200 likes | 313 Views
Validation studies : projects using French data. Assessing the consistency of ESEC with theoretical framework à la Goldthorpe Assessing the explanatory power of ESEC on various outcome variables. Assessing the consistency of ESEC.
E N D
Validation studies : projects using French data • Assessing the consistency of ESEC with theoretical framework à la Goldthorpe • Assessing the explanatory power of ESEC on various outcome variables
Assessing the consistency of ESEC • The data : 1998 working conditions survey (supplementing the LFS) => Information available on autonomy, routine, authority over employees, wage, tenure, short term contract (details below) • Compare various definitions of ESEC … => based on French PCS (first definition of ESEC) => based on beta-matrix => based on matrix V2 • … In terms of explanatory power => regarding autonomy, routine, authority, wage…
Method of analysis • Carry out analysis of variance for previous variables • Criterion : definition of ESEC performs better, the higher the variance between groups, & the smaller the variance within groups • Run the regression by OLS of variables on ESEC group indicators, and recover the R²
Method of analysis 2 sets of variables : • “global variables” assumed to be relevant for every group of ESEC : autonomy, wage, tenure, short term contract • “local variables” relevant for given pairs of groups : routine, authority => assess the ability of a given definition of ESEC to separate particular pairs
Autonomy variables (employees) • “You carry out instructions strictly” vs “Sometimes you do otherwise” & “Most of the time you do otherwise” & “Not relevant” • “You are told what to do & how to do it” vs “You are told what to do but you decide how to do it”
Wage variable (employees) • Monthly earnings + yearly bonus • Lack of sufficient information on structure of earnings (monthly bonuses) • More information available in other sources than LFS, yet 4-digit occupation not available in these sources • Further analysis using additional sources requires measure of ESEC based on 2-digit occupation
Short term contract • Unlimited duration labour contracts (“contrat à durée indéterminée”) • Vs • Limited duration contract (“contrat à durée déterminée”) • Temporary work (“intérim”) • Subsidized jobs, training schemes, work placement (“emplois aidés”, “stages”)
Routine • “Does your job consist in repeating the same series of operations ?” • => local variable : group 7 vs group 8 in ESEC first definition, groups 6 & 8 vs group 9 in ESEC second definition
Authority • “Do you have one employee or more under your authority” • Local explanatory power : => group 1 vs group 2, group 2 vs group 3, in both ESEC definitions => group 6 vs group 7 & 8, in first ESEC definition => group 7 vs groups 6, 8 & 9, in second ESEC definition
Conclusion • Summary of above results : • most of the time, relatively poor explanatory power • Most of the time ESEC based on ISCO 2 better than ESEC based on ISCO 1 • On autonomy, PCS based definitions slightly better than ISCO 1, close enough to ISCO 2 • On wage, PCS based definitions slightly better • On routine, PCS based definitions slightly better but weak explanatory power • On authority, ISCO based definitions better in upper part of ESEC (1/2, 2/3), PCS based definition better in lower part (6/7 & 8)
Conclusion • Future research on outcome variables : • Testing various ESEC definitions to explain cultural and sportive activities (EPCV 2003, Y Lemel) • To be confirmed : health, consumption • Important restriction : requires definition of ESEC built on 2-digit PCS (Catégories socioprofessionnelles)