When translation is not enough: Background variables in comparative surveys

摘要: Background or socio-demographic variables like age, education, or occupation are often referred to as ‘objective variables’, reflecting ‘hard facts’. Under this assumption, the measurement of these variables should not pose any difficulties in comparative perspective. Of course, it is true that, for example, age can be measured in years, days, or even seconds. However, we are not ‘young’or ‘old’at the same ‘age’in different societies. It is therefore important to realize that those constructs and measures, even if they seem to reflect ‘natural’or ‘objective’states, are always socially constructed and consequently are linked to institutional structures and context-dependent interpretations. With respect to background variables, theorizing is oftentimes limited, resulting in variables not optimal for testing specific hypotheses. For example, education is frequently included in models as a ‘control variable’without considering its substantive meanings. This is even more and Christof Wolf a problem because countries have different systems of education, including for example vocational training paths. Other problems may arise from treating education as a continuous variable, ignoring that, in some cases, its categorical characteristics could be of importance. Another aspect jeopardizing comparability of background variables arises when a category of such a variable is strongly linked to the institutional structure of a given society. An example with respect to occupation is the position of ‘Cadre’in the French context or ‘Professional’in the Anglo-Saxon one. A literal translation of such terms does not produce comparability in the occupations measured. A second case …