Page 71 - Cultures and Organizations

56 DIMENSIONS OF NATIONAL CULTURES

3
this cluster, we selected the three that were most strongly related. From
the mean scores of the standard sample of IBM employees in a country on
these three questions, a power distance index (PDI) for the country was cal-
culated. The formula developed for this purpose uses simple mathematics
(adding or sub tract ing the three scores after multiplying each by a fi xed

number, and finally adding another fixed number). The purpose of the for-

mula was (1) to ensure that each of the three questions would carry equal

weight in arriving at the final index and (2) to get index values ranging
from about 0 for a small-power-distance country to about 100 for a large-
power-distance country. Two countries that were added later score above
100.
The three survey items used for composing the power distance index
were as follows:

1. Answers by nonmanagerial employees to the question “How fre-
quently, in your experience, does the following problem occur:
employees being afraid to express disagreement with their manag-
ers?” (mean score on a 1–5 scale from “very frequently” to “very
seldom”)
2. Subordinates’ perception of the boss’s actual decision-making style
(percentage choosing the description of either an autocratic style or
a paternalistic style, out of four possible styles plus a “none of these”
alternative) 4
3. Subordinates’ preference for their boss’s decision-making style (per-
centage preferring an autocratic or a paternalistic style, or, on the
contrary, a style based on majority vote, but not a consultative style)

Country PDI scores are shown in Table 3.1. For fifty-seven of the coun-
tries or regions (see Table 2.2) the scores were calculated directly from the

IBM data set. The remaining cases were calculated from replications or
based on informed estimates. Because of the way the scores were calcu-
5
lated, they represent relative, not absolute, positions of countries: they are
measures of differ ences only. The scores that were based on answers by IBM
employees paradoxically contain no information about the corporate culture
of IBM: they show only to what extent people from the subsidiary in country
X answered the same questions differently from similar people in country Y.

The conclusion that the score differences reflect different national cultures
is confi rmed by the fact that we found the same differences in populations
outside IBM (the validation process as described in Chapter 2).