The indicators for describing cultural dimensions came from two databases: the European Social Survey (ESS) and the European Values Study (EVS), which include questions pertaining to all four cultural dimensions considered here. The EVS is a multi-country survey covering almost all European countries that is repeated every nine years (1981, 1990, 1999 and 2008). Here, the fourth wave for year 2008 is used. The ESS is a new biennial multi-country survey covering an increasing number of European countries. The first round was conducted in 2002 and this article uses the fourth wave from year 2008.

Both surveys include questions that can be used to describe cultural dimensions. Some aspects are covered in both surveys (general trust or giving men priority, for example), but mostly the questions from the two surveys complement each other. Hence, combining data from these two surveys gives a much better coverage of cultural dimensions than either does alone. Also, the two surveys are similar in the number of respondents. The ESS usually includes 1,500 to 2,500 respondents per country. The EVS usually includes 1,000 to 2,000 respondents per country. Also, both surveys provide weights that we use in the present study to ensure that the regional-level data represent the demographic structure of a region.

In Europe, the NUTS – Nomenclature of Territorial Units for Statistics – is widely used not only for statistical purposes, but for a wide range of administrative and policy decisions by the European Union. Established by Eurostat, it is a hierarchical classification of countries and regions within countries. The classification subdivides each country (NUTS0 level) into one or more NUTS1 regions, each of which is, in turn, subdivided into one or more NUTS2 regions and so on. The NUTS0 level coincides with the country level. At the NUTS1 and NUTS2 levels, countries are divided into regions based partly on lower and upper limits for the population size for each level. Apart from population size, the classification favours a state's administrative divisions in forming regions. Therefore, some smaller countries are represented as whole countries at the NUTS1 or even NUTS2 level as well, whereas some smaller countries like Switzerland are represented by many regions.

Both the ESS and EVS include respondents from all NUTS1 and NUTS2 regions of a particular country. Although our intention was to include all countries at the NUTS2 level, the ESS data were only available at the NUTS1 level for Belgium, France, Germany and the United Kingdom. For Norway the EVS, and for Russia both databases provided data only at the first regional level (comparable to the NUTS1 level). Thus, these countries had to be included in the analysis at this level, and therefore the level of deepest division is here and hereafter referred to as the NUTS2(1) level analysis. The average number of respondents per region at the NUTS1 level was 443.28 and 245.93 at the NUTS2(1) level. In the ESS the average number of respondents per region was 545.74 at the NUTS1 level and 303.21 at the NUTS2(1) level.

The initial regional-level and country-level indicators were obtained by aggregating individual-level data using the weights provided by the ESS and EVS.

More information can be found in the article.