Towards a more scientific way of studying multicultural management

In recent years, the management of cultural diversity has become a popular topic within management in general and organisational behaviour and human resource management in particular. The progressive movement of the world into the post-modern era calls for a comprehensive, practical and realistic strategy to change not only the demographics of organisational members and their attitudes, but also the way in which multicultural organisations are managed. Controversies and uncertainties regarding current outcomes still prevail, possibly because of a lack of application of more scientific management styles and methodologies. Management approaches often appear to be based on unfounded assumptions, superficial observations and simplistic assessments. As a result, the derived outcomes are incongruent, oversimplistic, unrealistic and unreliable. To ensure the success of organisations in the growing global market, managers need to establish an effective approach to managing cultural diversity. This article examines some of the traditional concepts of culture and cultural diversity and employs illustrative empirical data to suggest a more scientific approach to the measurement of cultural diversity and its impact on the work environment and management.

• Miscommunication, misunderstanding and conflict within multicultural organisations are due to a single factor while ignoring the potential contribution of other cultural factors, such as differences in sex, age, educational level, political conviction, social class, and so on. 2The lack of empirical literature on the dynamics of culturally diverse work groups and the effective management of such groups have led most prior researchers to use short-term research subject groups that existed only for the duration of their studies.Also, the tasks employed by the subjects have been overly simplistic and/or have had a gamelike quality, with no significant impact on group members' well being, so that validity of the studies was limited.
Such approaches have resulted in derived outcomes being incongruent, over simplistic, and unreliable.As evidence of incongruent outcomes, some studies have concluded that racial diversity inhibits group performance (e.g., Fenelon & Mergagee, 1971;Ruhe & Allen, 1977); some find no performance differences between racially diverse and racially homogeneous groups (e.g., Matsui, Kakuyama & Onglatco, 1987); and some conclude that racial diversity enhances the performance of groups (e.g., Ruhe & Eatman, 1977;Watson et al., 1993).Also, some specialists in the field argue that national or regional culture is rarely present in the firms (Maurice, Serge & Warner, 1980;Maurice, Sellier & Silvestre, 1992;Amadieu, 1993) or that it is empowered by organisational culture (Ivanier, 1992), while others affirm that national or regional culture is dominant when compared with organisational culture (e.g., Laurent, 1983;D'Iribarne, 1986;Hofstede, 1989;Meschi & Roger, 1994). 4ith regard to over simplistic and unrealistic research in the area of multicultural management, the need for comparisons between sub-groups in various contexts has resulted in a need to reduce complexity (Evers, 1991).But, though these analyses serve as useful background information, oversimplification of the problem has sometimes rendered the results ineffective as recommendations for action.
The purpose of this article is to present an assessment of previous studies of multicultural management, to propose improvements to the study of multicultural management following the approach described by Rijamampianina (1999), and to illustrate the proposed improvements using empirical data collected at two Malagasy firms by Rijamampianina (1999).

Towards scientific concepts
The concept of culture Most management researchers view 'culture' as ideas or common theories of behaviour or mental programmes that are shared by a group of individuals (e.g., Allaire & Firsirotu, 1984;Hofstede, 1984Hofstede, , 1989Hofstede, , 1991Hofstede, , 1994)).Culture is therefore not considered to be an individual characteristic.Researchers tend to agree that the majority of the components of culture (e.g., beliefs, values, norms, perceptions, attitudes and priorities) are essentially invisible (Kotter & Heskett, 1992;Schein, 1993), which makes them more difficult to understand and cope with successfully.They also tend to agree that in any given culture, some values are regarded as more central than others (e.g., Kluckhohn & Strodtbeck, 1973;Rokeach, 1973).Lachman, Nedd and Hinings (1994) used the terms 'core' and 'periphery' to represent, respectively, the relatively high and the relatively low positioning of values in the value hierarchy and the extent to which they influence social control.
None of the prior studies cited, however, consider the interrelationship between cultural visibility and centrality of values, and hence they are not able to consider the impact of cultural visibility and centrality of values on cultural change and conflict in intercultural interactions.
As an extension of the work of Lachman, Nedd & Hinings (1994), Rijamampianina (1999) proposed the relationship described in Figure 1, in which values higher in the hierarchy -that is, core values -are more important, more enduring and more resistant to change, particularly if they are invisible.For example, North Americans attach primary values to freedom and independence, and Arabs highly value religious belief and devotion.Such values are highly accepted within the cultural groups; thus they are more likely to be resistant to change and to cause conflict5 in intercultural interactions.Resistance to change lessens and the chance of conflict decreases when core values become visible.By contrast, values of low priority, low consensus, and lower importance -that is, peripheral values -are relatively susceptible to change (Shils, 1961).They are more easily changed and are less likely to cause conflict when they are visible (e.g., housing, clothing) than when they are not.Thus it is not all cultural differences that are likely to generate conflict in intercultural interactions.Small differences in core values are more likely to generate conflict than a large difference in peripheral values.Similarly invisible values are more likely to cause conflict than visible values.

The concepts of cultural diversity and cultural distance
According to Cox (1993), since a cultural group refers to an affiliation of people who collectively share certain norms, values or traditions that are different to those of other groups, 'cultural diversity means the representation, in one social system, of people with distinctly different group affiliations of cultural significance' (Cox, 1993: 6).Mazrui (1994) provides a more specific definition by labelling three types of societies as follows: 1.A homogeneous society is one in which over 80% of the population is of the same cultural tradition.
2. A preponderant society is one in which over 50% of the population belongs to the same cultural tradition; and 3. A heterogeneous society is one in which all cultural groups constitute less than 50% of the population.
Although Cox (1993) and Mazrui (1994) both consider all social systems as being culturally diverse, Mazrui (1994) argues that the degrees of diversity differ depending on the proportions of the cultural groups within the social systems.
By contrast, Rijamampianina (1999) argued that cultural diversity is defined by the significance of the cultural distances6 between existing cultures rather than by the number of cultural groups within the social systems, or their respective proportions.It follows then that where cultural distances between group members are not significant, the social system could not be viewed as multicultural, whereas significant cultural distances indicate the existence of a multicultural group.
Rijamampianina (1999) also affirms that cultural distances may be based on gender, race, nationality, tribal or ethnic group, age, educational level, occupational level and so on.
It is therefore misleading to state that cultural diversity is caused by a single factor, though indeed, one of them may be more significant than the others.

Toward scientific measurement of cultural diversity
Very few empirical studies have been conducted into the dynamics of culturally diverse work groups and the effective management of such groups -a serious failing given the centrality of these concepts in management.Empirical studies that have been conducted in these areas (e.g., Hofstede, 1989Hofstede, , 1991) ) measured cultural distances between countries or ethnic groups by comparing the extent of the difference in average scores on specific dimensions of culture.In other words, such studies typically considered the differences in the mean scores, without considering the distributions of the scores around those means.Such an approach fails to give an indication of the extent to which the cultural dimension is "core" or "periphery" for the group concerned (as illustrated in Figure 1).In addition, differences in average scores do not reflect the extent to which values are visible or invisible, and therefore offer no indication of the difficulty that could be encountered in surfacing, measuring and managing those issues.Rijamampianina (1999) suggested that data should be analysed using a framework that takes into consideration the extent to which the dimensions are core or periphery, visible or invisible.To establish the extent to which a cultural dimension is core to a sample of people, the variance of their scores around the mid-point of the distribution has to be analysed.In the case of comparing the scores of cultural sub-groups, the differences in mean scores need to be tested to determine whether or not they are significantly different to zero.When a cultural distance has high absolute value but is not significantly different to zero, it suggests that the cultural dimension relates to visible and peripheral values, and when the cultural distance has a low absolute value but is significantly different to zero, it suggests that the difference comprises invisible and core values.Differences that are significant are more likely to cause resistance to change and conflict associated with attempts to bring change in intercultural interactions.
Practically, two-sample tests or Analysis of Variance (ANOVA) procedures can be used to identify those cultural distances that are significant.In the case of ANOVA procedures, multiple comparison methods can elicit detailed information about the significance of the differences between individual sub-groups in a study and allow the control of error rates for a multitude of comparisons.
Various theorists have suggested dimensions relevant to the description of cultures.While it is possible to use any one of a number of frameworks offered in the literature, Rijamampianina (1999) suggested a combination of those developed by Kluckhohn & Strodtbeck's (1973) and Hofstede (1989), as described below.
Power Distance (PD) -defines the extent to which a group of people accepts the unequal distribution of power in institutions and organisations; Uncertainty Avoidance (UA) -defines the extent to which people in a culture feel threatened by uncertainty and ambiguous situations and try to avoid such situations; Masculinity Orientation (MO) -indicates the extent to which the dominant values of a group of people are 'masculine' (e.g., assertion, competition, self-achievement); Group Orientation (GO) -implies the extent to which a group is a tightly knit social framework in which the 'ingroups' are expected to take care of their members; Task Orientation (TO) -indicates the extent to which the dominant values of a group of people are task related; Space Orientation (SO) -defines the extent to which the dominant values of a group of people are space related; Human relation Orientation (HO) -indicates the extent to which the dominant values of a group of people are human relation oriented.

Toward scientific measurement of the effects of cultural diversity on organisational performance
Empirical studies into culture typically assume a static environment (e.g., Hofstede, 1984Hofstede, , 1989Hofstede, , 1991Hofstede, , 1994)).They do not investigate interactions between cultures or dimensions of culture, nor do they investigate the impact of dimensions of culture on organisational processes.Though some researchers (e.g., Evers, 1991) claim that it is scientifically difficult to measure the interactions of various cultures in a dynamic system, numerous analytical tools are available to support a more scientific and realistic analysis.Rijamampianina (1999) proposed the use of the Covariance Analysis of Linear Structural Equations (CALISE) procedure -available, for example, within the SAS package.
The CALISE procedure can be used to estimate parameters for constrained and unconstrained problems in multiple and multivariate linear regression, linear measurement-error models, path analysis and causal modelling, simultaneous equation models with reciprocal causation, exploratory or confirmatory factor analysis of any order, and canonical correlation.
In order to assess the impact of cultural diversity on organisational performance, Rijamampianina (1999) analysed the impact of cultural diversity on each of the four inter-related organisational processes described below:

Motivation:
The process by which employees share in the success or failure of the organisation.In other words, the process by which internal and external forces act upon employees, predisposing them to act in a manner beneficial to the achievement of the objectives of the organisation.

Interaction:
The process by which employees share their mental models.In other words the process by which employees share their own views of the world, generate an understanding of the assumptions driving other stakeholders in the workplace, and give meaning to organisational communications and actions.

Visioning:
The process by which employees articulate and / or internalise an understanding of the essential intentions of the organisation; and Core competence development: The process by which employees are empowered to meet the full range of challenges they face at work.

Illustrative example
In this section the elements of the proposed approach described above are illustrated in their application to data collected by Rijamampianina (1999) in his study of culture in the Madagascan firms, KRAOMA and STAR.

KRAOMA and STAR: cultural context
Kraomita Malagasy (KRAOMA) and Star-Madagascar (STAR) are both limited liability companies in Madagascar.KRAOMA employs some 400 workers whose main activities are to extract, treat and commercialise mineral chrome (chromite).With its 1 400 employees, STAR's main activities are to produce and commercialise gasified and non-gasified alcoholic and non-alcoholic beverages.
Madagascar comprises six provinces (Antananarivo, Antsiranana, Fianarantsoa, Mahajanga, Toamasina, and Toliary) separated by virtual borders.Although 18 discrete ethnic groups can be identified in Madagascar, Covell (1987) categorised the three main ethnic groups according to their geographical locations, activities, work ethics, and cultures, and defined the groups as: Coastal, Merina, and Highland.
The cultural diversity evident in Madagascar is explained to some extent by the diversity of the peoples settling there during the course of the last 2000 years; from Indonesia, India, Arabia, Europe and Africa (Rijamampianina, 1999).
While Madagascans can all speak Malagasy, important linguistic, cultural and physical differences are evident in groups concentrated in different parts of the country.
Each of the provinces contain people from each of the three groups, but the proportion of Coastal people decreases as the distance from the coast increases.The proportion of Highlanders is dominant midway between the coast and the innermost regions of the Island, and the proportion of Merina people is highest in the innermost regions of Madagascar.In recent history, the Merina group appears to have been dominant with regard to social, economic and political power.Tensions that have developed between the three groups, particularly through the colonial era beginning in the late 1800s, are described in more detail by Rijamampianina (1999).Due to their multi-ethnicity (i.e., the fact that they employ members of each of the three ethnic groups), KRAOMA and STAR can be considered representative of Malagasy organisations, and therefore serve as appropriate illustrative cases of culturally diverse contexts.

Data collection
The empirical data for this research were collected via a questionnaire, distributed to a sample of respondents from each region employed by each of the two companies.The questionnaire was designed to measure the cultural diversity of the workforces of the two companies, assess the multicultural management practised within the two firms, and appraise the relationship between cultural diversity, management, and performance in the organisations.
One hundred and fifty questionnaires were distributed to employees of each of the companies.One hundred and two completed questionnaires were received from KRAOMA employees, and 103 completed questionnaires were received from STAR employees, representing a 68,3% response rate.After obtaining each cultural group's standardised means for each of the cultural dimensions described in Section 3, the means of each sub-group were compared using the ANOVA procedure, to determine which cultural distances were significantly different to zero.The results are presented in Table 2.
Although KRAOMA and STAR are both multiethnic companies, the differences between the standardised means of each cultural group for Uncertainty Avoidance, Group Orientation and Human relation Orientation were not significant (despite the apparently high values of some of the distances).That is, in both companies, differences in these dimensions are not likely to cause conflict because they are mainly composed of peripheral and visible 7 values.
Differences between the standardised means of each cultural group for Power Distance, Masculinity Orientation, Task Orientation and Space Orientation were significant.This shows that differences in these dimensions are likely to consist mainly of core and invisible values, and thus are likely to cause conflict in intercultural interactions.
In order to determine the impact of cultural diversity on organisational performance, questionnaire responses relating to each of the four organisational processes described in Section 4 were analysed using the CALISE procedure to test the following alternate hypotheses.These hypotheses 7 It is likely that core values are less visible than peripheral values because as values become more visible, they are opened to critical examination, which stimulates or allows cultural dynamism.This could have the effect of moving the values along the centrality dimension towards periphery.
propose a causal impact of dimensions of culture on successful operation in each of the organisational processes (described as paths 1 to 4 in Figure 4), and they propose that the impact on performance will differ across ethnic subgroups.
Hypothesis 1a: Cultural distances between cultural groups will affect their success sharing.

Hypothesis 1b:
The effect of cultural distances on success sharing will not be the same for all the cultural groups.
Hypothesis 2a: Cultural distances between cultural groups will affect their mental model sharing.

Hypothesis 2b:
The effect of cultural distances on mental model sharing will differ for all cultural groups.
Hypothesis 3a: Cultural distances between cultural groups will affect vision sharing.

Hypothesis 3b:
The effect of cultural distances on vision sharing will differ for all cultural groups.
Hypothesis 4a: Cultural distances between cultural groups will affect their core competence development.

Hypothesis 4b:
The effect of cultural distances on core competence development will differ for all cultural groups.The coefficients presented in Table 3 indicate the impact of the core cultural dimensions (power distance, masculinity orientation, task orientation and space orientation) on each of the four organisational processes (motivation, interaction, visioning and core competence development) represented by the paths p1 to p4.For example, reading the top left cell of the table it is evident that the power distance coefficient (-0.4386) of the path p1 for the Merina group is significant and negative; as power distance increases, so motivation levels in the Merina ethnic group at KRAOMA decrease.
Similarly, an increase in the power distance index will result in a significant decrease in the quality of the interaction processes (p2) for Merina employees at KRAOMA.By contrast, an increase in power distance will result in a significant decrease in motivation and core competence development for Highlanders, and a significant decrease in interaction and visioning for Coasters at KRAOMA.In summary, Table 3 demonstrates that: • Coefficients of each of the paths 1 to 4 are significantly different to zero with respect to at least one of the cultural dimensions.This means that alternate hypotheses 1a, 2a, 3a and 4a can be accepted.In other words, changes in the indices of the cultural dimensions significantly affect the levels of success sharing, mental model sharing, vision sharing, and core competence development, and therefore organisational performance.In both companies, performance of the Coasters', the Highlanders' and the Merina's will be affected by the manner in which power distance, masculinity orientation, task orientation and space orientation are managed.
• Different cultural groups feel the impact of changes in core dimensions of culture differently.The coefficients of some paths are significantly different to zero for some cultural groups on some organisational processes, when they are not significantly different to zero for other cultural groups on those paths.This demonstrates that alternate hypotheses 1b, 2b, 3b and 4b can be accepted.In other words, when different cultural groups interact in the workplace, some of them are likely to be more negatively affected (i.e., disadvantaged) by the changes that occur in the dimensions of culture that are 'core' for that group of employees.This could impede the integration and combination of employees' ideas and competencies, attitudes and behaviours, ways of thinking and doing, because the cultural groups that feel disadvantaged may elect not to collaborate with other groups.
These results suggest that understanding the centrality and visibility of values of cultural groups, and the manner in which these values impact on key organisational processes in a workplace, could enhance the management of those groups towards high performance, generate more congruent and more reliable research results, and form a basis for practical multicultural management guidance.

Conclusion
Although the pace of globalisation is sufficient to encourage practitioners and scholars to find a way to manage culturally diverse groups, a scientific approach for analysing the issues is fundamental to the design of effective strategies.Indeed, without any scientific foundation, is there any difference between research in this area and building a house on the sand?
The issues of multicultural organisations are not as new as has been claimed, since many countries (and, thus, organisations) have, for some time, been multicultural.However, they are expanding now because of the pace of development in technologies and service opportunities, and the evolution of new domestic infrastructures and competitive environments.Since the era of knowledge has already been entered with great fanfare, it is time to move from the subjective-tacit-knowledge world to the objectivecodified-knowledge universe.
Figure 1: The structure of culture

Figure 3 :
Figure 2: Sources of conflict at KRAOMA Table 1 details the number of respondents who classified themselves into each ethnic group in KRAOMA and STAR.

The influence of cultural diversity on the workplace environment and management
Results of the CALISE analysis and hypothesis tests are presented in Table3, for only those dimensions of culture that were assessed as 'core' in this study.