Culture

11

satisfaction, combined with the actual effectiveness and efficiency of the product. As it is based on a literature survey, Hornbaek’s model has the advantage of consistency with other HCI research. Furthermore, it is a clarification of the ISO9241 definition of usability, and comes a long way in helping resolve the issue of connecting usability’s definition with its measurement. However, it fails to incorporate the effect of usability factors such as user, task, and product as described in the MUSIC model.

2.2 Culture

2.2.1 Definitions

The concept of culture is difficult to describe precisely, and there is little agreement on culture’s definition. By 1951, 164 different definitions for culture had been identified (Olie, 1995), suggesting that currently the number of definitions for culture must be significantly higher. Of these definitions, Straub, et al. (Straub, et al., 2002) have identified three groups of definitions of culture based on either values or cognitive style, or all-encompassing definitions of culture.

Studies of culturability however, base themselves on slightly different definitions of culture. These include differences either in nationality, communication style,

meanings, or a comprehensive set of differences including cognition and values. Badre (2000) loosely describes culture as no more than “a means of distinguishing between different countries”, indicating that culture is simply based on the nationality of the user. Cultural differences have also been based on differences in communication (e.g.

(Hall, 1976), or meaning (e.g. (Bourges-Waldegg and Scrivener, 1998, Marcus, 2006).

For Marcus, culture is based on differences in the meaning of group “behaviors, leaders/followers, values, artifacts, and signs”. Another type of definition of culture is more comprehensive and attempts to combine cognitive, emotional, communicational, and behavioral differences that help to distinguish groups of people (e.g. (Borgman, 1986, Hofstede, 1984). For Hofstede, culture is “the collective programming of the mind that distinguishes the members of one group or category of people from another, where the mind stands for thinking, feeling and acting, with consequences for beliefs, attitudes and skills”. In studies of culturability, the concept of culture is commonly separated into objective and subjective culture (Ford and Kotzé, 2005, Hoft, 1996).

Objective culture includes expressions of culture such as institutions, language, and lifestyle etc, while subjective culture includes internal, psychological manifestations of culture, such as attitudes held in common with others.

2.2.2 Models

Models of culture used in studies on culturability usually describe dimensions along which national cultures vary (Badre, 2000, Choi, et al., 2005, Ford and Kotzé, 2005, Marcus and Gould, 2000). Typically, such studies are based on Hofstede’s cultural dimensions (Clemmensen and Roese, 2010). However, given the large number of cultural dimensions identified (Lee et al. (2008) reported 36 such dimensions), discussion of the relative merits of the more common cultural dimensions is required.

These include Hall (1976), Triandis (1972), Trompenaars (1994), Schwartz (2004), Hofstede, and the World Values Survey.

A number of these models suggest descriptions of culture that are difficult to quantify. For example, Triandis suggests cultures vary according to their complexity, for example, a traditional agrarian society may have a simple structure compared to that of a society based on a developed service-oriented economy. Criteria such as these would be extremely difficult to assess numerically. Similarly Hall describes a number of characteristics by which a culture can be classified, such as cultural complexity or tightness, but no means to gauge the degree a culture can be associated with such a characteristic.

Table II-3 Models of Culture (1)

Hall Triandis Trompenaars

 Universalism vs. particularism (rules or relationships)

 Neutral or emotional

 Individualism vs. collectivism

 Specific vs. diffuse (assignment of responsibility)

 Achievement vs. ascription (how is status obtained)

 Sequential vs. synchronic (how tasks are performed)

 Internal vs. external (control over environment) On the other hand, a number of models provide quantifiable descriptions of culture that can be used to investigate culture’s effect on usability. For example, Trompenaars provides seven dimensions of culture, five relating to interpersonal relationships, the other two relating to time and space. However, apart from

individual-collective dimensions, Trompenaar’s dimensions have little correlation with those of Schwartz, Hofstede, and the World Values Survey. The dimensions provided by the latter three researchers overlaps in the area of individual-collective values, hierarchical values, and control-acceptance values. Schwartz’s model of cultural

13

which culture varied. These axes consisted of embeddedness versus autonomy, hierarchy versus egalitarianism, and mastery versus harmony. Embeddness vs.

autonomy refers to the extent an individual is embedded in a group. Hierarchy vs.

egalitarianism naturally deals with acceptance of differences based on status. Mastery vs. harmony concerns attitudes towards controlling one’s social and physical

environment. While not exactly matching Hofstede’s results, Schwartz’s showed some consistency with Hofstede’s values of individualism-collectivism, power-distance, and masculinity-femininity, respectively. Similarly, Hofstede found that the WVS

dimensions well-being versus survival correlated strongly with his own dimensions individualism and masculinity, and that the WVS dimension secular-rational versus traditional authority correlated negatively with power distance. This overlap between these three models of culture suggests confirmation of the accuracy of these

dimensions.

Table II-4 Models of Culture (2)

Schwartz Hofstede World Values Survey

 Embeddedness vs.

 Need for survival vs. the need for self-expression

 Traditional vs. secular-rational values

Given the similarity of these three cultural models, it is worth questioning why Hostede’s work is so prevalent in academic literature compared to others. Hofstede’s study was conducted among IBM employees in 49 countries in 1967 and 1973.

Hofstede describes five main values by which cultures can be distinguished. These are power-distance, masculinity-femininity, individual-collective, uncertainty avoidance, and long-term orientation. Power-distance refers to the degree to which people with less power in a group accept the authority of more powerful members. Masculinity-femininity describes the degree to which the masculine values in a culture, such as assertiveness, independence, and ambitiousness, differ from feminine values, such as caring and cooperation. Individual-collective values describe the extent to which individuals identify themselves as a group member or as an individual. Uncertainty avoidance refers to the need for an individual to have certainty, for example, provided by rules or tradition. Additionally, Hofstede describes long term orientation, identified

in cooperation with Bond (Hofstede and Bond, 1998), which describes certain Confucian-oriented values related to long and short term views of the world such as persistence and the need to save face.

A number of criticisms of Hofstede’s methodology have been made, which perhaps have implications for other cultural dimension models as well. One point raised by McSweeney (2002) was that the idea of national culture being shared by all of a country’s citizens ignores the fact that there exist a number of distinctive cultures within one nation. The existence of national subcultures, McSweeney argues, makes it hard to say that there exists for each country a single shared culture that makes that country’s culture unique. In fact, cultures exist on every level of social interaction, below and above the level of a nation, and distinctive cultures within a nation help to make up the culture on a national level. For example, the cowboy culture of the American West, while clearly distinct from suburban culture in America’s cities, is still an important part of American national culture. In addition, it could be said that there exists a North American regional culture, comprising the cultures of the USA and Canada. To study the impact of culture on usability, therefore, researchers need to consider the possibility that nationality may not be the most reliable indicator of cultural background.

In addition, many of McSweeney’s criticisms concern Hofstede’s methodology.

Hofstede’s conclusions were based on a survey of IBM employees, who do not necessarily represent the values of the country they represent. In fact, in a developing country, the values of educated, well-off IBM employees may even be that of a minority for that country. McSweeney also notes that the sample size of the survey varies between countries. For example, a country such as Singapore with a smaller local office provided much fewer results per head of population, compared to the USA with its large number of IBM offices.

Lastly, McSweeney argues that Hofstede’s survey of company employees may be overly influenced by the working context it was conducted in. Alternative values, such as family or societal values may have been ignored. In fact, Lee et al. suggest that instead of considering differences in work-related attitudes, or other attitudes such as those concerned with family or society, usability research should focus on product related attitudes. They raise the point that cultural dimensions which measure values related to human-to-human interaction may not be relevant to interaction that is, in fact,

15

human-to-interface. To support their argument they identified ten cultural dimensions which vary between machine types as well as between cultures.

Unfortunately, later, similar research by Schwartz on cultural dimensions may have repeated the methodological errors made by Hofstede. Although Schwartz’s model of culture was based on more recent research over a larger sample than

Hofstede’s, with research conducted students in 67 countries in the period 1988-2000, Schwartz restricted his survey to school teachers and college, thus making it hard to say his surveys represented a cross-section of society and contexts.

On the other hand, research conducted by the World Values Society (WVS) (http://www.worldvaluessurvey.org/) shows greater methodological rigor than both Schwartz’s and Hofstede’s studies. The WVS conducted research into global values in five waves of surveys between 1981 and 2008, with the latest survey covering 64 countries. Participants were chosen randomly, once quotas were decided, based on the demographic features of that country. National sub-cultures were also investigated.

The greater spread of WVS survey suggests that the cultural values identified by the WVS more accurately represent the culture of the country surveyed than those identified by Hofstede. The World Values Survey identified two axes of variance in global values – the need for survival vs. the need for self-expression, and traditional vs.

secular-rational values. Societies with low incomes cannot take survival for granted and values in such societies are based on the need for economic and physical security.

On the other hand, countries with sufficient income can afford to focus on other needs such as personal well-being, an improved quality of life, and the need for

self-expression. Traditional values are strongly influenced by religion and include such values as the importance of traditional family values and national pride. Secular-rational values are based less on traditional religions and attitudes, and are more liberal on social issues such as divorce or euthanasia.

Given the differences in methodology between Hofstede, Schwartz, and the WVS, it is interesting to note that their results support each other. Schwartz’s (2004) research into regional groupings of cultural values shows that regions identified by the WVS, Hofstede and Schwartz’s own research are similar, and that the values identified by each study, while differing in nature, show some correlation with each other. This point supports the reliability of cultural dimension models such as those developed by Hofstede, Schwartz, and the WVS. Research conducted by the WVS showed the greatest methodological accuracy, yet arrived at similar conclusions to Hofstede’s.

Therefore, it appears reasonable to select cultural dimensions developed by WVS for study into the effect of culture on usability. In addition, the widespread use of

Hofstede’s model of culture allows comparison between studies, thus providing two models of culture for use in this study.