90 likes | 101 Views
This document discusses the current state of wave implementation in the French LFS, including the rotation pattern, data collection methods, and questionnaire complexity. It also explores the implementation of the AHM module and IT challenges faced during wave 6. Additionally, it addresses specific weightings used in the LFS and the need for changes due to new regulations regarding structural variables and AHM. Contact Insee for more information.
E N D
Wave implementation in the French LFS: present state, questions, future needs for change due to new regulation LAMAS, June 2015 meeting
Main characteristics of the current French LFS • The rotation pattern : • Dwellings are interviewed 6 consecutive quarters • Sample rotated 1/6th quarterly • 5/6th overlap Q-to-Q, 1/3 overlap Y-to-Y • The LFS sample is made up of clusters • Composed by ~20 nearby main residences • Each cluster is assigned to the same reference week (and same interviewer) during the 6 quarters • Clusters replaced by adjacent clusters to minimize variance of evolutions • Data collection • CAPI (1st, 6th wave), CATI (2nd-5th wave) (~ CAPI by phone) • Fixed reference weeks • A short collection period (2 weeks and 2 days) • Questionnaire • A rather complex questionnaire (16 mn 1st ; 4-5 mn 2nd-5th ; 6 mn 6th) Wave implementation in the French LFS – Lamas meeting
Wave implementation • EU Structural variables • Collected in 1st wave • Excepted some variables collected both in 1st and 6th wave • Working conditions (night, evening, we …), wages (INCDECIL)… • Excepted some structural variables collected each quarter • Asked every quarter as REGISTER…, or • Updated each quarter (EDUCFIELD, HATYEAR, HATFIELD..), or • Incremented each quarter • Main status (MAINSTAT) asked for every month for the former 12 months in 1st wave ; for the former 3 months in other waves • Participation in training during the 12 past months available because asked for the 3 past months at each wave (longitudinal weighting is necessary) • French Structural variables • About health/disability (waves 1 & 6) • About the parents (social origin) (wave 1) Wave implementation in the French LFS – Lamas meeting
Wave implementation : AHM • AHM • Collected in 6th wave • Without proxy • Addition of supplementary french questions until 2014 • Why? • Continuity with past (annual survey, complementary survey in the 3rd and last wave) • Risk of jeopardising LFS response rate • Length of the whole interview (LFS + AHM) • The duration of the 1st interview (CAPI) is much longer as the 6th interview (because of dependent interview) • Choice of non proxy interview to enhance quality of answers Wave implementation in the French LFS – Lamas meeting
AHM : wave implementation and IT questions • 2 (Blaise) datamodels • 1 for the LFS • 1 for the AHM • Because of complexity of the LFS (filters, dependent interviews…) • Because of changes every year of the AHM • It would be very risky or demanding a lot of work in France to scatter the AHM sub-modules in the LFS modules • Some difficulties due to the implementation in the 6th wave • When AHM questions make reference to LFS questions not asked/coded during the interview • Example : AHM 2016 • HATLEVEL : we encountered 2 difficulties • - updated question : need to get back the former answer in order to build the LFS variable question (XQ) needed to filter the AHM interview • get the answer from Q-1, …,Q-5 (must be known very early), or … ask again • - codification : after the data collection and not during the data collection Wave implementation in the French LFS – Lamas meeting
AHM : wave implementation and organisation of tests • PAPI test • November Y-2 • With interviewers : 100-200 interviews • CAPI test • June Y-1 • With interviewers • As a 7th wave for dwellings interviewed a few months before with the LFS : LFS 6th wave interrogation and AHM • Final rehearsal • November Y-1: check the whole IT process LFS+AHM Wave implementation in the French LFS – Lamas meeting
Wave implementation : specific weightings • Several weights LFS • Q weights 1 to 6 • Y weights for wave 1 (in France, also Y weights for waves 1 and 6) • Consistent with quarterly results (Sex, Age, Acteu) • Consistent with mean annual results (CS, diploma) • Method • Cf. Spain presentation at the 9th WS on LFS Methodology in Roma, May 2014 • Cf. methodological presentation : • http://jms.insee.fr/files/documents/2015/S07_4_ACTE_V1_OURADOU_JMS2015.PDF • Specific weight AHM (wave 6) • Not consistent with mean annual results Wave implementation in the French LFS – Lamas meeting
The new regulation • Structural variables, AHM • Need to change according to new regulation • AHM linked to structural variables (ex : JOBDET block) • AHM in 1st or 2nd wave ? • Same kinds of technical difficulties for 2nd wave as now in 6th wave • Proxy interviews, CATI (CAWI)? • No more (or very limited) supplementary questions : less burden Wave implementation in the French LFS – Lamas meeting
Insee 18 bd Adolphe-Pinard 75675 Paris Cedex 14 www.insee.fr Informations statistiques : www.insee.fr / Contacter l’Insee 09 72 72 4000 (coût d’un appel local) du lundi au vendredi de 9h00 à 17h00 Thank you for your attention ! Contacts : Sylvie Le Minez, Nicole Roth (Insee) Courriels : sylvie.le-minez@insee.fr; nicole.roth@insee.fr