ESS/EVS-based Indicators of Cultural Dimensions

These indicators were created in Kaasa, A.; Vadi, M.; Varblane, U. (2014) Regional Cultural Differences Within European Countries: Evidence from Multi-Country Surveys. Management International Review, 54(6): 825 - 852.

When using material and data on this webpage, please cite the article.

In order to construct latent variables reflecting cultural dimensions, we used a theory-based rather than an empirically-driven approach. The aim was to build cultural dimensions using questions that while different from those used by Hofstede, still conceptually contain aspects of what Hofstede's dimensions represent. The analysis relies on Hofstede’s theory and the choice of indicators was based on Hofstede’s (2001) overview of the characteristics and differences of dimension extremes.

Each of the four cultural dimensions appeared to be described using six initial indicators. The list of initial indicators is here. Some aspects of the cultural dimensions were covered by questions in both databases. In these cases, only one was included in the analysis. The choice was made with the intention to involve questions as equally as possible from both databases

A confirmatory factor analysis of regional-level or country-level indicators was conducted using the principal components method. Confirmatory factor analysis was chosen because the indicators describing a particular latent factor are predetermined based on theoretical considerations taking previous cultural studies into account.

The factor scores of latent variables were saved as variables.

The analysis was repeated on three levels: country level, NUTS1 and NUTS2(1) (see explanation here) level. However, the analyses performed separately for the country level and two regional levels will results in different scales of factor scores. For comparing the country-level factor score with scores of different regions in a particular country, comparable scores had to be obtained. Hence, the same analysis was also performed for the combined sample of all three samples/levels. The resulting factor scores, all now in the same scale, were also saved as variables. This change had no effect on relative sizes of scores at the same level. Hence, these comparable indicators are presented in the dataset.

As the indicators of cultural dimensions are obtained from the confirmatory factor analysis, they are in a standardised form: the mean is equal to 0 and standard deviation equal to 1 for all dimensions.

More information can be found in the article.