Work package 9: ESeC and EGP compared
In European social science the most commonly used social class schema at present is the EGP (Erikson-Goldthorpe-Portocarero) class schema. The new UK National Statistics Socio-economic Classification (NS-SEC) is to a considerable extent based on the same principles as EGP and so is the proposed ESeC. In order for European researchers to trust and use ESeC they must be assured that ESeC is as good a social class indicator as EGP. There are therefore sound reasons to try to establish both the relation between EGP and ESeC and to find out to what extent ESeC has a similar relation to various criteria as EGP.
Hypotheses, data and analyses
A validation should thus be done in two ways. If both EGP and ESeC are coded on the same data set it will be possible to discover on the one hand the extent to which they overlap and on the other where there are cases that are treated differently by the two approaches. Such a study will highlight the differences between the two classification principles, but also indicate where the coding procedures for one or the other are less than acceptable.
The second form of validation includes a study of the discriminatory power of the two indicators of social class. While it is a clear mistake to try to find the indicator that ‘explains’ most of some measure of, say, people’s life chances (such an indicator would typically be a composite measure of several variables and could in the end hardly be used for any other purpose than to show that there are differences between groups of people), the use of an indicator that does not catch the basic differences between classes or other groups would lead to an underestimation of the degree of stratification in society. The ‘classless society’ is one where there are no differences between classes, but if no differences were found, it seems more sensible to believe that there is something peculiar with the measurement of class rather than that Utopia has arrived. Thus, measures of stratification should have a clear conceptual base, but given that, one ought to strive for high discriminatory power. Therefore, if one of two indicators of social class, like EGP and ESeC, which are both founded on the same theoretical base, clearly has less discriminatory power than the other, the second one is clearly to be preferred. If the discriminatory power is about the same, other criteria like ease of coding could determine which indicator to choose.
Therefore, it seems essential to code both EGP and ESeC to some data sets in order (a) to cross tabulate the two indicators against each other and (b) to find out their relation to various criteria, both attitudinal and behavioural. It is not possible at the outset to determine which data sets to use, but we should certainly try to code both indicators to the data of the European Social Survey, since this would make it possible to find out about the relation between the two approaches to class assignment in a large number of European countries. This would also serve to complement the UESSEX-ISER validation study. It would also be of great value if the occupational information in the European Community Household Panel could be coded according to both schemes. However, at the outset it is not clear to what extent the information available for these two surveys is precise enough to allow a complete coding of both class schemas. Therefore, both schemas should be coded to some national data sets like the British Household Panel Survey, the German Mikrozensus and the Swedish Level of Living Survey, since both the occupational information in these surveys is typically of very high quality and a large number of suitable criterion variables are available in these studies.
- Step 4: UNIMANN-MZES: Work package 5: ESeC and Education
- Step 4: UESSEX-ISER: Work package 6: Construct validation using the European Social Survey
- Step 4: INSEE: Work package 7: Criterion validation of ESeC using the ECT
- Step 4: EUR: Work package 8: Construct validation of the ESeC in relation to health inequalities
- Step 4: UNIMIB: Work Package 10: Unemployment and Earnings Inequalities
- Step 4: ESRI: Work Package 11: Poverty and the Deprivation