
and Cohn, 1997). Melville and Debasish (1998) published
the result of using Data Envelopment Analysis (DEA) to assess the relative efficiency
of 45 Canadian universities. These results were obtained from nine different
specifications of inputs and outputs. Sarrico and Dyson
(2000) explored the contribution of DEA methodology to inform management.
They described Warwick University's performance. The Australia universities
technical and scale efficiency was also carried out by Abbott
and Doucouliagos (2003). Martine (2003) also used
this method to investigate Zaragoza University's teaching departments' performance
assessment. Chapple et al. (2005) extended evidence
on the relative performance of U.K. university Technology Transfer Offices (TTOs)
using Data Envelopment Analysis (DEA) and Stochastic Frontier Estimation (SFE).
DEA MODELS
Data envelopment Analysis model is a nonparametrical model which it is usable
for different systems performance assessment. One of the most important scores
of DEA is its usability for systems having multiinputs and multioutputs. If
the number of inputs and outputs are more than one, sum inputs and outputs multiplications
are used:
• 
Virtual output = u_{i}y_{io}+u_{2}y_{2o}+...+u_{s}y_{so} 
• 
Virtual input = v_{1}x_{1o}+vx_{2o}+...+v_{s}x_{so}. 
To compare homogenous systems with identical inputs and outputs, we need a
mathematical model. This model was propounded as the CCR model by Charnes
et al. (1978). The optimum coefficients for each of the DMUs are
separately calculated so that the most efficiency for the viewed DMU is obtained.
In this model, the maximum amount in relation to virtual output to virtual input
is assessed for each DMU, so that the efficiency of none of DMUs should not
be more than a (100% efficiency) while all the variables multipliers should
be positive. This model is formulated as follow:
Max:
Subject to:
The ε≥0 is a nonArchimedean constant. In the light of the above model, the efficiency score of each DMU is bigger than zero and smaller or equals one. A DMU is a perfect efficient, if and only if the improvement possibility of none of inputs and outputs exist without worsening other inputs and outputs (ParetoKoopmans Efficiency). The optimal values v_{i}* and u_{r}* may be interpreted as weights (v_{i} and u_{r} are the variables of the above models). These values are determined through solution of the model. θ_{o}* is highest rating that allow for a DMU.
It also should be noted that DEA assesses relative efficiency, i.e., estimation the influence of DMUs that have complete efficiency.
The model (1) is changed to a linear planning through changing a variable as follow:
Maximize:
Subject to:
As it is stated by Charnes et al. (1978), this
implies that the conditions for ParetoKoopman optimality, because increase
in this maximum value is obtained only when some values input (x_{ij})
is increased or if some of the values output (y_{rj}) are reduced (Cooper
et al., 2007).
EDUCATIONAL EVALUATION THROUGH DEA METHOD
To evaluate educational system cannot be used of market evaluation mechanisms such as benefit assessment to determine DMU performance or inputs and outputs economic value, because inputs and outputs generally stand in the education, research and service departments which the measurement or presentation of an assessment unit is very difficult. DEA method also emphasizes university targets for inputs and outputs choice and makes possible the choice of qualities input and output indicators to the system. There is also the permission for the choice of several inputs and outputs.
Indicators (variables): The choice of indicators or in other words, effective factors on efficiency assessment in a university has great importance, because to succeed in performance assessment, precise choice and suitable one in selecting important factors and comparable in units under consideration should be carried out.
In this study, the application of DEA method in educational assessment, Islamic Azad University, Zahedan unit's educational departments in 20082009 academic years is studied. The definition of indicators which show produced output in one educational unit or inputs to be viewed are not simple due to numerous effective factors. The Islamic Azad University, Zahedan unit's educational departments are viewed as DMUs. The effective indicators on educational department's efficiency can be the number of teaching staff (including: professors, assistant professor, lectures and educational experts, who are fulltime and halftime), the performed research work by the teaching staff of each educational departments (including: books compilation and translation, published articles or presented in authentic conferences and reports and research projects) the number of registered students, the number of graduates, the number of passed students to higher levels (such as a twoyears course to BA, BA to MA and to PHD), educational possibilities etc.
Table 1: 
The gathered information to assess educational departments
of Islamic Azad University, Zahedan Unit, Academic year 20082009 

Table 2: 
Efficiency of DMUs in CCR model 

In the light of the abovementioned, the gathered information are written in the Table 1.
By using the above table's information in the DEA linear programming models (CCR model) the following results are obtained (Table 2).
The above table yields that eight DMUs are efficient but their rank is undefined in comparative together, also two DMUs have average ratings greater than 0.90. The Arabic literature department has minimum efficiency equal 0.366.
THE EFFICIENCY ASSESSMENT IN PESSIMISTIC VIEWPOINT
In most DEA models, weights for each DMU are so calculated that the most efficiency
for that DMU is obtained, whilst weights are calculated separately for each
DMU. So this kind of assessment can be called optimistic assessment (Liang
et al., 2008). In this section, we deal with pessimistic viewpoint
of efficiency and through using this method, interval efficiency each DMU is
calculated.
We, in this article, use the multiplication efficiency method to calculated pessimistic efficiency. In this method, DEA model is solved for each DMU to get the optimistic efficiency. Now, if the value of obtained variables for other DMUs to be viewed in dependent objective of fractional DMU with the selection of least pessimistic efficiency in identical conditions for all DMUs is calculated, i.e., θ_{kj} scores which are the jth DMU efficiency scores are determined against kth DMU weights:
All matrix components multiplication efficiency is between zero and one, 0θ_{kj}≤1
and diagonal multiplication efficiency matrix (θ_{kk}), is the
DEA efficiency score usual for (optimistic efficiency). If DMU_{k} is
efficiency θ_{kk} = 1 and otherwise inefficiency, the least existing
value in this matrix can show the worst conditions for the efficiency of each
DMU (θ_{o}^{l*}). In this method the pessimistic efficiency
is smaller than optimistic efficiency and because efficiency of each DMU even
in the worst conditions will not be zero, so none of the DMUs have interval
efficiency [0, 1].
This method (Crossefficiency evaluation) not only provides a ranking among the DMUs but also eliminates unrealistic DEA weighting schemes without requiring a priori information on weight restrictions.
INTERVAL RANKING
When the efficiency of each DMU is merely calculated optimistically, the efficiency score of each DMU is the DMUs rank which is of course methods such as AndersenPetersen's method should be used to rank efficient units. Now, that we have obtained the efficiency of each DMU in the form of interval, it is obvious that ranking on the basis of optimistic efficiency causes the loss of a lot of information.
Definition:
• 
DMU_{o }is a strong efficient whenever θ_{o}^{l*}
= θ_{o}^{u*} = 1 
• 
DMU_{o }is efficient whenever θ_{o}^{l*}
<1, θ_{o}^{u*} = 1 
• 
DMU_{o }is inefficient whenever θ_{o}^{u*}<1 
For interval efficiency ranking many methods can be used: First DMUs are ranking
on the basis the upper limit and if DMUs have identical upper limit, ranking
is carried out with comparison to low ranking. If [θ_{o}^{l*},
θ_{k}^{u*}]≤[θ_{t}^{l*}, θ_{t}^{u*}],
DMU_{t} is rated as more efficient than DMU_{k}.
Optimistic coefficient method, in this method to assess the efficiency of each DMU, a set of multiplications from the upper limit and the lower limit are used. The coefficients are determined in the light of decision maker and the existing conditions. Meanwhile the collection of two coefficients (the upper limit coefficient and the upper limit coefficient) should become one.
INTERVAL EFFICIENCY IN PERFORMANCE ASSESSMENT OF EDUCATION
By using the above method, the section 3 data, interval efficiency of educational departments is as Table 3.
If ranking is first carried out on the basis of the upper limit and about departments having identical the upper limit, the ranking will be like the above table.
If optimism coefficient is used for ranking, the upper limit and the lower limit interval efficiency, each should be multiplied by an appropriate number explaining the decision maker's optimism, so that the collection of two chosen numbers become one as coefficient.
By using the questionnaires and assessing the views of authorities and the students, optimism coefficient for each department is determined and then the average of these coefficients were chosen as optimism coefficient, the resulting value 0.64 is for optimism and 0.36 for pessimism.
CONCLUSION
Universities are in the present arena, one of the major sponsors of education
and provision of efficient manpower needed for the country. So, assessment of
their performances and determination of their weak and strong points can continuously
be effective in reaching to their aims. Researchers in their studies concerning
the performance of the set of training organizations which train society's manpower
are looking after indicators which can compare them on the basis of these indicators.
These indicators can be quantitative or qualitative. The indicators should be
chosen so that they reflected the institution's performances. Generally speaking,
the trained force in the institution and also the rate of learning concepts
and the application of science by graduates and individuals' thought results
in the form of articles, books and other tools, reflects graduates' knowledge,
awareness and the various sciences that they can be suitable indicators for
further assessment of an educational institution. Therefore, performance assessment
indicators: number of service receiving students in an academic year, number
of graduates, number of passing students, scientific board's concession, guest
lecturers, the concession of chosen research work were selected. The unit under
investigation was all Islamic Azad University, Zahedan unit's all educational
departments in all levels of two year post university course, B.A and M.A. to
carry out a more accurate assessment and also efficiency was determined with
two optimistically and pessimistically viewpoints.
Table 3: 
Interval efficiency of DMUs and Ranking of DMUs with using
optimistic coefficient 

Then interval efficiency is calculated to determine optimistic efficiency
from CCR model and pessimistic efficiency through crossefficiency. This method
is better in relation to methods used so far to determine the pessimistic efficiency,
because despite the IDEA model, a DMU which stands both the border of efficiency
and inefficiency does not occur, while in other model DMU is resulted with interval
efficiency which shows the best efficiency in optimism case that is opposes
the efficiency CCR's model concepts. Finally, we have dealt with using interval
ranking concepts to educational departments ranking.