Attrition in EUSILC in the Czech Republic Presentation
Attrition in EU-SILC in the Czech Republic Presentation for ESRA Conference June 28 th 2007 Martina Mysíková, Štěpán Tourek, Martin Zelený CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Design of EU-SILC § Four year rotational panel § households are visited continuously for four years § each year ¼ dropped and ¼ newly selected § sample persons moved to another private households are traced and visited § First wave: 4351 households responded (response rate 64, 6%) § Second wave: 87, 7% Response, 11, 2% Nonresponse, 1, 1% Out of survey population CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Design of EU-SILC 2005 2006 2007 2008 2009 1 2 1 3 2 1 4 3 2 1 CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Households CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Household distribution NUMBER OF HOUSEHOLD MEMBERS § The distribution by number of household members is the same at „response“ and „non-response“ § Among „out“ prevail households with one member CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Household distribution NUMBER OF CHILDREN § The distribution by number of children is similar at „response“ and „non-response“ § Among „out“ prevail households without children CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Household characteristics SIZE OF MUNICIPALITY § The „non-response“ is highest in big cities (18%) Small – less than 49999 inhabitants Middle – 50000 -99999 inhabitants Big – more than 100000 inhabitants CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Household characteristics INTERVIEWER § The change of the interviewer has some impact on „nonresponse“ (15%) CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individuals CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individual distribution NET INCOME DECILES § The greatest share of „non-response“ falls in higher deciles CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individual characteristics GENDER § No differences by gender CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individual characteristics AGE § Higher „non-response“ at middle-age groups § Highest „out“ at oldest group CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individual characteristics MARITAL STATUS § Higher „non-response“ at separated (18%), divorced (13%) and single (13%) § Highest „out“ at widowed CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individual characteristics ECONOMIC ACTIVITY § Higher „non-response“ at self-employed (17%) § Highest „response“ at retired (not working) and those receiving parental allowance § Highest „out“ at retired (not working) CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Individual characteristics EDUCATION § „non-response“ rises with higher educational level § „out“ increases moderately with lower educational level CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Summary § Household level § „Non-response“ § Higher in households with working members § Lower in households with retired (not working) members § Higher in bigger cities § „Out“ § Mainly one member households (85%) – retired (institutions or decease) CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Summary § Individual level § Tight connection with time and economic activity § „Non-response“ § No impact of gender § Higher at middle-age groups, self-employed, higher educated, high-income groups § Lower at old groups, widowed, retired, parental leave probably spend more time at home § „Out“ § Mainly old, widowed and retired (institutions or decease) and partly young (move out of country) § Mainly low-income groups CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights § Focus on impact of attrition on weights and on selected variables 1. Take the values from the 1 st wave 2. Simply eliminate those who didn't response in the 2 nd wave (the same weights) 3. Recalculate the weights excluding those who didn't respond in the 2 nd wave CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights 1. Original sample Original weights “Non-response” and “Out” households eliminated 2. Exclusion Original weights Weights recalculated 3. Exclusion New weights Optimally: § the 3 rd step should bring the values closer to the original values CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights § Dwelling § 3 rd step lowers the difference by one half CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights § Problems with the dwelling § the 3 rd step brings the values closer in all cases CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights § Can not afford. . . § the 3 rd step can also increase the difference CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights § Regular monthly inter-household transfer § great decrease at received and only slight increase at paid transfer CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Impact on weights § Annual household income § 3 rd step brings the values closer to the original ones CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
Conclusion § The 3 rd step § brings the values back in most cases (e. g. dwelling) § increases the difference in some cases (e. g. affordability) § The influence of the attrition on weights and results in next wave is relatively negligible but can be significant for certain variables (diagnostics) CZECH STATISTICAL OFFICE Na padesátém 81, CZ - 100 82 Praha 10, Czech Republic www. czso. cz
- Slides: 25