Dimension data matrix

Dimension data matrix

  • Here are, in downloadable .xls, .doc, .csv and .sav (SPSS) format, the base culture data for six dimensions of culture as presented in Cultures and Organizations 3rd edition 2010. Researchers can use them without asking for permission. Those who are considering commercial use should contact us.

Researchers wishing to work with our data are strongly advised to read Culture's Consequences 2nd ed 2001. The research and VSM section of this web site also contains some advice.

These are the data as used in our books (version 2015 12 08):

Some of the dimension scores obtained in replication studies fall outside the 0-100 continuum. Researchers who cannot live with that range can use modified data that we prepared.

For those that prefer to work with scales of 0 to 100, we have in 2014 brought some outlying values, obtained in replication studies, within the 0-100 range. they are: PDI of Malaysia and Slovakia104 to 100; MAS of Slovakia 110 to 100; UAI of Greece 112 to 100, of Portugal 104 to 99, of Guatemala 101 to 98, of Uruguay 100 to 98. The resulting files (version 2015):

As of 2012 we also keep track of seriously researched VSM scores that are brought to our attention and that are plausible but cannot be integrated into the wider model with confidence.

This could be, for example, because insufficient comparison is possible with reference countries that are already in the Hofstede dimension data matrix. Often these are also studies with a very limited number of countries. So far we have Nepal and Sri Lanka here:

Dimensions do not exist

Geert likes to say that 'dimensions do not exist'. They are a product of our imagination, used for understanding.

Hofstede's dimensions were not postulated but found inductively. Each new study uses new respondent sets and different countries. Even if it used the same questions, these questions might have come to mean different things. So we should take dimension scores with a grain of salt. Actually it is rare for different studies on different data sets to yield the same dimensions.