The hard-to-survey in EU-SILC Nadja Lamei Statistics Austria, - PowerPoint PPT Presentation

The hard-to-survey in EU-SILC Nadja Lamei Statistics Austria, Challenges and potential solutions for the Directorate Social Statistics Austrian case Summer school 'Reaching out to hard-to- survey groups among the poor' 30 May – 3 June, 2016 www.statistik.at We provide information

Outline • _ EU-SILC : the project, its political context and some important facts of fieldwork • _ The „hard-to-survey“ : who are they and what is so difficult about them? • _ Any solutions?: strategies for different groups www.statistik.at slide 2 | 3 June 2016

EU-SILC European Community Statistics on Income and Living Conditions http://ec.europa.eu/eurostat/web/income-and-living- conditions/overview Folie 3 | 30.03.2016

Objectives Comparable statistical information on income and living conditions Some typical questions at hand: How are household incomes composed and distributed? What about the  income situation of families, singel-parents households, older persons…? For which groups of the population are certain goods and services not  affordable , e.g. going to the doctor‘s, owning a car, having a holiday…? How often are children socially excluded because of their parents‘ economic  situation? How often are social risks passed on between the generations? How much are tenants burdened by their rents ?  How satisfied are persons with their life, what roll does the financial  situation play in this? How much risk of poverty and social exclusion is there really in a wealthy  country like Austria? EU-SILC as data source for (social) politics in MS and the EU! www.statistik.at slide 4 | 3 June 2016

Legal Background  Regulation (EC) No 1177/2003 of the European Parliament and of the Council of 16 June 2003 concerning Community statistics on income and living conditions (EU-SILC)  plus implementing regulations: Definitions Fieldwork and imputation procedures Sampling and tracing rules List of permanent variables Quality reports New material deprivation items from 2016 onwards Yearly Modules Ongoing revision of the European legal documents > Integrated European Social Statistics (from 2019?)  National Regulation of the Federal Minsistry of Labour, Social Affairs and Consumer Protection (ELStV, BGBL. 277/II/2010) www.statistik.at slide 5 | 3 June 2016

Political Background 1999 Treaty of Amsterdam : Social Politics on the EU‘s Agenda 2000 European Councils of Lisbon and Nice: Poverty must me reduced until 2010 2001 European Council of Laeken : Decision on common indicators in the field of social protection and social inclusion 2010 Europe 2020-Strategy for smart, sustainable and inclusive growth: Emphasis on social situation than just economic indicators Targets and indicators: Europe-2020-Targets Fighting poverty and social exclusion in the EU: At least 20 million fewer people in or at risk of poverty and social exclusion Translated into National Action Plans www.statistik.at slide 6 | 3 June 2016

Measuring Poverty in Austria and the EU „Official“ reporting on poverty since 1990ies 1995: AT‘s EU-accession, European Community Household Panel ( ECHP ) 2003: new instrument EU-SILC , continous reporting 2004: start of the integrated (cross-sectional and longitudinal) rotational design, 18 participating countries. Now: 28 EU-MS + Norway, Island, Turkey, Switzerland, Macedonia, Serbia and Montenegro. Further development of concepts : • OECD vs. EU-scale for equivalised income • Risk of poverty > risk of poverty and social exclusion Further development of data and methods : • Survey and administrative data combined • CAPI, CATI and in the future also CAWI • Optimizing field work, sampling, weighting > better representativeness and validity Ever faster, many stakeholders! www.statistik.at slide 7 | 3 June 2016

Latest results Short English content: http://www.statistik.at/web_en/statistics/PeopleSociety/social_statistics/poverty_and_s ocial_inclusion/index.html More detailed German Version: http://www.statistik.at/web_de/statistiken/menschen_und_gesellschaft/soziales/armut _und_soziale_eingliederung/index.html www.statistik.at slide 8 | 3 June 2016

Latest results (2) Risk of poverty or social exclusion Risk of poverty Severely materially deprived Low work intensity www.statistik.at slide 9 | 3 June 2016

Latest results (3) Risk of poverty or social exclusion Q: STATISTIK AUSTRIA/EUROSTAT, EU-SILC 2008-2015. www.statistik.at slide 10 | 3 June 2016

Fieldwork for EU-SILC in Austria – overview Rotational design 2003 Cross sectional survey 2004 Start of integrated rotational design 2005 2006 2007 Beginning of in-house field work 2008 Mixed Mode Design, field work 100% in-house national financing (before: 2/3 Eurostat) 2009 2010 National regulation Some numbers… 2011 Register use (for pension variables only) Min. effective sample size: 4,500 (cross-sectional) 2012 Extensive register use 3,250 (longitudinal) HHs. 2013 Actual sample size reached: ~6,000 HHs net/ year 2014 4 wave panel: ~1,200 HHs 2015 Voluntary participation, all persons aged 16+ are surveyed 2016 2017 New survey infrastructure Proxy rate ~ 10% 2018 Modular questionnaire design and CAWI Since 2012: 85% of total sum of household income from registers 2019 New European Legal Act? www.statistik.at slide 11 | 3 June 2016

First wave  Dwellings /households – see in further detail next slide  Personal computer assisted interviewing (CAPI)  Gross sample 2016 : 3,528 adresses  Expected response rate : 65% www.statistik.at slide 12 | 3 June 2016

Sampling: Selection of first wave sample Type of sampling: one-stage stratified probability sample Sampling units: dwellings registered in the central residence register (ZMR) Stratification criteria: _ Interviewer units (geographical units below NUTS2 level) _Since 2016: also information on Household income from registers (first quartile or above) _Disproportional allocation per NUTS2 level according to expected response rates (based on average response of two preceding year) www.statistik.at slide 13 | 3 June 2016

Follow up waves (2-4)  Persons sample _ Tracing of sample persons (movers, splits) _ All households with at least one sample person take part in the survey (i.e. sample persons and “co-residents” are interviewed) _ Non eligible if moved to institutional household or outside Austria  Computer assisted telephone interviewing (CATI) except for: _ Households with no valid telephone number _ Households explicitly asking for a CAPI (f2f)-interview _ Method changes are possible throughout the fieldwork period  Gross sample 2016: 4,787 households, 2-3% split off-households per year • therof CAPI: 965 therof CATI: 3,822 (80%) •  Expected response rate : 85% www.statistik.at slide 14 | 3 June 2016

Fieldwork  Fieldwork period: February – June/July  Information for households _ Seprate letters for each wave also in Bosnian/Croatian/Serbian or Turkish _ Leaflet on EU-SILC _ First wave: Booklet „Austria in Figures“ _ Follow up waves: EU-SILC newsletter  Website and information video: www.statistik.at/silcinfo  Hotline and e-mail  Information for local administrative units  15 Euro incentive (voucher) for each household upon successfull interview www.statistik.at slide 15 | 3 June 2016

Mode (only follow up waves) planned response rate actual response rate by response rate by assignment mode changes by planned assignment mode changes actual mode to mode mode to mode 98% response rate: 84% 100% response rate: remain CAPI 27% CAPI 83% response rate: 2% changes to CATI response rate: 62% 40% CAPI 81% total follow-up response 100% 100% gross 82% rate follow- sample response rate: 86% remain CATI up waves: 100% 100% response rate: 84% 73% CATI 84% response rate: 60% CATI 86% 18% response rate: 75% changes to CAPI n=4.699 (follow-up) n=3.941 (n=7.936 total cross section) (n=5.909 total cross section) S: Statistics Austria, EU-SILC 2014. *Follow -up w aves excl. 8 households w hich w eren't processed (reported refusals betw een the w aves etc.). www.statistik.at slide 16 | 3 June 2016

Who are those hard-to-survey? … and what effect might they have on the statistic‘s outcome? Folie 18 | 30.03.2016

Nonresponse and Total Survey Error _ Item-Nonresponse: Missings in Variables Counter-measure -> Imputation Item Non-Response 2014: 2% for Employment Income (from admin data > lack of identifier for linking but income receipt seems likely), <1% for Unemployment benefits (same reason as for empl. income), 10% for Self-employed income (surveyed) _ Unit-Nonresponse: Missing Persons or complete Households Counter-measure -> Weighting Is Nonresponse selective? www.statistik.at slide 19 | 3 June 2016 Groves et al. 2004:48