Figure 4.1: Results of the Experiment from Chapter
4.2 CULTURAL VALUES FRAMEWORK
Accepting the importance of societal values in influencing a decision maker’s thought processes, I now consider what framework of cultural values should be used. A
threshold consideration must first be discussed. The entire concept of studying national culture has been criticized. Myers and Tan (2002) have criticized the entire enterprise. The concept of “national culture” is problematic they argue. The concept of “nation state” is recent; changes in form and makeup; and is perhaps being made obsolete by
globalization. It is also not true that each nation has its own culture; there are
subcultures to be found within any geographic boundary. The concept of national culture is not that accepted by anthropologists who find that the relationship between national culture values and work-related issues and attitude is not well explained by Hofstede’s model. Baskerville (2003) places similar criticisms in the context of Hofstede’s work. She contends that Hofstede’s definition of culture is outmoded and therefore his work has had little uptake in the fields of sociology or anthropology and is not well cited in those fields. Similar to Myers and Tan, she criticizes the concept of national culture. Finally, she argues that the concept of using scales and outside observer status has been rejected by the anthropological discipline and therefore is questionable. Hofstede (2003) in a response to Baskerville, defending this approach, argues that his work has been cited widely in organizational sociological literature, but has not been cited in the anthropological literature due to little interest in that field in business related activities and general antipathy to the business management field. This led to an ignorance of his work in the social anthropology field. In terms of the national culture argument, he
agrees that the national level is perhaps not the best level of analysis for culture and that nation states cannot be identified with national culture, but he finds that empirically his scales correlate well with data from other sources. Baskerville, however, remains
unconvinced and argues that further examination of the replications that Hofstede claims as validating his scales needs to be done (Baskerville-Morley 2005). Given that I am undertaking a replication of a positivist study and that Hofstede style scales have been replicated, I will adopt the use of this type of methodology for this study. The question then is which set of scales to use.
The most prominent cultural values framework is that proposed by Hofstede (2001). Based on a factorial analysis of internal survey results of IBM employees in 50 countries
conducted in the late 1960s and early 1970s, he derived four cultural values: Individualism, Power Distance, Uncertainty Avoidance, and Masculinity. Later, after additional research in Asian cultures, he added another dimension: Confucian
Dynamism or Long-Term Orientation. The Hofstede measures have been extensively used within the IS literature including a study of IS project escalation (Keil et al., 2000) and a study of the Mum Effect (Tan, et al. 2003). However, the Hofstede measures have been criticized as lacking construct validity and other psychometric properties (Spector, Cooper and Sparks 2001), and as being US/European or IBM centric (Baskerville 2003; McSweeney 2002; Myers and Tan 2002). They have also been criticized as atheoretical (House, et al. 2004). Also, the items used in the Hofstede measures mix organizational and societal values with values of the individual and the collective. Since, in this study, I am interested in investigating the effect of societal cultural values on the occurrence of the Deaf Effect, Hofstede’s framework; with its items mixing organizational and societal levels does not meet these requirements.
Recently, the Global Leadership and Organizational Behavior Effectiveness Research Program (GLOBE) collaboration (House, et al. 2004) has proposed a framework of nine cultural values that overcomes most of the criticisms of the Hofstede framework. This framework appears to improve on that of Hofstede by, first, being theoretically based versus simply being empirically derived. The GLOBE framework is based on a
conceptual model that reflects an integration of implicit leadership theory, value-belief theory of culture, implicit motivation theory and structural contingency theory of
organizational form. Second, it seeks to eliminate societal and organizational centrism by being developed by an international team of 170 researchers in 62 different countries with middle manager subjects in those countries within 951 organizations in three
different industries. Thirdly, they employed a rigorous validation process that conforms to the prescription of Karahanna, Evaristo and Srite (2002) for establishing cross-cultural equivalence of measures. To avoid construct bias, Karahanna, et al. first suggest informants in each culture “should be asked to describe the construct and associated behaviors” (p. 49). For each construct, items were derived through interviews and focus groups in several countries (Hanges and Dickson 2004, p. 124). These items were first reviewed for appropriateness through the use of Q-sorting, item evaluation, and
translation and back translation by U.S. PhD students and then by Country Co-
Investigators (CCI) in 38 countries. Item evaluation reports were provided by the CCIs to inform GLOBE as to any words or phrases that were ambiguous or could not be
adequately translated. Problematic items were rewritten or dropped. Secondly,
Karahanna, et al. suggest that factor analysis and multi-dimensional scaling be used to assess internal structural congruence. In two pilot tests involving results from 43
countries, the GLOBE investigators applied exploratory factor, reliability and aggregation (rwg(J), ICC (1) and (2), Muthen multilevel confirmatory factor) analyses to assess the validity of the items. To suppress item bias, the GLOBE researchers used the qualitative methods described above to eliminate poor translation, and complex or inappropriate wording. Fourth, the GLOBE framework presents four scales, two for measuring societal values, two for organizational values. Within these, they have a scale for measuring existing cultural values and one for measuring how the respondents believe the values should be. Thus they remove the concern of cross measurement of society and
organization and how the society is now versus how it should be.
I therefore believe that these measures represent an improvement over the Hofstede measures and so propose to use them in this study to measure cultural attributes.
4.2.1 Selection of the Cultural Values Scale
As noted above, GLOBE provides four different instruments that measure the constructs at two different levels, at the societal/organizational level and at the practices (as-is)/ values (should-be) level. The societal measures ask respondents to rate the society in which they live as to the various cultural values. Similarly, the organizational scale requests them to evaluate their organizations. The practices scales, of which there is one for the society and one for the organization, ask respondents to rate a set of
practices relating to the surrounding society or organization, as it exists now. The values scales, which also exist in a form for the society and one for the organization, requests the respondents to assess what the practices ‘should be’ which gives a view of societal values. In this study, I will use the societal practices scales to measure culture, as I am interested in assessing practices (as-is) at the societal level. In this scale the
respondents are not providing their own values, but acting as individual raters to provide their perception of the society’s practices. This measures precisely what is needed in this study.