Work package 6: Construct validation using the European Social Survey
A thorough test of ESeC requires use of a dataset offering the widest possible coverage of EU member states and new member states and with a broad range of outcome and criterion variables relevant to (a) class variation and (b) issues concerning the knowledge based society. The new European Social Survey (ESS) would appear to serve all these purposes. In addition it contains all the operational variables necessary to the construction of ESeC.
ESS is a nineteen country annually repeated cross-sectional survey with both core and rotating questionnaire modules. Areas of the core questionnaire relevant to ESeC construct validation include employment, education, health, family background and social exclusion. It covers all the current EU member states except Austria, Belgium and France. It also covers several new member states: the Czech Republic, Hungary, Poland and Slovenia.
Hypotheses and analysis
Using ESS, UESSEX-ISER will undertake a similar type of analysis and with similar hypotheses to those proposed by EUR and UNIMIB, examining ESeC in relation to earnings, unemployment, social exclusion and limited long-standing illness (LLI). In all these areas we would hypothesise that class varies montonically with outcome variables so that the risks of low earnings, unemployment experience, feeling excluded and suffering LLI increase moving from higher to lower classes. It may be possible to construct a household version of ESeC from ESS data so that this may also be tested against outcomes.
Finally, given the range of employment variables, it might also be possible to undertake some limited forms of criterion variable analyses of ESS data. The ESS contains at least two variables that appear to measure aspects of employment contracts.
Analyses will be undertaken using ordinary least squares and logistic regression techniques.
In addition, and with the aid of WARWICK, this work package will also include a replication of research on estimates of the family gap in pay which used ECHP data (see Davies et. al. 2003). Here, however, we shall use BHPS and SOEP data, each of which will allow the construction of ESeC without any difficulties.
- Step 4: UNIMANN-MZES: Work package 5: ESeC and Education
- 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: SOFI: Work package 9: ESeC and EGP compared
- Step 4: UNIMIB: Work Package 10: Unemployment and Earnings Inequalities
- Step 4: ESRI: Work Package 11: Poverty and the Deprivation