« LFS series breaks with the adoption of the IESS FR How is Statistics Portugal planning to tackle the issue? 13th Workshop on Labour Force Survey Methodology.

Slides:



Advertisements
Similar presentations
United Nations Workshop on Revision 3 of Principles and recommendations for Population and Housing Censuses and Census Evaluation Amman, Jordan, 19 – 23.
Advertisements

Eurostat Data collection. Presented by Johan Erikson Statistics Sweden.
National design, fieldwork and data harmonization for Labour Force Survey Irena Svetin Statistical Office of the Republic of Slovenia September 2014.
Centraal Bureau voor de Statistiek Challenges of redesigning household surveys and maintaining output quality Menno Cuppen Paul van der Laan Wim van Nunspeet.
1 South Africa Design and Implementation of Labour Force Surveys Yandiswa Mpetsheni South Africa.
1 Web interviewing at Statistics Netherlands: practices so far Labour Force Survey The Netherlands Björn Janssen Hendrika Lautenbach April 2010.
Planning, preparation and conducting TQS in Tajikistan Agency on statistics under the President of Tajikistan.
Quantitative Methods for Business Studies
Pali Lehohla Statistician-General.
Data and information on disability and employment
Multi-mode data collection
Confidentiality in Published Statistical Tables
Design and Implementation of Labour Force Surveys
SEM II : Marketing Research
Annual labour force surveys
Weighting issues in EU-LFS
The usage of web interviewing in Lithuanian Labour Force Survey
ESCB experiences with breaks in labour market time series
The second wave of the new design of the Dutch EU-SILC: Possibilities and challenges Judit Arends.
LIVESTOCK PRODUCTION AND PRODUCTIVITY
Regression composite estimation for the Finnish LFS from a practical perspective Riku Salonen.
Linking Population and Housing Censuses with Agricultural Censuses
LAMAS Working Group 7-8 December 2016
Generic Statistical Business Process Model (GSBPM)
The European Statistical Training Programme (ESTP)
LFS Workshop Reykjavik, May 2018
Working Group on Labour Statistics for MEDSTAT countries October 2013
Annual labour force surveys
Estimating breaks in time series in the Austrian LFS
13th Workshop on Labour Force Survey Methodology
Chapter 14: Mixed-mode datacollection
Redesigning the LFS at Statistics Netherlands
The impact of using web in The Danish LFS
LAMAS Working Group 7-8 December 2015
Hungarian practice on chain-linking and its implication for SA
Assessing Quality of Paradata to Better Understand the Data Collection Process for CAPI Social Surveys François Laflamme Milana Karaganis European Conference.
Production of break free time series for the Italian LFS
Effects of attrition on longitudinal EU-LFS estimates
Estimation of Employment for Cities, Towns and Rural Districts
National needs for AES Purpose - describe participation in learning during a 12 months period. The main parameters are; Participation rates in different.
LAMAS Working Group 29 June-1 July 2016
LAMAS Working Group 29 June-1 July 2016
Albania 2021 Population and Housing Census - Plans
Passenger Mobility Statistics 2017
LAMAS Working Group 29 June-1 July 2016
LAMAS Working Group 29 June-1 July 2016
Debriefing from the December 2017 LAMAS meeting Item 4
Meeting Of The European Directors of Social Statistics
The change of data sources in the Spanish SILC
Point 5 : Progress of the work on the EHS since the last DSS meeting
LAMAS Working Group June 2017
Policy needs for rural development statistics and data analysis
LAMAS Working Group 6-7 December 2017
The modules of the EU Labour Force Survey
Implementing mixed mode questionnaire in FI-SILC
Agenda item 5.2 Methodology
LAMAS Working Group 6-7 December 2017
The Swedish survey on turnover in the service sector
DIAGNOSTIC FRAMEWORK: National Accounts and Supporting Statistics
13th LFS Workshop on Methodology Reykjavík 17th and 18th of May 2018
Mode effects in mixed-mode data collection WP2
LAMAS Working Group June 2018
Deciding the mixed-mode design WP1
Meeting of the Directors of Social Statistics February 2016
Item 5 Modernisation of the EU-SILC Production
ՀՀ ԱԶԳԱՅԻՆ ՎԻՃԱԿԱԳՐԱԿԱՆ ԾԱՌԱՅՈՒԹՅՈՒՆ
Item 4.1: Annual labour market flows
Meeting Of The European Directors of Social Statistics
LAMAS Working Group June 2019
Workshop on best practices for EU-SILC revision, −
Presentation transcript:

« LFS series breaks with the adoption of the IESS FR How is Statistics Portugal planning to tackle the issue? 13th Workshop on Labour Force Survey Methodology Session 3: Series Breaks « Sónia Torres (sonia.torres@ine.pt) Statistics Portugal / Labour Market Statistics Unit Reykjavik, 18th May, 2018 «

« « The LFS transition into the new regime in Portugal 1. Changes to be introduced in 2021 Changes directly implied by the IESS FR adoption (all countries): New questionnaire: variables design & frequency (e.g. redefinition of current variables; core/standardized variables; new variables; regular modules) + flowcharts and model questions for the labour status measurement. New concepts: although the main concepts (E and U) do not change in substance, the 19th ICLS Resolution provided more clear specifications (e.g. coverage; inclusions/exclusions). Expected sizeable change in Portugal, as the subsistence agriculture will be no longer considered as employment. In 2017: 3.4% of total employment; 53% of employment in section A. «

« « The LFS transition into the new regime in Portugal 1. Changes to be introduced in 2021 (cont.) Statistics Portugal will take the opportunity to introduce other changes (to reduce the respondents’ burden and to concentrate all changes at once): Wave approach: large questionnaire (all variables) in the 1st interview (CAPI) + simple questionnaire (quarterly variables) in the 2nd-6th interviews (CAWI or CATI), under certain conditions. Dependent interviewing (reinforcement): from a dozen of demographic variables to a larger list of non-demographic variables (based on testing) + simplified rules for ind. 75+. Web interviewing (alongside with CAPI and CATI): in the 2nd-6th interviews. In the 1st interview (CAPI), the interviewer “teaches” how to fill in the LFS questionnaire and informs about its smaller size in the following quarters. «

« « The LFS transition into the new regime in Portugal 2. Statistics Portugal strategy to tackle the series break All changes will likely translate into a series break in the Portuguese LFS estimates (like it did in 2011). Need to provide users with (at least for the main indicators): Linking factors (to understand the break). Back-calculated series (for short-term analyses). Also required by Eurostat for 14 indicators (E and U by sex and age group) + other indicators on a voluntary basis, until December 2021. Up to a maximum of 1 annual + 4 quarterly factors per indicator. And to publish both in the day of dissemination of the new series results (1Q2021, in May 2021) (to be useful). «

« « The LFS transition into the new regime in Portugal 3. The parallel sample design Set up of a Task Force, which compared several alternatives to compute accurate linking factors (pros and cons of each): Time-series models to make forecasts or nowcasts with external data → Not feasible in Portugal: LFS series patterns not easy to forecast. Administrative data not available for all variables. Risk of transition into the new regime (questionnaire and collection mode) in 1Q2021 with the same sample. Use of a parallel sample, from 2Q2020 to 4Q2020, alongside with the current LFS (core LFS) → Diagram next slide «

« The LFS transition into the new regime in Portugal CAPI: Computer Assisted Personal Interviewing CATI: Computer Assisted Telephone Interviewing WI: Web Interviewing Ii: interview [i = 1 to 6] Qjk: questionnaire [j = 1 (current) or 2 (new); k = L (large: all variables) or S (small: quarterly variables)] «

« « The LFS transition into the new regime in Portugal 3. The parallel sample design: characteristics, advantages and drawbacks New sample (another source of break): the one that will proceed (smoothly) into the new regime in 1Q2021. Big size: total of 12 rotations (5 in CAPI); 2 almost complete quarters → Expensive solution. Start in 2Q2020. Why? To min. costs (vis-a-vis 4Q2019 or 1Q2020). To min. the Rotation Group Bias (vis-a-vis 3Q2020 or 4Q2020). To provide enough time to conclude all upstream tasks (e.g. questionnaire; wave approach; dependent interviewing; implement and test IT tools). «

« « The LFS transition into the new regime in Portugal 3. The parallel sample design: characteristics, advantages and drawbacks (cont.) Linking factors calculation flexibility: as many as needed, by using 2 almost complete quarters (3Q2020 and 4Q2020). No impact in the LFS core results during the transition period. Independence of the LFS core sample (respondents must also have never replied to LFS questionnaires before). Data collection modes (CAPI/CATI/CAWI), rules for proxy, imputation rules and weighting = LFS under the IESS FR (1Q2021). Hard to collect in 2020 (Census 2021 fieldwork preparation). Implications for the MUR dissemination in 5 months. «

« « The LFS transition into the new regime in Portugal 4. The linking factors calculation By differences between the estimates obtained from the parallel sample and the estimates obtained from the LFS core sample → Justified by the similar characteristics of both samples, except for the changes to be introduced. For 2 quarters (3Q2020 and 4Q2020), then averaged. 1 factor per indicator (variable/breakdown), to avoid complexities in linking the series & interpreting the results and the need to run a parallel sample for at least 4 quarters. The same factor will be applied to all quarters and years in the back-calculation exercise. «

« « The LFS transition into the new regime in Portugal 5. Communication strategy 1Q2021 results to be published in May 2021, alongside with: The linking factors for a selection of indicators. The back-calculated series for a selection of indicators. A methodological paper describing both the changes introduced in the LFS and the methodology followed to calculate the factors. A briefing with journalists (to be evaluated). Meetings with main LFS users (e.g. Ministry of Employment; Ministry of Economy; Bank of Portugal; National Accounts Department; Statistics Council) during the transition process. Advance information note (in December 2020) for general LFS users. «

« « The 2011 LFS series break in Portugal 6. Notes regarding the 2011 LFS series break In May 2011, Statistics Portugal released the results of a new LFS series, starting in 1Q2011 (break with the 1Q1998-4Q2010 series). Main change: mixed mode data collection (from CAPI to CAPI+CATI) → Deep revision in the questionnaire. Purposes of the change: sustain the decrease in response rates (e.g. by reducing the interviewing time and by offering a friendlier interviewing mode); reduce costs; improve efficiency (e.g. interviewers/supervisors procedures; data collection management; monitoring tools; HOMECATI). All other characteristics unchanged: sample size and design; representativeness; weighting; rotation pattern; and concepts. «

« « The 2011 LFS series break in Portugal 6. Notes regarding the 2011 LFS series break Long-lasting process, involving: Several Departments at Statistics Portugal and an external consultant. Study visits to 2 countries (Spain and Finland). Parallel sample in place for 1 year (24 rotations): to test and implement the new CATI logistic, to test several effects, and to calculate linking factors, without harming the LFS core. Meetings with the LFS main users. Process and results generally welcomed by users, although too expensive and time-consuming. Impacts: ↓ E, ↑ U, and ↑ UR (+1.1 p.p.). «

The 2011 and 2021 LFS series break compared «