INTRODUCTION
Organizational Development (OD) as a definite field and a distinguished research
area in management science does not have a long history. Unlike medicine, accounting,
law and other professions and scientific fields, all of which have clear and
fixed goals, the area of OD is fairly unique due to its inherent and fundamental
vagueness. OD, despite its steady and constant growth, is a discipline that
is still challenging an unresolved problem pertaining to its inherent nature,
which is a combination of organizational sciences and individuals. Although,
the theoreticians of organizational improvement and development have contemplated,
complied and discussed its fundamental principles, there is no consensus as
to its scope and parameters. In a more moderate sense it could be said that
there is much disagreement on the nature of OD. The term OD was used for the
first time in 1957 by Blake and Mouten. The OD consists of numerous concepts
and methods for improving the organization. Also, several definitions have been
suggested for OD and the diversity of approaches and techniques has caused the
indefiniteness of its limits and the difficulty of its analysis (Boonstra,
2004). Beckhard (1969) defined it as an attempt which
is planned, takes place all over the organization and which is directed by the
top level of organization, in order to enhance the effectiveness of the organization
through planned intervention in the organizational processes by exploiting behavioral
sciences. Schein (1969) defines it as an interpersonal
process in order to facilitate group dynamics. Philbin and
Mikush (1999) defined it as a process through which an organization enhances
its internal capacity so it can reach its maximum efficiency specified in the
organization's mission and can maintain it in the long run. Burke
(1982) defined it as a planned process of changing the organization's culture
by exploiting the theory, research and technology of behavioral sciences. Porras
and Robertson (1992) defined it as a series of programmed and knowledge
based behavioral actions in the working environment in order to improve the
organization's operation and growth and promotion of the individuals. Jamieson
et al. (1984) defined it as planned and longterm changes in the
culture, technology and management of an organization. French
and Bell (1978) define it as an attempt to improve problemsolving methods
and renewing the processes in the organization, especially through the effective
management of the organizational culture, with the helper of the change expediting
factors and with the use of the theory and technology of applied behavioral
sciences.
Warrick (1984) defines it as a series of planned and
longterm systems consisting of behavioral sciences strategy in order to understand,
develop and change the organization so as to promote its effectiveness and health
at present and in future. Harvey (1974), too, in a different
definition likens OD to a revolutionary movement that is carried out by a militant
leader in order to democratize organizations. Weick and
Quinn (1999), in their research organizational change and development, examine
the different types of organizational change and development. They suggest that
the recent analyses of organizational change imply the increasing importance
of the speed and time of the changes. Also they believe that there are two types
of changes: Emergent and Continuous, which are in contrast from
the viewpoint of implicational metaphors and concepts, such as organization,
analytical frameworks, ideal organizations, intervention theories and the role
of change factors. In the emergent change, the sequence is freeze, transition,
unfreeze, while continuous change follows the sequence unfreeze, rebalance,
freeze. In their view, the way of understanding the necessity of change determines
whether the change is considered emergent or it is seen continuous.
Therefore, with regard to the cited definitions we suggest the following characteristics
for a developed organization:
• 
Information and knowledge are not exclusive; the organization
opens the communication doors to all who live or work in it. It cause the
growth and flourishing and enhancement of knowledge and awareness of all
people and it encourage its members to acquire and increase their knowledge
and awareness. It also adopts the open door policy and elucidates all its
operations to its personnel, customers or clients and even the society 
• 
Trust which means believing in other people's honesty; if the members
of a group, family or organization trust each other and believe in each
other's thoughts and deeds, the power and strength of that group will increase.
Trusting others is a characteristic, which is related to the openness of
a system. That is, the more lucidly an organization behaves in its operations,
the more it gains others' trust and the more it increases the sense of trust
in the working environment 
• 
It gets feedback from all the sources and factors that affect its destiny
and based on the feedback, it modifies its actions and movements in order
to achieve predetermined objectives. That is, the strategy of feedback
which by means of it, an individual, a family or an organization can gain
information from inside or outside and use it in order to conduce its
movement forward or to correct it in order to attain its goal

• 
It opens the doors of participation, give and take of information and
thoughts to all its members and encourages the members to cooperate and
participate with each other. As a result, managing an organization does
not fall on a small number of members; rather all the members according
to their effort and their mental and physical capabilities help to carry
on the operations of the organization and they regard the organizational
objectives as their own objectives

• 
It cares about training and empowering its members and by providing the
right means, it prepares them to achieve operational independence. In other
words educating, giving awareness and training the members to become independent
and autonomous and finally giving them the authority to manage independently,
free of other's views or supervision, are some of the developed organization's
characteristics 
• 
It naturally uses a few organizational layers and it reduces the gap between
the higher and lower layers. This causes the members of the organization
to communicate more easily; it also causes the work relationships and associations
to assume a higher human nature 
At last, we can specify the following common characteristics for developed
organizations:
C_{1} = Lucid information
C_{2} = Lucid communication
C_{3} = Learning
C_{4} = Being honesty
C_{5} = Team working
C_{6} = Competency
C_{7} = Individual independence
C_{8} = Lucid tasks
C_{9} = Better supervision
In this study, in order to prioritize organizational change situations with regard to these characteristics criteria, we used group AHP with satisfying Pareto optimality.
MATERIALS AND METHODS
In the organization under study, Governmental Educational Organization of Tabriz
in Iran in 2008, there are different situations for organizational change, which
we want to prioritize by the criteria mentioned in the previous section. These
situations are as follows:
A_{1} 
= 
Further adaptation of the organizational space to individual
needs 
A_{2} 
= 
Changing the cultural norms 
A_{3} 
= 
Changing the structure and composition of roles 
A_{4} 
= 
Enhancing the intergroup cooperation 
A_{5} 
= 
Transparenting the communication system 
A_{6} 
= 
Changing the motivation of the workforce 
In order to prioritize these situations (alternatives), a number of teachers
holding bachelor or higher degrees were chosen, using random sampling (Eq.
1):

Fig. 1: 
Hierarchical tree 
A questionnaire was used to collect the data. The questionnaire was devised
on the basis of the information required for AHP model. The hierarchical tree
related to the problem is shown in Fig. 1. In each of the
questions in the questionnaire, the respondents were asked to make a pairwise
comparison between two factors and specify the priority of one over the other.
In order to determine the reliability of the questionnaire the retest method
was used. For this purpose, the final questionnaire was randomly distributed
among 20 members of the statistical population at two different times within
an interval of two weeks. After gathering the necessary data at two times, first
the geometrical mean resulting from the subjects' attitude toward each factor
at each time was calculated by means of Eq. 2 (Aczel
and Saaty, 1983):
Finally, Spearman's rank correlation coefficient ρ (rho) was calculated and the significance of it in the statistical population was tested (Eq. 3):
In order to generalize the findings of the sample to statistical population,
we can approximate the rank correlation coefficient ρ (rho) with normal
distribution for values n>10, in which case expected value and variance will
be needed. Expected value and variance ρ can be obtained from Eq.
4 and 5:
Aczel and Saaty (1983) showed that the only mathematical
method to mix pairwise comparisons is the use of geometric mean provided that
the resulting matrix also has all the characteristics of one pairwise comparison
matrix. Saaty (1994) suggested that in order to compare
the convergence of the views of each individual to the views of the group (geometric
mean), we can use compatibility index. If V=(V_{1}, V_{2},…,V_{n})^{T}
is the weights resulting from an individual's judgments and W=(W_{1},
W_{2}, …,W_{n})^{T} is the weights resulting from
the geometric mean, we can calculate compatibility index by using the hadamard
product of the two matrices (Eq. 6):
Saaty suggests that this value should not exceed 1.01. Exceeding this value
leads to violation from Pareto's optimality. According to this principle, if
all members prefer A to B, then the group should also prefer A to B. What leads
to this in practice is the dispersion of judgments around the geometric mean.
In other words, lack of dispersion in each one of the judgments around the geometric
mean leads to compatibility at the group level, but viceversa is not true.
Saaty and Vargag (2007) defined geometric dispersion
as follows (Eq. 7):
They computed the geometric dispersion of randomly generated samples of size 20,000 under the assumption that the judgments are distributed according to a continuous reciprocal uniform distribution:
for groups consisting of n individuals. They have concluded that with the increase
in the value of n, the distribution dispersion moves toward a distribution of
threeparameter gamma with the situation parameter (γ=1) and values α
(shape) and β (scale) (Eq. 8):
The values of α(shape) and β(scale) with γ = 1 and for n = 4, 5, …, 50 are shown in Table 1.
They have also presented a linear regression model to calculate the values of α and β for other values of n as follows (Eq. 9):
Thus, the geometric dispersion value obtained from the judgments of individual members on each pairwise comparison can be tested against the statistical hypothesis by using the gamma function and at the confidence level of, % (1α) (Eq. 10):
For example, for a group of size 6, whose judgments on a given issue are equal to {2, 3, 7, 9, 1, 2}, the geometric dispersion of the group is equal to 1.9052169. Taking the usual significance level of 5 percent, we observe that:
Thus, the pvalue corresponding to the sample geometric dispersion indicates
that it seems rare to observe values of the geometric dispersion smaller than
the sample geometric dispersion and hence, the geometric dispersion of the group
is not unusually large, which in turn implies that the geometric mean can be
used as the representative preference judgment for the entire group.
Table 1: 
Gamma distribution parameters 

RESULTS
The present study, statistical population consists of 2516 teachers who hold bachelor, or higher degrees. With regard to (Eq. 1) and the following values, the sample volume is:
The required calculations to determine the reliability of the questionnaire,
in which each pairwise comparison has been shown in Table 2
in order to facilitate the computer calculations and optimal use of the collected
data.
The reliability test of the questionnaire relating to the local weights of
the primary factor (lucid information) with the confidence level of 95% was
carried out and given as a sample as follows:
• 
Statistical hypothesis: 
• 
Spearman's correlation coefficient: 
Table 2: 
Pairwise comparison of alternatives over C_{1} 

• 
Critical value: Since, n>10; therefore, the distribution
is approximately normal and we will have: 
Table 3: 
Data relating to the local weights of the lucid information
criterion 

Table 4: 
Statistic test calculated for all the factors influencing
the organizational development 

• 
Decisionmaking: Since, z is calculated (z = 0.2.35),
it is located in the critical zone; therefore, the null hypothesis is rejected
(Table 3) 
Thus, it can be claimed that there is a significant correlation at the recognition
level (α = 0.0.5) between the two variables X and Y. Likewise, the statistic
test (z) has been calculated for all of the factors influencing the organizational
development according to Table 4.
Because all of the calculated statistics are placed in the critical zone, then the null hypothesis is rejected. Thus, it can be claimed there is a significant correlation at the recognition level of α = 0.0.5 between the first and second times.
In order to test the geometric mean values obtained from the group members' judgments, we apply the gamma test explained in the previous section to each of the comparisons made in the pair comparison table. For instance, Table 5 depicts the judgments of the group in the comparison of the two items, adaptation of the organizational space to individual needs over changing the cultural norms relating to the criterion information.
Table 5: 
Individual judgments of group in comparison of A_{1}
and A_{2 }over criterion C_{1} 

Table 6: 
Revised individual judgments of group in comparison of A_{1}
and A_{2 }over criterion C_{1} 

The parameters of gamma distribution function to n = 111 will be as follows
(Eq. 11):
The geometric mean value will be 1.0229, the geometric dispersion value will be 3.6738 and the value of p obtained testing the statistical hypothesis will be as follows (Eq. 12):
In which case the statistical hypothesis is rejected and the value of geometric mean cannot be used as group's judgment.
This has been done for other matrices as well and in the case of rejection, the matrices revised by the group members have been used (Table 6).
The geometric mean value will be 1.0272, the geometric dispersion value will be 2.4028 and the value of p obtained from testing the statistical hypothesis will be as follows:
And the statistical hypothesis will be accepted. This is done for all other
pairwise comparison matrices.
Table 7: 
Global weights of organizational change situations 

The results of testing the statistical hypothesis against each pairwise comparison
matrix value show that all the judgments were homogeneous and there is no dispersion.
The results of testing the statistical hypothesis against each pairwise comparison matrix value show that all the judgments were homogeneous and there is no dispersion. Thus, we proceed to calculate the global weights of the alternatives (Table 7).
CONCLUSIONS
The extreme closeness of the priorities determined by the group indicates the
absence of complete consensus on the issue by the group. Nevertheless, in the
studied organization, the teachers gave the highest priority to promotion of
the communicative system and the lowest priority to the enhancing the intergroup
cooperation. But the third and fourth priorities, which are related to changing
the cultural norms and changing the structure and composition of roles, are
not much different. Finally, the fifth priority is related to changing the motivation
of the workforce.