Page 49 - Cultures and Organizations

34 THE CONCEPT OF CULTURE

Replications of the IBM Research

In the 1970s, while IBM survey data continued to come in, Geert admin-
istered some of the same questions to an international population of non-
IBM managers. These people, who came from different companies in fi fteen
countries, attended courses at a business school in Switzerland where Geert
8
was a visiting lecturer. At that time, he did not yet have a clear concept
of dimensions in the data, but the replication showed that on a key ques-
tion about power (later part of the power distance dimension), the countries
ranked almost exactly the same as in IBM. Other questions indicated coun-
try differences in what we now call individualism versus collectivism, again

very similar to those in IBM. This was Geert’s first indication that the
country differences found inside IBM existed elsewhere as well.
In later years many people administered the IBM questionnaire—or
parts of it, or its later, improved versions, called Values Survey Modules
9
(VSMs)—to other groups of respondents. The usefulness of replications
increases with the number of countries included. The more countries, the
easier it becomes to use statistical tests for verifying the degree of similar-
ity in the results. As of this writing, along with many smaller studies, we
count six major replication studies, each covering fourteen or more coun-
tries from the IBM database. Those six are listed in Table 2.1.
Four of the six replications in Table 2.1 confirm only three out of the

four dimensions—and each time the one missing is different. For example,
data obtained from consumers did not replicate the power distance dimen-
sion. We assume this is because the respondents included people in differ-
ent jobs with different relationships to power or people without paid jobs
at all, such as students and housewives.
Most of the smaller studies compared two or three countries at a time.
It would seem too idealistic to expect confirmation of the IBM results in

all these cases, but a review of nineteen small replications by the Danish
researcher Mikael Søndergaard found that together they did statistically

confirm all four dimensions. The strongest confi rmation was for individu-
10
alism. Most small replications start from the United States, which in the
IBM studies was the highest scorer on individualism, and any comparison
with the United States is likely to show a clear individualism difference.
The success of the replications does not necessarily mean that the coun-
tries’ cultures did not change since the IBM research, but if they changed,
they changed together, so that their relative position remained intact.