VALUES SURVEY MODULE 1994
by Geert Hofstede
MAY BE FREELY USED FOR RESEARCH PURPOSES
FOR REPRODUCTION IN COMMERCIAL PUBLICATIONS,
PERMISSION IS NEEDED
Copyright ( Geert Hofstede BV
How to use and not to use the VSM 94 3
1. Formulas for index calculation 4
History of the VSM 94 7
2. Comparison of the VSM 94 with the VSM 82 8 3. A note on translations10
1. How to use and not to use the VSM 94
The Values Survey Module 1994 (VSM 94) is a 26-item questionnaire developed for comparing culturally determined values of people from two or more countries or regions.
It allows scores to be computed on five dimensions of national or regional culture, on the basis of four questions per dimension: for this, it needs 5 x 4 = 20 questions. The remaining six questions are demographic; they ask for the respondent’s gender, age, education level, kind of job, present nationality, and nationality at birth.
Experience has shown that the answers to the 20 content questions vary substantially between nationalities. This is not to say that every respondent of one nationality gives one answer and everyone of another nationality gives another answer, but on average, a sample of respondents of nationality A will (nearly) always score higher, or always score lower, than a comparable sample of people of nationality B (in statistical terms, an analysis of variance shows a significant country effect).
However, answers to the 20 content questions will also be influenced by other characteristics of the respondents, such as gender, age, level of education, occupation, kind of work, and year that the survey was held. Therefore, comparisons of countries or regions should in as far as possible be based on samples of respondents who are matched on all criteria other than nationality or region. They should be matched on any criterion (other than nationality) that can be expected to affect the answers.
The 5 x 4 content questions were selected because, when matched samples from different countries are compared, the mean scores for the countries on the four questions belonging to the same dimension usually vary together (if one is high, the other is high, or low if it is a negatively formulated question; if one is low, the other is low, etc.). In statistical terms, the country mean scores are strongly correlated. The mean scores for the countries on questions belonging to different dimensions usually do not vary together (are uncorrelated). Therefore, the 20 questions form 5 clusters of 4 questions each. The five clusters stand for the five dimensions of national culture identified in research by Hofstede and Bond.
When samples of respondents of the same nationality but with different occupations or different employers were compared (matched on criteria other than occupation or employer), the same dimensions were not found. Nor were they found when the answers of individual respondents were compared. The answers to most of the questions do vary somewhat from one occupation to another and sometimes from one employer to another, and the answers to all of the questions vary from one individual to another. However, the mean scores on the questions for different occupations or for different employers will not form the same five clusters. Nor will scores from individual respondents (instead of mean scores for groups of respondents of the same nationality) form the same clusters.
If the questionnaire is used to compare responses from individuals, from respondents with different occupations or employers, or from respondents belonging to any category other than nations or regions, the answers should be...