Research on national cultures

Hofstede's analysis defined four initial dimensions of national culture that were positioned against analysis of 40 initial countries. As a trained psychologist, he began his analysis of the survey data he had collected at IBM at the individual respondent level. At the end of two years, he realized he needed an "ecological" analysis, in which respondents were contextualized by their countries. By aggregating individuals as societal units, he could examine national cultures rather than individual personalities.

Hofstede's model explaining national cultural differences and their consequences, when introduced in 1980, came at a time when cultural differences between societies had become increasingly relevant for both economic and political reasons. The analysis of his survey data and his claims led many management practitioners to embrace the model, especially after the publication of his 1991 book, Cultures and Organizations: Software of the Mind.

In 1980, Hofstede co-founded and became the first Director for the IRIC, the Institute for Research on Intercultural Cooperation, located at Tilburg University since 1998. Much of Hofstede's research on the basic dimensions of nations came through the IRIC. In 2001, Hofstede published an entirely re-written second edition of Culture's Consequences. In 2010, a third edition of Cultures and Organizations: Software of the Mind was published with Gert Jan Hofstede, Michael H. Bond and Michael Minkov [es] as co-authors. In this book, there were two new dimensions that were added, and the number of countries covered was between 76 and 93. This book also introduced the topic of organizational cultures as a separate and different phenomenon.

Critiques

Despite the popularity of Hofstede's model, some critics have argued that his conceptualization of culture and its impact on people's behavior might be incorrect. The most cited criticism of his work is by Professor Brendan McSweeney (Royal Holloway, University of London and Stockholm University), who argues that Hofstede's claims about the role of national culture indicates too much determinism that might be linked to fundamental flaws in his methodology.^[14] Hofstede replied to this critique,^[15] arguing that the second edition of his book had responded to many of McSweeney's concerns and that he viewed the resistance to his ideas as a sign that he was shifting the prevalent paradigm in cross-cultural studies.^[15] McSweeney has rejected Hofstede's reply, arguing that the same profound methodological flaws that characterize the original analysis of the IBM data remain in the second edition.^[16]

Another key critique, which largely focuses on level of analysis, is by Professor Barry Gerhart (University of Wisconsin-Madison) and Professor Meiyu Fang (National Central University, Taiwan), who point out that among other problems with Hofstede's research (and the way it is widely interpreted) is that his results actually only show that around 2 to 4 percent of variance in individual values is explained by national differences – in other words 96 percent, and perhaps more, is not explained. And that there is nothing in Hofstede's work that pertains to individual-level behaviours or actions.^[17]

In a 2008 article in the Academy of Management's journal, The Academy of Management Review, Galit Ailon deconstructs Culture's Consequences by mirroring it against its own assumptions and logic.^[18] Ailon finds several inconsistencies at the level of both theory and methodology, and cautions against an uncritical reading of Hofstede's cultural dimensions.

Philippe d'Iribarne, director of research at the CNRS (Centre national de la recherche scientifique) in Paris expressed concern that "a theory of culture that considers culture to be 'shared meaning' does not allow for representation of the forms of unity and continuity."^[19] Part of d'Iribarne's objections have been with the weaknesses of Hofstede's terminology in general and category names specifically (e.g., power distance as a culture as whole versus a culture's acceptance of hierarchy only within organizational settings). More pointedly, d'Iribarne questions the generalized conclusions that Hofstede draws from the data, imposing Hofstede's own value system on what the data show. For instance, d'Iribarne questioned Hofstede's conclusions from the uncertainty avoidance statistics, arguing that Hofstede superimposes his own view that data. For d'Iribarne, Hostede simply presumes that showing high stress at work correlates with weak uncertainty avoidance, while d'Iribarne asserts that the presence of high stress could just as readily indicate high stress results from high uncertainty avoidance, since no external control exists in low uncertainty avoidance cultures.^[20] Finally, d'Iribarne questions Hofstede's implicit assumption of uniformity in complex organizations, let alone entire national cultures. Such assumptions of uniformity are useful, d'Iribarne writes only "if one thinks of a culture specific to a close-knit community."^[21] Instead, though, d'Iribarne notes that in most situations, "society is split into more or less antagonistic groups"^[21] and in any case, "meaning is not only received but produced";^[21] in short, Hofstede does not allow for the fact that people do not remain static in how they interact with one another. Philippe d'Iribarne fills out the bare bones of Hofstede's simplified structure, a point with which Hofstede himself acknowledged when he wrote that, "The two approaches are complementary -- mine is more quantitative, d'Iribarne's more qualitative. I provided a skeleton for the countries he studied, and he provided the flesh. The skeleton I proposed is a worldwide structure in cultural differences among cultures."^[22]

Other academics also point to a fundamental flaw in the common application of Hofstede's culture dimensions. Hofstede's culture dimensions and scores are national or "ecological" in nature and do not apply to individual people living in the sampled countries:^[23][24] In Hofstede's analysis, the correlations of his culture variables are significant when aggregated to the national level but not significant at the individual level. This means that no cultural implications can be drawn about individual people living in a certain country; to do so is to commit an “ecological fallacy”. To avoid this fallacy and resulting confusion Brewer and Venaik recommend avoiding the use of the Hofstede dimension scores in management research and training.^[25] The same authors compare the Hofstede culture dimension scores with equivalent dimension scores from the GLOBE culture model^[26] and show severe problems in face, discriminant and convergent validity across the two models.^[27]

In a re-analysis of the cross-national value data, based on Hofstede, Shalom Schwartz and Ronald Inglehart and his own factor analysis of recent World Values Survey data, Arno Tausch [de; es; fr; it; pt; ru; tr] (Corvinus University of Budapest) found however a large-scale confirmation of Hofstede's value scales with other value survey research results. Especially the dimensions Power Distance, Individualism vs. Collectivism, Long-Term Orientation and Indulgence versus Restraint are closely correlated with value dimensions reported by Inglehart, Schwartz and the current data from the World Values Survey.^[28]