7/31/2019 The Impacts of Culture on System Development Methodology in Practice
1/21
1
Research Paper
The Im pacts of Culture on System D evelopm ent M ethodology in Practice
Author: Lena Salim (u4431957)
ABSTRACT
Information system methodologies are undoubtedly one of the most important topics in
information system field (Fitzgerald et al, 2002). M any researches had been conducted to
provide both business practitioners and academia with deeper understanding in this field;
however, this particular topic is still considered remains in unclear, less structured grey area.
This paper aims for contributing a clearer vision towards development methodologies by
inviting the readers to view this topic from the culture point of view. Arguments and wide
discussion will be covered in this paper to assist project practitioners to answer the question
how organization culture can affect the development practice.
1. Introduction
Culture is defined as the whole range of learned values that guide the patterns of
thinking, feeling and acting of people coming from that culture (based on Hofstede, 1997).
It assists people in categorizing and projecting their world by understanding the habits,
regulations, and expectations from others attitudes (based onOlson and Olson, 2003). M any
crucial decisions in business are m ade with a high regard on culture; often time,
organizations have to pay a great expense after a less careful consideration on culture.
Information system development methodology (ISD M ) is a project framework that
provides a clear step by step guideline for approaching a development situation (based on
Fitzgerald et al, 2002; Turbit, 2005). This project framework will be applied onto personnel
who are involved in a project throughout life cycle; therefore, it should be accepted as a partday to day pattern of thinking, feeling and behaviors in an organization. Or in other words,
regardless to which ISD M that is adopted by an organization, it should be well accepted or at
least does not contradicted with the organization culture.
M any of IS researches have been conducted into the effects of culture but only few
of them that particularly touched the area of development methodologies (based on M yers
and Tan, 2002). Some of these researches are either too theoretical or too specific that focus
only on some parts of the culture. A paper written by M oores and Gregory, for instance,
focused only on two small areas of culture: the avoidance of group debates and the use of
multiple languages (M yers and Tan, 2002). This paper aims for providing a wider perspective
of culture and developing arguments that will assists the readers to have a betterunderstanding on the how culture could effect the development methodologies in practice.
By the end of this paper, the author aims to deliver:
The fundamental theory of both culture and ISD M s that will be used as the basic for
developing the whole ideas covered in this paper.
Developing hypotheses that postulates how certain type of culture can affect the certain
pratices of ISD M .
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
2/21
2
To testify the developed hypotheses by using the global survey results pre-collected by
Cusumano et al in 2003.
To produce a guideline for assisting developers bridge the gap between organization
culture and ISDM s.
2. Background Theory
To reveal the relationship between culture and the implementation of ISDM in
practice, one must first have a sound understanding on the basic culture theories and the
variety of ISDM s exist in software industry. This section provides discussions on key cultural
dimensions and different types of ISDM s that are very crucial for the following sections. An
overview of paper by Cusum ano et al will also be provided here to sum marize the results
from the survey of global software development practice.
2.1.H oftedes cultural dim ension
Plenty of researches have been conducted by social scientists to answer the question
on how cultures differ among social groups (based on Olson and O lson, 2003). Hofstede de
Geert is arguably the most popular social scientists in this area (based on Gorlenko, 2006).
Indeed most of IS researches that analyzed the effect of national culture had been conducted
based on Hofstedes dimensions of culture (M yers and Tan, 2003). The five cultural
dimensions suggested by Geert Hofstede are sum marized as follows:
Low versus high power distance (PDI).
Power distance refers to what extent the less powerful members of organizations and
institutions (such as family) accept and expect that power is distributed unequally (Hofstede
and M C Crae, 2004). All social groups have som e inequalities attached to them, but som e are
more unequal compared to the others (Hofstede and M C Crae, 2004). Countries with higher
power distance tend to have more inequalities in their societies.
Individualism versus collectivism (IDV)
Individualism versus collectivism refers to what extent individuals are integrated into
groups (Hofstede and M C Crae, 2004). Individualists expect to look after themselves and
tend to value individual interests over the ones of a group (based on Hofstede, 1991).
Collectivists, on the other hand, are very dependent to their group and perceive that
allocating individual over group interests is evil (based on Hofstede, 2001).
M asculinity versus feminity (M AS)
M asculinity versus fem inist refers to the distribution of em otional roles between sexes
Hofstede and M C Crae, 2004). The m asculinity side is characterized with the more assertivebehavior whereas caring and modest are determined towards feminity (Hofstede and M C
Crae, 2004).
Low versus high uncertainty avoidance (UAI)
Uncertainty avoidance refers to what extent a society can deal with ambiguity (Hofstede and
M C Crae, 2004). Societies with low uncertainty avoidance are m ore tolerant to ambiguity and
accept it as a part of their lives (based on Hofstede, 2001). Societies with high uncertainty
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
3/21
3
avoidance on the other hand, feel very uncomfortable with ambiguity and tend to minimize
it (Hofstede and M C Crae, 2004).
Long term versus short term orientation (LTO )
Societies with long term orientation are characterized as being m ore preserving, ordering
relationships by status, thrift and having a sense of shame (Hofstede, 2001). On the other
pole, societies with short term orientation are perceived as societies with personal
steadiness, self-face protecting, respect for tradition, and reciprocation of greetings, favors
and gifts (Hofstede, 2001).
2.2.Form alized Inform ation System D evelopm ent M ethodologies (ISD M )
In their book, Fitzgerald et.al. (2002) use the term formalized methodologies to
refer to all development methodologies that are formally documented, including both
com mercial methods and in-house methods. Formalized m ethods provide guidelines for
developers to approach development situation in a methodical way (Fitzgerald et al, 2002).
Some examples for these formalized m ethodologies are waterfall, Extreme Programming
(XP), and M ultiview. There are plenty of formalized methodologies available in softwareindustry nowadays and most of them can be categorized into these two following
dimensions: flexibility and paradigm .
Figure 1 Categorization of Formalized M ethodologies
Flexibility : Rigid versus Agile
The rigid methodologies assum e that user requirements are static and are predictable at the
beginning of the development process. W aterfall SD LC, for example, consists of no sub
cycle and user requirements are therefore gathered once only at the beginning phase of thelife cycle. Rigid methodologies rely heavily on extensive documentations and usually involve
a large num ber of project team members (based on Chavarat, 2003b). Agile methodologies,
on the other side, are particularly designed to adapt to the changing environment and
emphasize heavily on the importance of team work (Abrahamson et al, 2002). Agile
methodologies like XP and scrum consist of more than one sub cycles; therefore,
requirements are more flexible and being refined throughout the development process.
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
4/21
4
Good relationship with customer is more preferable in agile m ethods, as it is com pared to
the strict agreements written on the contract (Abrahamson et al, 2002).
Paradigm : Scientific versus System atic (holistic)
The science paradigm approaches problems by reducing the com plexity in the real world and
discover explanation through validated and repeatable exam ination (based on Checkland in
Avison and Fitzgerald, 2006). The structured approach, for instance, divides the complex
requirements into smaller parts and assumes the break downs would not disrupt the system
as a whole (based on Checkland in Avison and Fitzgerald, 2006). The systematic paradigm ,
on the other hand, approach problems by requiring software developers to look beyond the
obvious boundaries of the system and to allocate more attention to the systems external
factors such as customers, competitors and governments (based on Avison and Fitzgerald,
2006). System atic approaches also acknowledge the possibility of more than one alternative
in problem solving and suggest the formation of development team that involves members
from variety discipline areas (on Avison and Fitzgerald, 2006). Exam ple for the systematic
ISDM s is SSM (Soft System M ethods).
2.3. Software D evelopm ent W orldwide: The state of the Practice
There is one survey in particular that is going to be used for testifying the hypotheses
developed in this research. During 2001 and 2002, Cusum ano et. al. conducted a global
survey of software development (Cusum ano et al., 2001). The sample com prised of 104
projects with 27 participating companies, originated from India, Japan, USA and W estern
Europe (Cusum ano et al., 2001). The result of this survey is summarized here in table1. The
key factors listed in the first column of the table indicate the factors being observed in the
survey and the num bers in the other colum ns show the percentage of projects in each
country that actually implem ent these key factors in practice.
Table 1. Software Development W orldwide (Source: Cusum ano et al, 2003)
The survey respondents were questioned whether they followed a m ore conventional
approach such as waterfall or a more flexible one that involves multiple mile stones or sub
cycles (based on Cusum ano et al., 2001). In their paper, Cusum ano et. al. defines firms that
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
5/21
5
followed a m ore conventional waterfall approach as those that did not use sub cycles in
system development practice. The respondents were also questioned whether they wrote
specifications and produced detailed design before they commenced the coding phase (based
on Cusumano et al., 2001). Finally, respondents were asked whether they used pair testing as
a technique to review their codes prior to submission (based on Cusumano et al., 2001).
3. Research M odel and H ypothesis
This research is conducted based on the idea that culture m ay play an influential role
in software development practice. In their book, Fitzgerald et al (2002) argue that
development context is the foundation of system development and therefore deserves more
attention and detailed understanding from development practitioners. Development context
here is described as the business context in which software development takes place,
including technology, culture, change strategies and other philosophical, practical and
method considerations (Fitzgerald et al, 2002).
Figure 2 The D evelopment Context (Fitzgerald et al, 2002)
As it is painted in Figure 2, culture is one element of development context that
shapes up the implementation of development methods in action. The word culturehere
mainly refers to the organization culture that is shared and reflected in the procedures andprocess within the organization (Fitzgerald et al, 2002). An organization culture will
determines what is possible or not to be conducted in the organization and therefore will
serve as either catalyzing or restraining factors for the implem entation of certain software
development practices (based on Fitzgerald et al, 2002). For example, in som e organization
cultures there is a strict em phasize on the importance of administration and bureaucracy.
This characteristic will eventually influence critical decisions such as which ISD M s to be
selected for an organization or how well will they adopt a particular development practice.
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
6/21
6
This paper will use the development framework painted in Figure 2 (particularly for
the area marked with red circle) as its research model and will draw some hypothesis by
using the five Hofstedes cultural dimensions. All these hypotheses are to be tested in section
4 by using the survey result from Cusum ano et al. M ost data that is currently available from
other researches conducted in this field are m ainly focus on national culture instead of
organizational culture; therefore, this paper will firstly analyze the hypotheses from the
national cultures perspective before projecting its result from organizational cultures point
of view. The developers model crafted in section 5 will provide a clearer framework on how
organizational culture would effect the implementation of development methods in action.
3.1.The adoption of form alized ISD M .
Despite the popularities of the formalized methodologies in software industry
nowadays, a research by Fitzgerald shows that there was no organization that had
implemented them rigorously (based on Fitzgerald, 1997). The diversity of development
context, along with the given time constraints had forced system developers to deliberatelyomit some part of development methodologies and tailored them to their development
situation (based on Fitzgerald, 1997). An exam ple for these omitted parts was the system
documentation (Fitzgerald, 1997).Indian software com panies are arguably one of the most
rigid adopters of formalized methodologies in the world. This is supported by the fact that
more than half of the CM M level 5 software firms in the world are located in the country
(M oitra, 2001; Arora et al, 2001). In 2001, there were m ore than 175 Indian software
companies that had acquired ISO 9001 certification (Arora et al, 2001). In order to be
eligible for either of these ISO or CM M certifications, a software firm must first adopt a
formalized system development methodology (based on Arora and Asundi, 1999).
One dimension of culture that presumably will affect the adoption of formalizedmethod is the power distance index (PDI). According to Hofstede, societies with high power
distance are characterized with the centralized decision-making style and the existence of
status inequalities among employees as the result from the organizational hierarchy
(Hofstede, 2001). In these societies, subordinates are more dependent on their supervisors
and expect to be told what to do (Hofstede, 2001). These characteristics lead to the
argument that societies with high power distance will have tendency to adopt the formalized
development methodology more rigorously than the others. The implementation of
formalized methodologies is usually imposed by people from upper management (Hiusman
and Livari, 2002). Therefore, managem ent from high power distance societies will be more
successful in convincing their developers to adopt formalized methodologies. This argument
can be summarized in a hypothesis as follows:
H1: The higher power distance score will have a positive impact on the
implementation of formalized methodology.
There is another cultural dimension that seems to bring impact on the adoption of
formalized m ethod. According to Hofstede in Narayanaswamy and Henry (2005), societies
with high uncertainty avoidance prefer sets of regulations in order to deal with uncertainties
and to provide m ore feelings of predictability. The implementation of formal control in
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
7/21
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
8/21
8
The last cultural dimension that may effect on the selection of development
methodology is the masculinity level. M asculinity determines to what extent a society values
their ego and material belongings over their relationship with others (based on Hofstede,
2001). Societies with high masculinity level are m ore competitive and are characterized with
mass production (Hofstede, 1997). Societies with low masculinity level, on the other hand,
show empathy to others and are usually characterized with custom made products
(Hofstede, 1997). The agile software development methodologies stress the importance of
relationship between developers and clients higher than the strict contracts and technology
(Abrahamsson et al, 2002; M iller, 2001 in Abrahamsson et al, 2002). Agile methodologies are
tailored to changes and corrections; therefore, they are more receptive towards changing in
user requirements (based on Favaro, 2002 in Abraham sson et al, 2002). These
methodologies are therefore more suitable for societies with low masculinity level which
values relationships over strict contracts and are more willing to tailor their products based
on customer requirements. This argument leads to the following hypothesis:
H5: The higher masculinity score will have a negative influence on the adoption of
agile methodologies.
3.3.Am ount of effort spent for detailed design
Not all developers have the same level of preferences when it comes to producing
detailed designs. Having a more detailed design -both the functional and the architectural
ones- instead of jumping directly from specifications to coding will lead to fewer software
defects (based on Cusum ano et al, 2003). This is the best practice, however, according to
Cusum ano et al was not always been implemented in the real world.
There is one dimension of the Hofstedes national culture that could possibly affect
the developers decisions on the am ount of effort spent for producing detailed designs. Thelong term orientation dimension, according to M ooij in Ryu and Eyugbolu (2003) refers to
what extent that a country embraces to the future- oriented perspectives as it is opposed to
the traditional short- term point of view. Societies with long term orientation put more
focus on the future state instead of current state in decision makings. This statem ent leads to
the argument that societies with higher long-term orientation are the m ost likely ones that
will go through detailed designs before jumping to the coding stage. Since detailed designs
oblige developers to describe functions and operations in details, discrepancies between
requirements and the system to be developed are more likely to be discovered soon enough
at the early stage of project life cycle. This will result in fewer amounts of changes required
during the development process and decrease the amount of the overall project cost.
Developers with long term perspectives are assumed to be more able to see how detailed
design will impact to the future state of the project and this can serve as their motivation for
allocating m ore effort in producing complete designs. For more details on the relationship
between cost and changes in software development process, please refer to Figure 2.
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
9/21
9
Figure 3 Cost of Change Curve: Traditional and Agile (Source: Ambler, 2002)
The following hypothesis is built based on the argument developed in this section:
H6: The higher long term orientation score will have a positive impact on the
amount of effort spent for detailed design.
3.4.The use of pair testing
Pair testing is one popular technique used for program testing in software projects.
Testers in this case work collaboratively as a pair, sometimes split up- working side by side at
two computers and then rejoin once they find defects in order to find the best solution(Ratzmann and Young, 2003). There is one cultural dimension suggested by Hofstede that is
closely related to the adoption of pair testing in system development. The individualism
index, according to Hofstede (2001), measures the extent to how people in a country look
after themselves and those of the social groups that they been part of. Societies with low
individualism index tend to expect opinions to be predetermined by in-group, contrary to
those who expect opinions to be coming personals (Hofstede, 2001). This way of thinking is
highly required in the implementation of pair testing, more especially when both testers
developers and tester have conflicting perspectives in defining a defect or simply finding the
best way to solve defects. Societies with low individualism disvalue the practice of placing
individual over group interests; therefore, it is more likely for them to put aside their
personal opinions and resolve conflicts based on consensus (based on Hofstede, 2001).
Hence, it is assumed that societies with low individualism will adopt pair testing more widely
than those are not.
H7: The higher individualism score will have a negative impact on the use of pair
testing.
Cost of Change: Traditional Cost of Change: Agile
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
10/21
10
3.5.The adoption of system atic (holistic) m ethods
The system atic (holistic) methods require software developers to have a system
perspective over the system to be developed. Since there is no such thing as systematic
dimension mentioned in the Hofstedes theory, this paper will draw out two other popular
cultural dimensions that presum ably will affect the adoption ofsystem perspective. These two
dimensions are: individualism and uncertainty avoidance.
The systematic perspective views an organization / system as a whole and
discourages the practice of emphasizing the detail of one particular domain or perceiving the
domain from one single point of view (Avison and Fitzgerald, 2006). From the cultural point
of view, individualistic societies are characterized with the tendency to look after them selves
or their immediate relatives rather than the society where they belong to (Hofstede, 1997).
The individual interests in this case sit on a higher priority as they are opposed to collective
interests (Hofstede, 1997). Such characteristics will restrain developers to allocate the
organization interests above the ones of their departments -or perhaps the ones of their
own. The system atic (holistic) approach is therefore considered less suitable for societies
with high individualism.
H8: The higher individualism score will have a negative impact on the adoption of
system atic (holistic) methods.
The other cultural dimension that presumably will bring a strong impact to the
adoption of systematic approach is the uncertainty avoidance. According to Hofstede (2001),
societies with high uncertainty avoidance prefer well-structured regulations and believe that
there is only one truth exists in this world: the one that they currently believe in. These
characteristics will restrain developers from acknowledging the existence of more than one
alternative in solving a problem , which is one of the most crucial success factors in
systematic perspective. Systematic methods encourage the formation of development teamwith multidisciplinary members; team members with high uncertainty avoidance in this case
will possibly fail to accept opinions from the others and will eventually cost a great expense
to the project success. This argument is framed in the following hypothesis:
H9: The higher uncertainty avoidance score will have a negative impact on the
adoption of systematic (holistic) methods.
4. Analysis
After the long discussion provided in the previous section, this paper will testify the
developed hypotheses by using data collected by Cusum ano et al sum marized in table 1.
There are nine hypotheses drawn in total but not all of them will be evaluated here due to
data insufficiency. There is no adequate information given in table 1 that could be used for
testifying the last two hypotheses; therefore, these hypotheses are not to be covered in this
section. Hypotheses accepted in this section will be documented in section 5 for
constructing a m odel dedicated to software practitioners.
4.1.Analysis: The adoption of form alized system developm ent m ethodologies
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
11/21
11
The adoption of formalized methodologies in organizations is usually imposed by
people from upper management (Hiusman and Livari, 2002). However in som e cases,
software developers decided to ignore such methodologies and implement the ones of their
own. There are five key factors used in this section for measuring to what extent software
projects have adopted the formalized methods: the amount of projects that produce
architectural specifications, functional specifications, code generation, design reviews and
code reviews. All these mentioned factors are also the basic parts of most formalized
methodologies; therefore, omission of one key factor may be perceived as the indicator of
methodology ignorance. The data allocated for each key factor here is gathered and averaged
from the survey results listed in table 1. Followings are the hypotheses to be testified in this
section:
H1: The higher power distance score will have a positive impact on the
implementation of formalized methodology.
H2: The higher uncertainty avoidance score will have a positive impact on the
implementation of formalized methodology.
Country Amount of projects
that perform
specifications, code
generation and
reviews %
PDI
*H1
IDV MAS UAI
*H2
LTO
India 87.48 77 48 56 40 61
Japan 75.56 54 46 95 92 80
Western Europe 73.62 42 64 45 67 31
USA 65.8 40 91 62 46 29
Conclusion SR MR MR
Table 2. The adoption of formalized system development methodologies vs. Hofstedes scores
PDI= Power Distance Index
IDV= Individualism Index
M AS=M asculinity SR=Strong Relationship
UAI= Uncertainty Avoidance Index M R=M oderate Relationship
LTO =Long term Orientation
The data figure in table 2 shows a strong relationship between the adoption of
formalized method and PDI. The PDI score listed in the table decreases along with the
declination of the % projects that implement formalized methods. India -the country withthe highest power distance score- is also ranked as the first when it comes to the
implementation of formalized ISD M s. On the other hand, USA -the country with the lowest
power distance score- ranked at the last in terms of the implementation of formalized
ISDM s. This data pattern indicates that there is a strong relationship between the two factors
being observed; which also m eans that H1 is supported in this analysis.
H1: The higher power distance score will have a positive impact on the
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
12/21
12
implementation of formalized methodology.
The second hypothesis to be evaluated in this section, H2, argues that there is a
positive relationship between the adoption of formalized ISD M s and UAI. This argument is
not supported by the data shown in table 2. Logically speaking, countries with high
uncertainty avoidance will prefer formalized methodologies as a tool to deal with project
uncertainties and to provide more feelings of predictability. One thing that seems to be
missing here is the fact that there are many types of formalized methodologies available in
the world, varying from the ones with low uncertainty level such as waterfall SD LC to the
ones with high uncertainly such as agile methods. Therefore, there is still a chance that
societies with high uncertainty avoidance will refuse to adopt formalized methods if they are
coming from the agile ones. Since the analysis results in this case do not appear to support
the hypothesis, H2 should therefore be rejected.
H2: The higher uncertainty avoidance score will have a positive impact on
the implementation of formalized methodology.
There are two new relationships revealed during this analysis that are not yet beinghypothesized in the previous section. The analysis results show moderate relationships
between the other two cultural dimensions IDV and LTO- and the adoption of formalized
methodology. The termmoderate relationshipis used here because there is one cultural score in
particular that deviates from the data pattern shown in table 2 (e.g. the IDV score or LTO
score for Japan). Omission of this score will result in a perfect relationship between both
factors being observed.
The first newly discovered relationship in this section is the relationship between
IDV and the adoption of formalized ISDM s. In his paper, Turbit (2005) suggested som e
restraining factors that might prevent software developers from adopting the formalized
methodologies. These restraining factors are: the formalized methodologies are beingperceived as the enemy of creativity and software developers in this case have their own
methodologies that are considered will work better. Societies with high individualism are
characterized with individual freedom and the importance of private opinions (Hofstede,
1997). Individual creativity / opinions are highly appreciated in these societies; therefore,
software developers in this case will have the tendencies to come out with their own
methodologies that they consider will work better or enhance their creativities. This
relationship leads to the following conclusion.
C10: The higher individualism score will have a negative impact on the
implementation of formalized methodology.
The second relationship that has just been revealed in this section is the relationship
between LTO and the adoption of formalized ISD M s. Based on a research conducted by
Chatzoglou in 1997, software projects that implement no formalized methodology will endup with higher cost, time and effort spent for the entire projects. The adoption of formalized
methodologies, on the other hand, will result in the increase in productivity by 10% -30%
(Turbit, 2005). Societies with high LTO are characterized with the tendency to focus on the
future status when it comes to decision making. Therefore, while deciding whether to adopt
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
13/21
13
a formalized methodology or not, software developers from LTO societies will automatically
predict the effect of their decisions on the future results of their projects. By holding upon
the research results from Chatzglou (1997), these software developers will therefore prefer to
adopt the formalized methodologies for their project.
C11: The higher long term orientation score will have a positive impact on
the implementation of formalized methodology.
4.2.Analysis: M ethodology preference (rigid vs agile m ethods)
There are three cultural dimensions hypothesized in section 3 to bring impact on the
methodology preferences of a society. These cultural dimensions, along with the nature of
their impacts are listed as follows:
H3: The higher power distance score will have a negative influence on the adoption
of agile m ethodologies.
H4: The higher uncertainly avoidance score will have a negative influence on theadoption of agile methodologies.
H5: The higher masculinity score will have a negative influence on the adoption of
agile methodologies.
Societies that less prefer agile methodologies will have the tendency adopt the m ore
rigid ones, and vice versa. The measurement used for indicating the adoption of agile
approaches in this section is taken from the percentage of sub cycles usage listed in table 1.
Software projects that adopt newer / more flexible m ethodologies are the ones that come
out with more iterative and incremental sub cycles (based on Cusumano et al, 2003; M iller,
2001 in A brahamsson et al, 2002).
Country Amount of projects that usesub cycles (%)
PDI
*H3
IDV MAS
*H5
UAI
*H4
LTO
Western Europe 86.2 42 64 45 67 31
India 79.2 77 48 56 40 61
USA 54.8 40 91 62 46 29
Japan 44.4 54 46 95 92 80
Conclusion SR
Table 3. M ethodology preference vs. Hofstedes Scores
PDI= Power Distance Index
IDV= Individualism Index
M AS=M asculinity SR=Strong Relationship
UAI= Uncertainty Avoidance Index
LTO =Long term Orientation
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
14/21
14
The data figure in table 4 shows that there is no relationship between the adoption of
agile methodologies and PDI. It was previously argued in H3 that societies with high PDI
will restrain software developers to express their opinions freely in front of their managers,
which in turn will prevent them to benefit from agile methods. One point that seems to be
missing here is: How if it is the management itself who insists on the adoption of agile
methods? It is believed that the implementation of methodologies is usually imposed by
people from upper management (based on Hiusman and Livari, 2002). Since high PDI
societies are characterized with obedient sub ordinates; this situation will result with two
conflicting behaviors created by the high PDI score: the unwillingness to express their
opinions and the obedience to adopt methodologies imposed. The success of methodologies
in this case will possibly be dependent on another external factor such as the management
support. In other words, the m anagem ents ability in lubricating the adoption of agile
methods can also make a significant difference here. For example, project managers from
high PDI may consider using one popular technique in m anagem ent such as devils advocate
in order to stimulate reactions from their subordinates. However, since the argument made
in H3 is not supported in table 3, this hypothesis should therefore be rejected.
H3: The higher power distance score will have a negative influence on theadoption of agile methodologies.
The second hypothesis to be testified in this section, H4, argues that there is a
negative correlation between UAI and the societies preference over agile methodologies.
This argument, however, is not supported in the analysis results. Logically speaking, societies
with high UAI will refuse to adopt agile methodologies with less certain work models,
significant amount of changes and risks attached to them. The information gathered and
data used in this research are still not sufficient enough to come out with explanations for
such analysis results. The author believes that a deeper research in the future will help
revealing the answer for this situation.
H4: The higher uncertainly avoidance score will have a negative influence
on the adoption of agile m ethodologies.
The last hypothesis to be testified in this section, H5, is supported by the data figure
shown in table 3. The analysis results come out with astrong relationshipfor describing the
correlation between M AS and the societies preference over the agile m ethods. Therefore,
H5 in this case is supported and should be accepted.
H5: The higher masculinity score will have a negative influence on the
adoption of agile methodologies.
4.3.Analysis: Am ount of effort spent for detailed design
Logically speaking, the m ore detail design crafted before the code generation will
lead to the fewer defects and less reworks throughout the project. There is one H ofstedes
cultural dimension that is postulated to have a relationship with the amount of effort spent
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
15/21
15
for detailed design: the long term orientation. Following is the hypothesis set for this
relationship:
H6: The higher long term orientation score will have a positive impact on the
amount of effort spent for detailed design.
Country Am ount of projects
that produce detailed
design %
PDI IDV MAS UAI LTO
H6
India 100 77 48 56 40 61
Japan 85.2 54 46 95 92 80
Western Europe 68.2 42 64 45 67 31
USA 32.3 40 91 62 46 29
Conclusion SR MR MR
Table 5. Detailed design versus vs. Hofstedes Scores
PDI= Power Distance Index
IDV= Individualism Index
M AS=M asculinity SR=Strong Relationship
UAI= Uncertainty Avoidance Index M R=M oderate Relationship
LTO=Long term Orientation
The data in table 5 shows a moderate correlation between the amount of projects
that produce detailed design and LTO. This correlation, however, is not the only relationship
shown in the analysis results. In addition to what has been postulated in H6, there are two
other unexplained relationships projected in table 5: the relationships between PDI and IDV
and the amount of projects that went through detailed design. There are two possible
scenarios that would likely be the explanation for this situation: either the information
gathered by this research is not yet sufficient for describing these new revealed relationships
or there is another unconsidered external factor that causes these results to take place. It is
strongly believed that the latter scenario is the m ost likely one in this research. There are
plenty of development methodologies available in the current software industry and different
methodologies may require different level of detail in design (based on S Fraser 2008, pers.
comm., 31 Oct). W aterfall SD LC, for example, requires software developers to produce a
more detailed design before coding compared to other methodologies such as prototyping.
The amount of effort spent for detailed design in this case is more related to the type of
methodology adopted rather than other factors such as LTO score. A deeper research in the
future is required in this case to discover a strong supported explanation for results shown intable 5.
H6: The higher long term orientation score will have a positive impact on
the amount of effort spent for detailed design.
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
16/21
16
4.4.Analysis: The use of pair testing
There is one H ofstedes cultural dimension that is postulated to contribute negatively
to use of pair testing: the individualism level of a society. Following is the hypothesis taken
from the previous section that is soon to be evaluated:
H7: The higher individualism score will have a negative impact on the use of pair
testing.
Country Amount of project thatuse pair testing %
PDI IDV
H7MAS UAI LTO
India 54.4 77 48 56 40 61
Japan 44.4 54 46 95 92 80
USA 35.5 40 91 62 46 29
Western Europe 31.8 42 64 45 67 31
Conclusion SR
Table 6. Pair Testing versus vs. Hofstedes scores.
PDI= Power Distance Index
IDV= Individualism Index
M AS=M asculinity SR=Strong Relationship
UAI= Uncertainty Avoidance Index M R=M oderate Relationship
LTO=Long term Orientation
The analysis results in table 6 show that there is no relationship between the adoption
of pair testing and the individualism score rated for each country. The PDI score, on the
other hand, is projected to have a strong relationship with the adoption of pair testing in
practice. Similar to the previous case, the author strongly believes that there is anotherexternal factor that may have contributed significantly here to these analysis results. Pair
testing is one technique that is recommended for agile development practices; therefore, the
adoption of pair testing in this case is very dependent on the type of methodologies chosen
for software projects (based on Nawaz and M alik, 2008). Since the data used in table 6 failed
to consider this important factor, results shown by the analysis should not be accepted for
the next section. A deeper research in the future is still required in this case to discover
explanation for results shown in table 6.
H7: The higher individualism score will have a negative impact on the use
of pair testing.
5. A fram ework for software practitioners
After a long discussion and analysis provided in the previous sections, this paper will
finally come out with a framework dedicated to software practitioners. There were nine
hypotheses drawn during this research but only seven of them were evaluated in section 4.
The results came out with two accepted hypotheses; by adding to the two other relationships
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
17/21
17
discovered during the analysis, there are four relationships being accepted in total. The
following framework sum marizes all these accepted hypotheses.
Figure 4. A Framework for Software Practitioners
This developed framework will assist project personnel in mitigating the potential
issues arise from the adoption of development methodologies. Implementation of ISD M
that is not suitable to culture of a society will result in user resistance and the excessive time
/ effort spent for educating m ethodology users. It is also important to note that software
practitioners in this case should refer to organizational culture instead of national culture.
The Hofstedes national culture scores provide a strong basic for evaluating the developed
hypotheses but when it comes to practice, one is strongly encouraged to recognize theirown
organizational culture. By using the five cultural dimensions from Hofstede for predicting
the culture of an organization, one may then use the framework painted in figure 4 fordecision makings. The national culture theory has long been criticized for its assum ption of
cultural homogeneity and stability (based on W alsham , 2004). If software practitioners insist
on referring to the national culture score from Hofstede, there is a chance that they will
neglect the m ulticultural issues that exists within a country and discount the dynamic nature
of culture.
6. Constraints and assum ptions
Just like other research papers written on the similar topic, this research is conducted
based on som e constraints and assumptions. These constraints and assumptions are stated asfollows:
There is a significant time gap exists between the two sets of data being analyzed in this
paper. The cultural score from Hofstede is collected in 1980s, which is about 23 years
before the survey results from Cusumano et. al.
Due to data insufficiency, there are only four countries being analyzed in this paper.
M ore data from other countries would have been very helpful for concluding a stronger
analysis result.
High Power distance
Low Individualism,
High Long term Orientation
Formalized
M ethodologies
Non Formalized
M ethodologies
High masculinity?
Rigid
M ethodologies
Agile
methodologies
Yes
Yes
No
No
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
18/21
18
There is no clear information provided in Cusum ano et al regarding to which countries
in W estern Europe that were being analyzed. The culture score for W estern Europe used
in this paper is calculated from the average from all countries in Hofstedes table that
belong to W estern Europe.
This paper discounts the other factors that may also bring impact to system development
practice. Factors such as the size of a project and project complexity may have playedinfluential roles in the survey results used for testifying the hypotheses.
7. Im plication for further research
From nine hypotheses drawn in this paper, two are left unevaluated and three
remains unsolved and are suggested for further research. The m ain issue faced by this
research is the lack of data available for testifying the developed hypotheses. In his book,
Geert Hofstede comes out with culture scores allocated for 74 countries around the world;
however, there is no similar amount of information made available in the current academic
literature that specifies the world practice of development methodologies. Such information
will not only be useful for testifying the hypotheses developed, but will also be a significant
contribution to all related researches conducted in the future.
8. Conclusion
Although there are some questions that are still left unsolved in this paper, the
conducted research has shown a significant role played culture in software development
practice. Certain type / dimension of culture will bring impacts to how well an organization
will adopt a certain practice of development methodology; some organization may accept a
development practice much better than the others. Project practitioners are recommended to
pay a high regard on organization culture before m aking any decisions regarding to
development methodologies. There are som e hypotheses that are left un-testified by the end
of this paper; a more complete of information on global system development practice will bevery helpful for research in the future.
7/31/2019 The Impacts of Culture on System Development Methodology in Practice
19/21
19
Reference
i) Ambler, S.W . (2002). Agile M odeling. http://www.agilemodeling.com /essays
/costOfChange.htm. Date accessed 13 September 2008.
ii) Abraham sson, P., Salo, O., Ronkainen, J. and Juhani W arsta (2002). Agile Software
Development M ethods Review ad Analysis. VTT. Finland.
http://www.vtt.fi/inf/pdf/publications/2002/P478.pdf.Date accessed: 4 September
2008.
iii) Arora, A. and Jai Asundi (1999). Quality Certification and the Economics of Contract
Software Development: A Study of the Indian Software Industry.
http://www.inomics.com/cgi/repec?handle=RePEc:nbr:nberwo:7260Date accessed:
4 September 2008.
iv) Arora, A., Arunachalam, V.S., Asundi, J. and Ronald Fernandes (2001). The Indian
Software Services Industry.Research Policy. 30, 8, pp. 1267-1287.
v) Avison, D. and Fitzgerald, G. (2006). Information Systems Development
M ethodologies, techniques and tools. 4th Ed. M c G raw Hill. Berkshire.
vi) Charvat, J., (2003a), Chapter 1: Understanding project methodologies, ProjectManagement Methodologies, Wiley: New Jersey, pp.1-29
vii) Charvat, J., (2003b). Chapter 4: Development M ethodology selection and utilization,
Project M anagement Methodologies, W iley: New Jersey, pp.102-103.
viii) Chatzoglou, P.D., (1997). Use of methodologies: An empirical analysis of their
impact on the economics of the development process. European Journal ofInformation Systems. 6. pp.256-270
ix) Cusumano, M ., M acCormack, A., Kemerer, C.F. and Bill Crandall (2003). Software
Development W orldwide: The State of the Practice. IEEE Software. 20. 6. pp. 28-34.
x) Fitzgerald, B. (1997). The Use of Systems Development M ethodologies in Practice: A
Field Study.Information Systems Journal, 7, 3, pp. 201-212.
xi) Fitzgerald, B., Russo, N.L. and Erik Stolterman (2002). Information System
Development: M ethods in Action. M c Graw Hill. Berkshire.
xii) Gorlenko, L. (2006). The Moment of Truth: How Much Does Culture Matter to
You?Interactions. 13, 2, pp. 29-31.
xiii) Hiusman, M., and Iivari, J., (2002), The organisational deployment of systemsdevelopment methodologies, in M. Kirikova, W. Wojjtkowski, S. Wrycza, and J.
Zupancic, (Eds), Information Systems Development: Advances in Methodologies,
http://www.agilemodeling.com/essays%20/costOfChange.htmhttp://www.agilemodeling.com/essays%20/costOfChange.htmhttp://www.vtt.fi/inf/pdf/publications/2002/P478.pdfhttp://www.vtt.fi/inf/pdf/publications/2002/P478.pdfhttp://www.inomics.com/cgi/repec?handle=RePEc:nbr:nberwo:7260http://www.inomics.com/cgi/repec?handle=RePEc:nbr:nberwo:7260http://www.inomics.com/cgi/repec?handle=RePEc:nbr:nberwo:7260http://www.vtt.fi/inf/pdf/publications/2002/P478.pdfhttp://www.agilemodeling.com/essays%20/costOfChange.htmhttp://www.agilemodeling.com/essays%20/costOfChange.htm7/31/2019 The Impacts of Culture on System Development Methodology in Practice
20/21
20
Components, and Management, KluwerAcademic/Plenum Publishers: New York,
pp.87-100
xiv) Hofstede, G. (1997). Cultures and Organizations Software of the M ind. M c Graw Hill.
Berkshire.
xv) Hofstede, G. (2001). Culture Consequences: Comparing Values, Behaviors,
Institutions and O rganizations Across Nations. 2ndEd. Sage Publications. California
xvi) M acgregor, E., Hsieh, Y. and Kruchten, P. (2005). Patterns in Software Process
M ishaps:Incidents in Global Projects.ACM SIGSOFT Software Engineering N otes.30(4).
xvii) M oitra, D. (2001). Indias Software Industry.IEEE Software. 18, 1, pp. 77-80
xviii)Myers, M. D. and Tan, F. B. 2003. Beyond models of national culture in
information systems research. In Advanced Topics in Global information
Management. IGI Publishing, Hershey, PA. http://www.igi-
pub.com/downloads/excerpts/Szewczak Excerpt.pdf Date accessed 13 September2008.
xix) Narayanaswamy, R. and Raymond M . Henry (2005). Effects of culture on control
mechanisms in offshore outsourced IT projects.Proceedings of the 2005 ACM SIGM IS
CPR conference on Computer personnel research. pp. 139-145.
xx) Nawaz, A. and Kashif M asood M alik (2008). Software Testing Process in Agile
Developm ent. M aster Thesis. http://www.bth.se/fou/cuppsats.nsf/all/249945527
e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_D evelopment.p
dfDate accessed: 2 N ovember 2008.
xxi) Olson, J.S. and G ary M . Olson. (2003). Culture Surprises in Remote Software
Developm ent Teams.Queue. 1, 9, pp. 52-59.
xxii) Ratzmann, M . and Clinton D e Young (2003). Software Testing and Internalization.
Galileo Press GmbH . http://wise.vub.ac.be/Members/mushtaha/PhD/phd-
activity/GalileoSoftwareTesting.pdfDate accessed 13 September 2008.
xxiii)Ryu, Sungmin and N ermin Eyuboglu (2003). The Influence of Long-Term Orientation
Culture on Channel Relationships: The Korean Case.The H awaii International Conference
on Business, Honolulu.
http://www.hicbusiness.org/biz2003proceedings/Sungmin% 20Ryu. pdf Dateaccessed 13 September 2008.
xxiv)Turbit, N. (2005),Project M anagement & Software Development M ethodology, The Project
Perfect W hite Paper collection.
http://www.igi-pub.com/downloads/excerpts/Szewczak%20Excerpt.pdfhttp://www.igi-pub.com/downloads/excerpts/Szewczak%20Excerpt.pdfhttp://www.bth.se/fou/cuppsats.nsf/all/249945527%20e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_Development.pdfhttp://www.bth.se/fou/cuppsats.nsf/all/249945527%20e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_Development.pdfhttp://www.bth.se/fou/cuppsats.nsf/all/249945527%20e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_Development.pdfhttp://wise.vub.ac.be/Members/mushtaha/PhD/phd-activity/GalileoSoftwareTesting.pdfhttp://wise.vub.ac.be/Members/mushtaha/PhD/phd-activity/GalileoSoftwareTesting.pdfhttp://wise.vub.ac.be/Members/mushtaha/PhD/phd-activity/GalileoSoftwareTesting.pdfhttp://www.hicbusiness.org/biz2003proceedings/Sungmin%20Ryu.%20pdfhttp://www.hicbusiness.org/biz2003proceedings/Sungmin%20Ryu.%20pdfhttp://www.hicbusiness.org/biz2003proceedings/Sungmin%20Ryu.%20pdfhttp://wise.vub.ac.be/Members/mushtaha/PhD/phd-activity/GalileoSoftwareTesting.pdfhttp://wise.vub.ac.be/Members/mushtaha/PhD/phd-activity/GalileoSoftwareTesting.pdfhttp://www.bth.se/fou/cuppsats.nsf/all/249945527%20e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_Development.pdfhttp://www.bth.se/fou/cuppsats.nsf/all/249945527%20e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_Development.pdfhttp://www.bth.se/fou/cuppsats.nsf/all/249945527%20e869a47c125746c0002f4e1/$file/Software_Testing_Process_in_Agile_Development.pdfhttp://www.igi-pub.com/downloads/excerpts/Szewczak%20Excerpt.pdfhttp://www.igi-pub.com/downloads/excerpts/Szewczak%20Excerpt.pdf7/31/2019 The Impacts of Culture on System Development Methodology in Practice
21/21
xxv) W alsham, G. (2004). Cross-Cultural Issues in Global Software Outsourcing.
http://www.almaden.ibm .com /institute/pdf/2004/Geoff_W alsham_1.pdf Date
accessed: 2 O ctober 2008.
http://www.almaden.ibm.com/institute/pdf/2004/Geoff_Walsham_1.pdfhttp://www.almaden.ibm.com/institute/pdf/2004/Geoff_Walsham_1.pdfhttp://www.almaden.ibm.com/institute/pdf/2004/Geoff_Walsham_1.pdf