Post on 08-Nov-2015
description
transcript
1National and Organisational Culture:
Their Use in Information Systems Design
Dr Ernest JordanDept. of Information SystemsCity University of Hong Kong
Tat Chee AvenueKowloon, Hong Kong
(currently at Macquarie Graduate School of Management)
Email: EJordan@work.gsm.mq.edu.au
Abstract
Traditionally the information systems (IS) designer has followed simple concepts of
organisational functioning based overwhelmingly on ideas of management control.
Data from operational activity will be processed to become information for
management to use in planning, decision making and supervision. While this
viewpoint fits some organisations it is not appropriate to all. Organisational culture
and national culture, may be readily overlooked by the IS designer. In this paper,
the major dimensions of the impact of culture on an organisation's information
systems are identified and illustrated by examples. The significance to the IS
designer is pointed out.
21. CULTURE AND INFORMATION
Inside working organisations, as in all areas of human activity, the behaviour of
people is affected by the values and attitudes that they hold. The collective patterns
of behaviour are important parts of the culture of the work-group or nation, which
form a backdrop against which values and attitudes are in turn developed. This
cycle is expressed succinctly in Fig. 1, taken from Adler (1986).
Fig. 1 The Influence of Culture on Behaviour (Adler, 1986:9)
CULTURE
VALUES
ATTITUDES
BEHAVIOUR
Data only becomes information when it is interpreted by a person, and this
interpretation of necessity takes place against the backdrop of the individual's
culture (Tricker, 1988). In decision-making information is a prerequisite (Simon,
1960) and the decision-making process is deeply affected by culture (Adler, 1986).
Thus the meaning of information and effectiveness of an information system can
vary substantially in different cultures. National cultures have long been associated
with differences in the organising and operating of businesses and, more recently,
cultures specific to organisations have been studied.
3In a wider sense, Ouchi (1981) signalled the importance of national values as
they impact upon corporate culture. He established a clear link from Japanese
national culture to the corporate cultures of major organisations and then to the
outstanding success of Japanese business. His interest was the possibility of
transferring or creating Japanese-like corporate values (and hence culture) in
American industry in order to generate similar successes. He also reported that
some American organisations already had cultures much like Japanese
organisations and, he argued, this was significant in their success. We now
examine alternative theoretical frameworks for culture that can be used to examine
information in an organisational setting.
2. THE TRANSACTION COSTS PERSPECTIVE
From the viewpoint of Williamson (1975), organisations come into their very
existence because of information. The uncertainty of the marketplace,
characterised by information about transactions, drives individuals into forming or
joining organisations, while the continuing uncertainty in the environment of the
organisation leads it to changes in its strategy and structure. This view reduces all
business activity to transactions between individuals and groups, with information as
the controlling resource. While such a simple and powerful mechanism is attractive
to some information technologists, giving primacy to information and economic
activity, it is too simplistic to deal with the real social, psychological and political
settings of most organisations. It is an example of what Bolman and Deal (1991)
term the "Structural Frame" of organisations and does not immediately link to the
other frames the human resource, political and cultural.
Williamson's ideas were extended by Boisot (1987). Boisot's aim is to
incorporate a cultural perspective into transactional costs approach. He looks at
4information in organisations through two attributes of the information, its
"codification" and "diffusion":
codification the degree of formal representation,
diffusion the degree of spread throughout the population,
and the ways that these two dimensions affect information transactions. This builds
on Williamson by realising that the effect of internalising the transaction within the
organisation is to reduce its diffusion. Thus the diffused information in the
marketplace becomes undiffused in the bureaucracy.
Dichotomising organisational forms based upon the two dimensions of
codification and diffusion leads to the categories shown in Fig. 2. Codified
information is the commonplace in formal business settings and so gives rise to the
major structural forms. If information is centralised (i.e. undiffused) a bureaucracy is
the form, while if it is widely distributed a market is in effect. Bureaucracies
correspond closely to Williamson's description of hierarchies. The additional
dimension of codification in particular the absence of codification produces fiefs
and clans. A fief is controlled by an individual in whose mind most of the real
("soft") information resides while a clan has diffused but uncodified information, such
as in a group of like-minded professionals.
5 Fig. 2 Organisation forms with Information Codification and Diffusion
(Boisot, 1987)
Bureaucracy
Fief
Market
Clan
DIFFUSEDUNDIFFUSED
CODIFIED
UNCODIFIED
(hierarchy)
INFORMATION INFORMATION
INFORMATION
INFORMATION
Although Boisot's model is interesting it adds little towards the solution of the
present problem. Formal information systems are all essentially codified at similar
levels and most organisations keep most of their information to themselves, that is,
the information is not diffused. These levels of codification and diffusion will vary
from one system to another, however in different departments of organisations or in
different organisations, most such variations are of negligible proportion. Clearly, if
a company is to develop or participate in an electronic marketplace it will change the
boundary of the proprietary information, but such a boundary will continue to exist.
A further criticism concerns the attributes themselves they are not widely
known in the literature. A more common concern is "standardisation," which is
simply a form of codification. So while Boisot considers an organisation with
codified and undiffused information to be classified as a bureaucracy, for Mintzberg
(1979) the nature of the codification is critical. Standardisation of inputs, processes
and outputs are used to distinguish coordination mechanisms and hence different
organisational structures. As Boisot and Child (1988) points out, the transactional
spectrum of Williamson (markets and hierarchies) becomes the norm for most
established, advanced industry and commerce. Thus the significance of the
6"uncodified information" is seen to be small. We will draw upon the ideas of Boisot
later in this paper, in an attempt to integrate the alternative approaches.
3. HOFSTEDE'S DIMENSIONS
Hofstede's (1980) massive study of national culture has since been replicated and
used extensively. Its authority is enhanced by its predictive ability and its synthesis
of previous partial results. Hofstede used a collection of 117,000 questionnaires
from 88,000 respondents in 66 countries, all employees of the same multinational
corporation, which enabled employees in different countries to be matched. Four
dimensions of national culture were found and index scores developed for each of
40 countries (Hofstede, 1980:85).
power distance the degree of inequality of power between a person
at a higher level and a person at a lower level, (being subservient the
boss),
uncertainty avoidance the extent to which future possibilities are
defended against or accepted (not facing the future or trying to
organise it),
individualism the relative importance of individual goals compared
with group or collective goals (looking after oneself),
masculinity the extent to which the goals of men dominate those of
women (assertion - nurturance).
Power distance and individualism are strongly negatively correlated and
represented only a single factor in a confirmatory factor analysis, however Hofstede
argues that they are conceptually different and were independently developed as
indices with reference to extensive literature bases. Uncertainty avoidance also
shows weaker correlations, positive with power distance and negative with
7individualism. Thus these dimensions are not orthogonal but nevertheless refer to
four "universal problems of mankind" (Hofstede, 1980:313).
Uncertainty avoidance is the domain of information systems, planning
systems, decision support systems and control systems. It is noteworthy that the
significant developments in these domains have come from countries that Hofstede
found to have low levels of uncertainty avoidance. In these countries, the USA, UK,
the Netherlands and Denmark, the future is accepted as having much uncertainty
with which one must live. In countries with much higher scores unfavourable
outcomes must be controlled against.
The second of Hofstede's dimensions that is critical to the IS designer is
power distance, which allows for varying relationships between superiors and
subordinates in organisations. If the boss is powerful and cannot be contradicted
then a formal, accurate management information system may be seen as
unacceptable, being "insubordinate" and possibly threatening. When uncertainty
avoidance is taken together with the power distance dimension, Hofstede shows
(1980:319) that levels of these establish four organisational types, as shown in Fig.
3.
8Fig. 3 Uncertainty avoidance, power distance and organisational forms
(from Hofstede, 1980)
Bureaucracy
Full bureaucracy
Implicitly-structured
Personnel bureaucracy
High uncertainty avoidance
Low
High
Workflow
Power distance
Power distance
Low uncertainty avoidance
Non-bureaucracy
These organisational types correspond to those found in the Aston studies
(Pugh and Hickson, 1989:13) through the association of dimensions:
Aston dimension Dimension associated by Hofstede
Structuring of activities Uncertainty avoidance
Concentration of authority Power distance
Thus for each of the four combinations (high and low values on both scales)
there is a preferred organisational form that varies with culture.
Furthermore, Hofstede specifies an underlying "implicit model" that describes
the nature of interactions taking place between people in each of the quadrants, that
is, the types of transactions. The names for the various implicit models correspond
closely to Boisot's four forms. We can then see some form of association between
Hofstede's uncertainty avoidance and power distance and Boisot's codification and
diffusion of information. Superimposing Boisot's model on Hofstede's, and also
incorporating the implicit models, yields a composite framework, shown in Fig. 4,
that links information and structure to culture.
9Fig. 4 Power Distance and Uncertainty Avoidance Impact on Organisational Types
(Hofstede, 1980:319)
Full bureaucracy Personnel bureaucracy
Low
High
Workflow bureaucracy
Power distance
Power distance
Implicitly-structurednon-bureaucracy
HOFSTEDEHigh uncertainty avoidanceLow uncertainty avoidance
HOFSTEDE
BOISOT
BOISOT
Undiffused information Diffused information
CodifiedInformation
InformationUncodified
IM: well-oiled machineB: bureaucracy
IM: pyramidB: fief
B: marketIM: market
IM: familyB: clan
KeyIM - implicit modelB - Boisot's form
The figure has been presented so that it is directly comparable with Boisot's
information codification and diffusion model (Fig. 2) with Boisot's labels added as
well as the "implicit model" field. This shows power distance and uncertainty
avoidance as empirically-established variables that explain more effectively the
variation found by Boisot. The information codification and diffusion dimensions
augment the power distance and uncertainty avoidance ones, establishing
configurations of culture, information and organisation that are self-reinforcing.
Hofstede goes further, to dichotomise national cultures on the basis of
uncertainty avoidance and power distance. Thus, for example, Anglophone
countries, Scandinavia and the Netherlands have low uncertainty avoidance and low
power distance whereas Mediterranean countries score high on both. These are
presented in Fig. 5 along with Boisot's categorisation. This is not intended to say
that, for example, all Finnish organisations will be bureaucracies, rather that, under
10
given circumstances, a Finnish organisation is more likely to be structured as a
bureaucracy.
Fig. 5 Association of Power Distance and Uncertainty Avoidance with National
Culture (Hofstede, 1980:319)
Low
HighPower distance
Power distance
HOFSTEDEHigh uncertainty avoidanceLow uncertainty avoidance
HOFSTEDE
B: bureaucracy
B: fief
B: market
B: clan
Key
B - Boisot's form
German-speaking,Finland, Israel
English-speaking, Scandinavia, Netherlands
Latin, Mediterranean,Islamic, Japan, some
other Asian
Southeast Asian(esp. HK, Singapore)
Hofstede (1980:316) produces a scatter diagram of uncertainty avoidance
against power distance, which shows the distribution of the national cultures. This
was used in his categorisation of countries that was shown above. This material is
widely quoted in the literature and is redrawn in Fig. 6 below. It was drawn in
Hofstede's original with four quadrants based upon apparently arbitrary divisions.
The dividing values for Hofstede's quadrants are neither the mean values nor the
medians. They were, however, values that placed the Netherlands, Hofstede's own
country, as nearest to the "norm." Using either of the usual criteria the bivariate
mean or median this central position would be accorded to Pakistan. Perhaps a
subtle cultural influence is at work. A more serious problem is that the groupings of
countries devised by Hofstede are based on language and ethnicity rather than
cluster analysis. Adler remarks that organisations scoring high in both the power
distance and uncertainty avoidance dimensions will stress the vertical lines of
11
communication, while those low in both dimensions will have less hierarchy and
"everyone talks with everyone" (Adler, 1986:41).
Fig. 6 National Power Distance and Uncertainty Avoidance Values
(Hofstede, 1980)
Uncertainty avoidance
Power distance
0
10
20
30
40
50
60
70
80
90
100
0 20 40 60 80 100 120
S.Africa
Thailand
TaiwanPak.
NZ
Italy
GBSweden Norway
FranceBelgium
Spain
Argentina
Germany
Portugal
USA
AUSNL
SingaporeYugoslavia
Greece
Philippines
Mexico
Israel
Ireland
Iran
India
HK
Denmark
Austria
Further analysis of Hofstede's indices on all four dimensions for the forty
countries shows that the country nearest the global mean is Iran, which supports the
suggestion that the "Islamic" group should be removed from the lower left quadrant
of Fig. 5. If this means that Iran is the country where the four conflicts (superior /
subordinate, male / female, individual / group and risk taking / risk averse) are most
in balance, many Anglophones and Europeans would be perturbed this is quite
contrary to received media reports.
On the other hand, viewing the data from the centre, where Iran's neighbours
are Brazil and Turkey, shows that the extreme positions are held by Denmark (low
masculinity, uncertainty avoidance and power distance but with high individuality),
12
Singapore (low individuality and uncertainty avoidance but high power distance) and
Venezuela (low individuality but high on power distance, masculinity and uncertainty
avoidance). Of all forty countries those pairs with closest neighbours are USA /
Australia, Taiwan / Pakistan, and Switzerland / Germany. The countries with no
close neighbours (in these four dimensions) are Japan (the most isolated), Israel
and Austria. It is possible that Austria is only isolated because its formerly
communist geographical neighbours, Czechoslovakia and Hungary, were not part of
the study.
Before looking at some further limitations of Hofstede's work, it is worth
considering the relevance of these points to the IS designer. High levels of
uncertainty avoidance can raise some organisational activities to the level of a ritual
that is carried out for its own sake, rather than for the results that are achieved. For
example, excessive reliance on planning and control systems, stronger belief in
accounting figures than in the grass-roots "gut-feel" for the organisation's
achievements, and excessive use of formalised reports and memos. A management
information system may attain an unwarranted status, with its reports being believed
too much. The IS designer may introduce some scepticism in interaction with users,
but may also have to look for ways to do it in the formal reports such as by
incorporating contradictory data or results obtained using alternative formulas. On
the other hand, a low level of uncertainty avoidance, with its attached fatalism, may
lead managers to have no interest in information systems. "Why bother to print out
these things, it won't change anything." The ability to control or plan in the
organisation may be seriously doubted. IS designers may not find themselves with
this problem, as they may well not be employed in such organisations. However,
outside consultants should well recognise that this is not fertile ground for their
services.
13
A high level of power-distance will mean that management information may
well be precisely that, information only available to management. Access to
information comes with the increased power of the manager. Pressure may well be
put upon the information providers to generate only the information that is
acceptable and subordinates may readily agree. Independent control and audit may
be curtailed. The Singapore government's rejection of "information" in The
Economist (1993) shows large power distance at work. The IS designer should
realise that, particularly in such settings, designing a system is above all a political
act and should then act accordingly. If power-distance is low the environment is
friendly, cooperative and egalitarian. Although there is then a danger of doing
nothing, it is in this situation that information systems can be readily developed.
However, it will not be constructive to use methodologies that are authoritative or
controlling (see Kendall and Kendall, 1993).
There are two significant cautions to be applied to any interpretation or use of
Hofstede's work. Firstly the values expressed are particularly those of the middle
class standardisation of the country samples was effected through the marketing
and service functions of a single organisation's personnel, with an emphasis on
managerial, sales, technical and administrative staff (Hofstede, 1980:73). Although
this limitation may not be critical to the IS designer who is facing precisely these
sorts of people as clients. The second caution concerns its significance. While the
differences between countries were highly significant statistically the
standardisation procedures removed many confounding sources of variation the
amount of variation between individuals explained by country is only 4.2%
(Hofstede, 1980:71); gender, age and occupation are more significant. The very
large sample size leads readily to statistically significant results. So, if IS designers
are going to be concerned about cultural impacts then they should take more
concern over gender, age and occupation.
14
Wilkins and Ouchi (1983) expand on this second concern, by examining the
conditions under which cultural issues assume greater significance, in particular
organisational culture. Taking a transaction cost approach they point out that both
the market and bureaucracy settings are dominant; they are the most efficient
arrangements under a range of conditions. The clan, the organisational form where
the group culture is most important, is only the most efficient form when the
environment in which the organisation is located is both complex and dynamic.
Decisions that are good for the organisation as a whole can then be made by
following the cultural norms; the goals held by the members are congruent. The
following circumstances are seen as necessary conditions for cultural values to be
dominant:
long history and stable membership (such as an army regiment,
without conscripts),
absence of institutional alternatives (only hiring raw recruits, for
example), and
full interaction among members (the culture can only spread to those
people who participate in it).
In addition there must be slack resources to allow for the inefficiencies of
devoting time to the passing on the culture, such as an army unit would have in
peacetime. These are thus seen as necessary conditions for the development of
strong organisational culture as an important part of the efficient functioning of the
organisation. The army at peace is thus able to develop a strong culture that
becomes the most effective manner in which it is to operate when the environment is
both complex and dynamic that is, at war. Wilkins and Ouchi further suggest that
the clan unit may only be applicable in part of an organisation.
15
4. INFORMATION REJE CTION
An alternative approach to those of Hofstede and Boisot is from Thompson and
Wildavsky (1986). They recognise the mythology of the age, that we are in the
"information era" in an "information society," however they note paradoxically that
the most significant behaviour of people with respect to information is rejection.
They describe four fundamentally distinct types of information rejection, which they
then associate with four strategies for forming organisations. The four types of
information rejection are seen to be mutually exclusive and exhaustive. They are
shown in Table 1 below.
16
Table 1 Information rejection types (Thompson & Wildavsky, 1986)
Information
rejection
type
Example Characterisation
Risk
absorption
"What you don't know
won't hurt you" (popular)
Fatalism; acceptance of a
world in which life is like
lottery
Networking "When I feel like reading a
book I write one"
(Benjamin Disraeli,
overburdened by data)
There is so much data that
it is all rejected in favour of
informal networking based
on personal relationships
Paradigm
protection
"These preposterous
theories of Prof. Ohm"
(the scientific
establishments reaction to
the first attempt to publish
Ohm' Law)
A powerful hierarchy
whose structure is
threatened closes its ranks
to reject information that
questions its foundation
Expulsion "If it was good enough for
Moses it's good enough
for me" (fundamentalist
song rejecting Darwin's
theory of evolution)
Not hierarchical; a sect-
like group closes ranks to
protect its vulnerable
members from the
predatory outsiders
Thompson and Wildavsky then draw parallels with the transaction costs
approach and the two fundamental organisational forms of Williamson, markets and
hierarchies, and Ouchi's clans. This now starts to fall into the same problems as the
Boisot approach, lacking sufficient models to describe current organisational
17
realities. We develop a richer approach by following Thompson and Wildavsky but
associating their categories of information rejection and the four dimensions of
culture found by Hofstede (1980). This is demonstrated in Table 2, below.
Table 2 Information Rejection and Culture (adapted from Thompson and
Wildavsky, 1986 and Hofstede, 1980)
Information
rejection type
Culture Hofstede category
Risk absorption Fatalism (clan) Low uncertainty avoidance
Networking Markets Low masculinity
Paradigm protectionHierarchies Low individuality
Expulsion Sects Low power distance
The full descriptions of the four cultural types proposed by Thompson and
Wildavsky are particularly precise characterisations of the (relatively extreme in
Hofstede's dimensions) national cultures in the countries indicated in Table 3,
shown as "corresponding countries."
18
Table 3 Information Rejection and Culture
Information rejection
type
Corresponding
countries
Opposite
countries
Risk absorption Singapore
Hong Kong
Greece
Portugal
Networking Sweden
Denmark
Japan
Paradigm protection Pakistan
Taiwan
USA
Australia
Expulsion Israel
Austria
Philippines
Mexico
Each culture identified by Thompson and Wildavsky represents a low level of
one of the dimensions of Hofstede. The limitation of Thompson and Wildavsky's
work is now made apparent by noticing the "opposite" countries, those with high
values on these dimensions. Among these there are also some strong cultures.
This shows that the multi-dimensionality cannot be ignored; a reduction to four
distinct alternatives is patently inadequate. There does remain, however, the very
significant finding that low levels on each of the Hofstede indices can be associated
with characteristic types of information rejection a most valuable warning for IS
designers.
5. CONCLUSIONS
The nature and role of information is central to the study of organisations, and, just
as emphatically, the culture has a critical impact on the selection, analysis and
design of information systems. The first step for the IS designer, before any
19
systems investigations are commenced, has to be a study of the organisation's
culture, and in the case of transnational and multinational systems, the national
cultures involved. These will give warnings of modes of information handling,
supervision and control that will be intimately concerned in any information system
to be introduced. We suggest that such an investigation could start with the
questionnaire of Hofstede (1980) and then use some of the above alternative
analyses.
National culture
Hofstede's dimensions of power distance and uncertainty avoidance may be of great
significance to the IS designer, especially if the values are extreme, that is, very
high or very low. Such extreme values can lead to systematic rejection of
information that conforms to recognisable types. Other extreme values may lead to
over-reliance on information to the detriment of the organisation. By being aware of
the environment the IS designer may be able to foresee some of the dangers.
These are most acute for an designer who is not a national of the country where
development is taking place. The dangers can be summarised briefly as:
If uncertainty avoidance is strong then an MIS is wanted to try to
reduce the uncertainty even if that is impossible; systems may become
rituals,
If uncertainty avoidance is weak, fatalism leads to scepticism about
MIS and resistance from users,
If power distance is large then the boss disagrees with the MIS and
the boss is right,
If power distance is small, authoritative approaches will be risky.
20
Organisational culture
Organisational culture is most valuable in a complex dynamic environment, such as
in what Mintzberg (1979) terms an adhocracy. In this situation information may
become part of the team; networking is critical and a formal management
information system is not important. Organisational culture may be a hindrance in
other environments the above attitudes will prevail although they are not
necessarily beneficial.
System development and change processes
Lederer and Nath (1990) have gone so far as to say that the traditional lifecycle
model used for systems development is inadequate because the fundamental
process is one of organisational change (rather than systems development). The
form that organisational change takes depends very strongly upon the culture in the
organisation. Kendall and Kendall (1993) consider the suitability of various system
development methodologies to the style of leadership that is prevalent in the
organisation. An investigation of culture will assist in an clear understanding of the
leadership characteristics in the organisation.
As a consequence of their direct impact on the working environment of
people, information systems are implicitly associated with organisational change, as
Keen (1981) pointed out. Current developments in group decision making support
technology will have significant organisational impacts, in the same way that end
user computing has had already. The influence of organisational and national
culture on the manner of organisational change and the ability of the organisation to
change readily, is of paramount concern, but unfortunately one that appears to be
little researched.
21
6. REFERENCES
Adler, N.J. (1986) International Dimensions of Organizational Behavior, Kent Publishing Co.: Boston, Mass.
Boisot, M. (1987) Information and Organizations: The Manager as Anthropologist,Fontana: London.
Boisot, M. and Child, J. (1988) The Iron Law of Fiefs: Bureaucratic Failure and theProblems of Governance in the Chinese Economic Reforms,Administrative ScienceQuarterly, 33, 507-527.
Bolman, L.G. and Deal, T.E. (1991) Reframing Organizations: Artistry, Choice and Leadership, Jossey-Bass: San Francisco, CA.
The Economist (1993) Singapore Saga, August 14-20 1993, 328, No. 7824, 4-6.
Hofstede, G. (1980) Culture's Consequences: International Differences in Work-Related Values, Sage: Beverly Hills, CA.
Keen, Peter G.W. (1981) Information Systems and Organizational Change, Communications of the ACM, 24, 1, 24-33.
Kendall, J.E. and Kendall, K.E. (1993) Metaphors and Methodologies: Living Beyond the Systems Machine, MIS Quarterly, 17, 2, 149-171.
Lederer, A.L. and Nath, R. (1990) Making Strategic Information Systems Happen,Academy of Management Executive, 4, 3, 76-83.
McGregor, D. (1960) The Human Side of Enterprise, McGraw-Hill: New York.
Mintzberg, H. (1979) The Structuring of Organizations: A Synthesis of the Research,Prentice-Hall: Englewood Cliffs, NJ.
Ouchi, W.G. (1981) Theory Z: How American Business Can Meet the Japanese Challenge, Avon: New York.
Pugh, D.S. and Hickson, D.J. (1989) Writers on Organizations, 4th ed., Penguin, London.
Simon, H. A. (1960) The New Science of Management Decision, Harper & Row: New York.
Thompson, M. and Wildavsky, A. (1986) A Cultural Theory of Information Bias in Organizations, Journal of Management Studies, 23, 3, 273-286.
22
Tricker, R.I. (1988) Information Resource Management A Cross-Cultural Perspective, Information & Management, 15, 37-46.
Wilkins, A.L. and Ouchi, W.G. (1983) Efficient Cultures: Exploring the Relationships between Culture and Organizational Performance, Administrative Science Quarterly, 28, 468-481.
Williamson, O.E. (1975) Markets and Hierarchies: Analysis and Antitrust Implications, Free Press: New York.