Title
European Union Statistics on Income and Living Conditions 2014
EU-SILC 2014 contains the 2014 Module on material deprivation.
Geographical Coverage
EU-28 member states and Iceland, Turkey, Norway and Switzerland. Test implementation: Montenegro, the former Yugoslav Republic of Macedonia (FYROM) and Serbia. (Eurostat: METHODOLOGICAL GUIDELINES AND DESCRIPTION OF EU-SILC TARGET VARIABLES. 2014 operation (Version October 2014, p. 18)
Time Period Covered
Reference period for core variables: constant, at selection, income reference persiod, usual week, last 4 weeks, last 12 months, since last year, working life, childcare reference period.
-
Country Specific Information: EU-SILC 2014
-
LV - Latvia
Producer
Central Statistical Bureau of Latvia
Universe
Persons living in private households within national territory
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Area units
Secondary Sampling Units
Dwellings
Stratification Criteria
The stratification was made depending on the type of municipality (Riga, cities, towns, rural areas). The Classification of Administrative Territories and Territorial Units (CATTU) of Latvia was used for stratification.
Number Of Rotational Groups
4
Type Of Data Source
Register,Survey
Interview Mode
- 64.1 % CAPI
- 32.4 % CATI
- 3.4 % PAPI
- 0.1 % Self-administered
Percentage Of Proxy Interviews
21.7 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Base weights were corrected by non-response in the primary sampling units. The 2012, 2013 and 2014 data were adjusted for returnees. New household members with RB110 = 3 (moved into from outside sample) and former household members with RB110 = 5, 6 or 7 (moved out, died, not registered in the previous wave and did not live in household anymore) had RB060 = 0. The newly born (household members with RB110 = 4) received the weight of their mother. For each year, each rotational group with adjusted design weights was calibrated on the corresponding year’s population.
Coverage Adjustment Weight: Method
For each year, each rotational group with adjusted design weights was calibrated on the corresponding year’s population. Weights were calibrated (in the household level) on the basis of demographic data by breaking them down by the type of municipality (Riga, cities, towns, rural areas), 16 age groups, sex and 6 regions of Latvia (NUTS 3). GREG calibration was used.
-
LU - Luxembourg
Producer
STATEC, Institut national de la statistique et des études économiques du Grand-Duché du Luxembourg;
LISER, Luxembourg Institute of Socio-Economic Research, formerly CEPS/INSTEAD, Centre d'Etudes de Populations, de Pauvreté et de Politiques Socio-économiques
Sampling Design
Probability
- random
- stratified
- one stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Sampling Method
Luxembourg uses a pure panel with a new sample each year.
Type Of Data Source
Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
32.3 %
-
LT - Lithuania
Producer
Statistics Lithuania
Universe
The target population of EU-SILC is all persons living in private households within the national territory of Lithuania at the time of data collection. Collective households and institutions are excluded from the target population.
Source Of Sampling Frame
Residents’ Register.
Sampling Design
Probability
- random
- stratified
- one stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Stratification Criteria
While selecting the new rotational group of the sample the country were grouped into 25 strata: 5 largest cities, other cities and rural area. Simple random sample of non–institutional persons aged 16 and over was selected from the Population Register in each stratum.
Number Of Rotational Groups
4
Start Date
02-2014
End Date
07-2014
Type Of Data Source
Register,Survey
Interview Mode
- 17.8 % CAPI
- 41 % CATI
- 41 % PAPI
- 0.2 % Self-administered
Percentage Of Proxy Interviews
16.9 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
To estimate household response probability logistic regression model are used. Response bias is estimated for responding and non-responding households.
Coverage Adjustment Weight: Method
Modified design weights are calibrated, seeking for the weights, which would remain as close as possible to sample design weights and allow obtaining some exact demographic estimates. Auxiliary variables: number of persons aged 0 and older (including newborn children) by different strata; number of persons by different age groups; number of males by different age groups.
-
HR - Croatia
Producer
Croatian Bureau of Statistics
Universe
All private households and their current members residing in the territory of Croatia at the time of data collection. Persons living in collective households and in institutions are excluded from the target population.
Source Of Sampling Frame
2011 Census data
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Area units
Secondary Sampling Units
Dwellings
Stratification Criteria
The sampling frame of segments for the first stage is divided into 4 strata which are defined according to statistical regions (NUTS-2) and the types of municipality. Beside explicit stratification, implicit stratification according to counties (NUTS-3) and municipalities was applied. It means that the list of segments within strata was sorted according to counties and municipalities before selection.
Number Of Rotational Groups
4
Type Of Data Source
Survey
Interview Mode
- 100 % CATI
Percentage Of Proxy Interviews
41.1 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The design weight of responded households is modified by a factor inversely proportional to the weighted response rate within a group of counties ("homogeneous group").
Coverage Adjustment Weight: Method
Both household and individual external information used in a single-shot calibration at household level. The new weights were calibrated on the number of private households in household size groups by regions / groups of counties, and the number of persons in gender–age groups by regions / groups of counties. External data was taken from the 2011 Census data.
-
RO - Romania
Corresponding National Study
Romanian Survey on Income and Living Conditions ("Quality of Life Survey")
Producer
National Institute of Statistics of Romania
Universe
The reference population is all private households and their current members residing in the territory of the Romania at the time of data collection. Persons living in collective households and in institutions are excluded from the target population.
Source Of Sampling Frame
Multifunctional Sample of Territorial Areas (EMZOT)
Sampling Design
Probability
- random
- stratified
- multi-stage
Primary Sampling Units
Census enumeration units
Secondary Sampling Units
Dwellings
Stratification Criteria
Stratification concerns only the first stage sampling. There are 88 strata, the criteria used being the area where a certain PSU is located (urban or rural area) and county (NUTS 3 level).
Number Of Rotational Groups
4
Type Of Data Source
Survey
Interview Mode
- 100 % PAPI
Percentage Of Proxy Interviews
13.9 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The non-response are not globally adjusted, at the entire sample level, but separately-at wave level, on groups of households, groups generated by the variables considered as explicative of the non response. For wave 1 as explicative variables for non-response region (NUTS II level) and area of residence (urban / rural) ar eused and for the second, third and fourth wave - the region.
Coverage Adjustment Weight: Method
The calibration is performed at the household level using the household variables and individual variables in their aggregate form as calibration variables. Adjustments were made using the SAS macro CALMAR. Calibration variables were: distribution of the population by age group, area of residence and gender using Romanian Population Estimates at the end of the income reference period and households totals by region.
-
NO - Norway
Producer
Statistics Norway
Universe
Persons aged 16 years or more at December 31 N-1 who are living outside an institution.
Source Of Sampling Frame
Central population register (BEREG)
Sampling Design
Probability
- random
- stratified
- one stage
Primary Sampling Units
Individuals
Secondary Sampling Units
Not used
Stratification Criteria
The new rotational groups are drawn as the proportion p of the population 16 years and over. In addition, each existing rotational group is then supplemented with new 16 year olds and new immigrants to ensure representativity. The sample is drawn from the population register, and this register is arranged to ensure geographical representativity. This is done by municipality and postal codes. The register is arranged by family number and personal code within the family before the actual selection of units.
Number Of Rotational Groups
4
Type Of Data Source
Register,Survey
Interview Mode
- 0.8 % CAPI
- 99.2 % CATI
Percentage Of Proxy Interviews
0 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Weights are calculated by stratifying the gross sample according to information in registers on sex, age, education and family size.
Coverage Adjustment Weight: Method
No adjustments are made, except for children's weights.
-
HU - Hungary
Producer
Hungarian Central Statistical Office
Universe
No difference to common definition
Source Of Sampling Frame
List of occupied dwelling units in census 2011 dataset
Sampling Design
Probability
- random
- stratified
- multi-stage
Primary Sampling Units
Localities
Secondary Sampling Units
Households
Stratification Criteria
At stage 1 the population of localities is stratified. Each of the larger localities is a stratum of its own. These are the self-representing localities, the number of which is 91. Smaller localities are stratified by NUTS3 regions and the size of locality. At stage 2 the households are stratified by the charachteristic of the head of household within each locality.
Number Of Rotational Groups
4
Start Date
03-2014
End Date
04-2014
Type Of Data Source
Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
11.2 %
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
In the new rotational group non-response weights were introduced to reduce bias caused by unit non-response on household level. Non-response adjustment was a simple expansion applied by the same classes as design factors were calculated by.
Coverage Adjustment Weight: Method
The aim of this adjustment was to improve the accuracy of data using socio-economical information available from the constantly updated Census 2001 and other surveys. Iterative raking scale methods were applied. For the integrative calibration population totals are used. Calibration was carried out with a self made SAS program. Calibration was carried ou in each rotational group.
-
FR - France
Corresponding National Study
Statistiques sur les Ressources et les Conditions de Vie
Producer
INSEE - National Institute for Statistics and Economic Studies
Universe
All private households and their current members residing in the territory of France at the time of data collection.
Source Of Sampling Frame
Statistiques sur les Ressources et Conditions de Vie (SRCV)
Primary Sampling Units
Municipality
Secondary Sampling Units
Dwellings
Number Of Rotational Groups
9
Type Of Data Source
Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
26.4 %
-
BG - Bulgaria
Producer
National Statistical Institute
Universe
All citizens officially living at Bulgarian territory (population de facto).
Source Of Sampling Frame
Population Census 2011
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Census enumeration units
Secondary Sampling Units
Households
Stratification Criteria
The sample is stratified by administrative-territorial districts in the country (NUTS3) and the household's location. As a result 56 strata are formed.
Number Of Rotational Groups
4
Start Date
04-2014
End Date
06-2014
Type Of Data Source
Survey
Interview Mode
- 100 % PAPI
Percentage Of Proxy Interviews
19.2 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Correction for non-response at the first year of subsamples was done with weighting within classes procedures: The design weights were modified by a factor inversely proportional to the response rate within strata. Coefficients of these corrections were computed separately according to classes of locality as ratios: the sum of design weights of selected units to the sum of design weights of responding units. The response probability for the households at wave 1, 2, 4 is obtained by a logistic regression model. The variables strata, size of household, sex, age group, activity, educational level and poverty indicators were used in the model.
Final Weighting: Method
The final weights were obtained applying the integrated calibration method. Combining the four independent subsamples, calibration is done on individual-level data, imposing equality of g-weights for individuals in the same household. Truncated linear function are used in order to limit g-weights close enough to 1. External information about the distribution of the population by administrative-territorial districts (NUTS 3) and residence and about the distribution of the population by age groups, sex and residence was used. This information was derived from the Information System Demography (ISD). Calibration was carried out with a G calib2.0 program (designed by Statistics Belgium).
-
BE - Belgium
Producer
Statistics Belgium
Universe
The reference population is all citizens living officially living at Belgian territory (population de jure).
Source Of Sampling Frame
Central Population Register
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Municipality
Secondary Sampling Units
Households
Number Of Rotational Groups
4
Type Of Data Source
Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
13.1 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
A logistic regression model was used to correct for non-response. The file was split by NUTS1 and the variables household size and region were used.
-
DE - Germany
Corresponding National Study
Leben in Europa
Producer
Federal Statistical Office of Germany
Universe
The sampling population for the whole sample comprises private households in Germany in their main residences.
Sampling Design
Probability
Number Of Rotational Groups
4
Type Of Data Source
Survey
Interview Mode
- 0.1 % CATI
- 99.9 % Self-administered
Percentage Of Proxy Interviews
18.9 %
-
DK - Denmark
Producer
Statistics Denmark
Universe
People residing in private households residing in Denmark at the time of the interview.
Source Of Sampling Frame
Central Population Register (CPR)
Sampling Design
Probability
- random
- not stratified
- one stage
Primary Sampling Units
Addresses
Secondary Sampling Units
Individuals
Number Of Rotational Groups
4
Type Of Data Source
Register,Survey
Interview Mode
- 28.8 % CATI
- 5.3 % Self-administered
- 65.9 % Websurvey
Percentage Of Proxy Interviews
7 %
Design Weight: Target
Household
Non-response Weight: Method
Calibration using external data
Coverage Adjustment Weight: Method
All external data are based on variables in administrative registers. Exactly the same variables are found in the sampled households. The following external variables have been used: Equivalised disposable income, size of the household, education level of the person with the highest professional status in the household and socio-economic status of main income holder in the household.
In addition to correcting for bias due to non-response on the household level, the SILC cross-sectional weights should reproduce certain demographic and poverty distributions on the personal level. Therefore some data on the personal level have been integrated in the calibration, that risk of poverty, age, sex, poverty, family type, education, income mass and income groups.
-
FI - Finland
Producer
Statistics Finland
Universe
Members of the private households permanently resident (the census definition) in Finland on 31 December 2013. For migrants in particular, permanent residency means that the persons have resided or intend to reside for at least 12 months and they have not a permanent recidence abroad. Persons living in institutions, in collective households or in residential homes have been excluded.
Source Of Sampling Frame
The Population Information System maintained by the Population Register Centre of Finland
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Individuals
Secondary Sampling Units
Dwellings
Stratification Criteria
Strata contain socio-economic groups grouped by income class (13 strata in the new sample). The register information used for strata construction is about all members of the selected sample person's household-dwelling unit available at the time of sample selection (updated by the final population register data later). Strata are created only for those who do not belong to the over-coverage.
Number Of Rotational Groups
4
Start Date
01-2014
End Date
05-2014
Type Of Data Source
Register,Survey
Interview Mode
- 2.6 % CAPI
- 97.4 % CATI
Percentage Of Proxy Interviews
10.1 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The household design weights were multiplied by a simple adjustment n(sample)/n(respondents) in every stratum.
Coverage Adjustment Weight: Method
The calibration can be interpreted as integrative, i.e. both the household and person levels were included in the process. The levels and variables included in the calibration were as follows: 1) Households: Region, type of municipality by degree of urbanisation (national classification), household size, total sums of 15 different income variables. 2) Persons: Sex and age classes. Calibrations were performed by the raking method with the SAS supporting CALMAR macro.
-
IE - Ireland
Producer
Central Statistics Office (CSO)
Universe
Persons aged 16 years and over living in private households.
Number Of Rotational Groups
4
Type Of Data Source
Register,Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
26.1 %
-
AT - Austria
Producer
Statistics Austria
Universe
"All private households and their current members residing in the territory of Austria at the time of data collection. Persons living in collective households and in institutions are generally excluded from the target population." (Eurostat (2015). National Reference Metadata in ESS Standard for Quality Reports Structure (ESQRS), AT, p. 6.)
Source Of Sampling Frame
Central registration register ZMR (Zentrales Melderegister), a constantly updated population register based on the register of residence.
Sampling Design
Probability
- stratified
- one stage
Primary Sampling Units
Dwellings
Secondary Sampling Units
Not used
Stratification Criteria
The sample of the first wave was stratified according to 205 interviewer units (Sprengel). These are regional divisions of federal territory which may be approximately combined to Austrian provinces (NUTS 2 units). (Eurostat (2015). National Reference Metadata in ESS Standard for Quality Reports Structure (ESQRS), AT, p. 2.)
Number Of Rotational Groups
4
Start Date
02-2015
End Date
09-2015
Type Of Data Source
Survey
Interview Mode
- 60.5 % CAPI
- 39.5 % CATI
Percentage Of Proxy Interviews
11.4 %
Weighting: Method
See Eurostat (2015). National Reference Metadata in ESS Standard for Quality Reports Structure (ESQRS), AT 2014, Annex 6, p. 1.)
-
CZ - Czech Republic
Producer
Czech Statistical Office
Universe
All persons with usual residence in Czech dwellings (their only place of residence or their main place of residence) were included in the survey.
Source Of Sampling Frame
Register of Census Enumeration Districts (small geographical areas covering the whole territory of the country)
Sampling Design
Probability
- random
- stratified
- multi-stage
Primary Sampling Units
Census enumeration units
Secondary Sampling Units
Dwellings
Stratification Criteria
The sampling of CEUs is stratified by region (NUTS4) and municipality size.
Number Of Rotational Groups
4
Start Date
02-2014
End Date
05-2014
Type Of Data Source
Survey
Interview Mode
- 44.2 % CAPI
- 55.8 % PAPI
Percentage Of Proxy Interviews
25.9 %
Design Weight: Method
none
Non-response Weight: Method
United calibration for all the waves was used as the method for correcting non-response. The achieved sample was re-weighted using the integrated calibration technique (producing the same weights on household and personal level). This technique ensures that the weighted sample structure corresponds to a set of known external population characteristics. The calculations were implemented using the CALMAR software in SAS.
Coverage Adjustment Weight: Method
For the adjustments to external data the following calibration variables were used: Number of inhabited dwellings in each NUTS3 region, subdivided into family houses (detached and semi-detached houses) and flats, based on the 2001 Census and 2011 Census continuously updated from administrative sources of construction authorities and Population characteristics (population totals in each NUTS3 region, economic activity in each NUTS3 region, demographic characteristics at the national level).
-
CY - Cyprus
Corresponding National Study
Survey of Income and Living Conditions
Producer
Statistical Service of Cyprus
Universe
All the households and their members living in the areas under the effective control of the government of the Republic of Cyprus. Population in collective households and institutions is excluded.
Source Of Sampling Frame
The sampling frame used for the 2013 was the list of households from the 2011, Census of Population
Sampling Design
Probability
- random
- stratified
- one stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Stratification Criteria
Geographical stratification criteria were used for the sample selection. The households were stratified in 9 strata based on District (Urban / Rural).
Number Of Rotational Groups
4
Start Date
02-2014
End Date
06-2014
Type Of Data Source
Register,Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
12.3 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The empirical response rate within each stratum provides an estimate of the response probability for all the households of the stratum.
Coverage Adjustment Weight: Method
The individual variables are aggregated at the household level by calculating household totals such as the number of male/female in the household, the number of persons aged 16+ etc. After that, calibration is done at the household level using the household variables and the individual variables in their aggregate form. At personal level the calibration variables used were the distribution of population by age and gender. Based on this calibration procedure and using the weight after non-response adjustment as the initial weight, the household and the personal cross-sectional weights were calculated.
-
SE - Sweden
Producer
Statistics Sweden
Universe
The refence population is the whole Swedish population except short term migrants. People who stay in Sweden for no longer than 3-12 months are not covered.
Sampling Design
Probability
- random
- stratified
- one stage
Primary Sampling Units
Individuals
Secondary Sampling Units
Not used
Stratification Criteria
mple is stratified by age in eight strata according to age.
Number Of Rotational Groups
4
Type Of Data Source
Register,Survey
Interview Mode
- 0.1 % CAPI
- 99.9 % CATI
Percentage Of Proxy Interviews
1.7 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The adjustment for non-response is done by straight expansion within each poststratum.
-
SI - Slovenia
Producer
The Statistical office of the Republic of Slovenia
Universe
The reference population is defined with the persons in the Central Register of Population which are aged 16 years or more. The individuals with Slovenian citizenship as well as foreigners were included in the sampling frame.
Source Of Sampling Frame
Central Register of Population (CRP)
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Census enumeration units
Secondary Sampling Units
Individuals
Stratification Criteria
The sampling frame of persons aged 16 years or more is divided into 6 strata, which are defined according to the size of the settlement and the proportion of agricultural households in the settlement.
Number Of Rotational Groups
4
Start Date
01-2014
End Date
06-2014
Type Of Data Source
Register,Survey
Interview Mode
- 47.5 % CAPI
- 52.5 % CATI
Percentage Of Proxy Interviews
19.6 %
Design Weight: Target
Person
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The non-response factor was calculated for each stratum. First the sample was divided into three categories: responses, non-responses and out-of-scope units.
Coverage Adjustment Weight: Method
By the calibration procedures the weighted sums of some key variables are set to the known population values. These population values are obtained from the different administrative sources. For the calibration of weights we used SAS Macro Calmar. We performed calibration for the level of households, as well as for the level of the persons.
-
SK - Slovakia
Producer
The Statistical Office of the Slovak Republic
Universe
Private households
Source Of Sampling Frame
Data set is based only on a survey
Sampling Design
Probability
- random
- stratified
- multi-stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Stratification Criteria
There are two criteria of area stratification in the sampling design:
- geographical stratification (8 standard administrative regions corresponding to the European NUTS 3 level.).
- degree of urbanization: 7 groups according to population size of municipalities and communes (number of inhabitants in municipalities and communes).
48 final strata were created (variable DB050) by using of those two stratification criteria.
Number Of Rotational Groups
4
Start Date
04-2014
End Date
06-2014
Type Of Data Source
Survey
Interview Mode
- 99.9 % PAPI
- 0.1 % Self-administered
Percentage Of Proxy Interviews
5.1 %
Design Weight: Target
Household
Non-response Weight: Method
Dividing households into strata (region and rotational group). Then the empirical value of the response rate within the stratum gives the estimate of the probability of response for each household in the stratum.
Coverage Adjustment Weight: Method
Calculation of the households crosssectional weights - has been implemented by system of simple calibration of weights of the households by using of variablesnumber of households by number of membership in administration regions.
Final Weighting: Method
The sample was designed as a selfweighting sample. Design factor for all sampled dwellings is equal to 1.
-
IS - Iceland
Producer
Statistics Iceland
Universe
People in private households
Source Of Sampling Frame
Iceland population register
Sampling Design
Probability
- random
- not stratified
- one stage
Primary Sampling Units
Individuals
Stratification Criteria
The sample is post stratified.
Number Of Rotational Groups
4
Type Of Data Source
Register,Survey
Interview Mode
- 100 % CATI
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Post stratification weights are used to adjust the data to the population. The information on the population comes from the national register. The weights both adjust for nonresponse and sampling error. The post stratification weights are based on age , sex and residence.
Coverage Adjustment Weight: Method
Results are only calibrated with numbers from the national register.
-
UK - Great Britain
Corresponding National Study
Family Resources Survey
Producer
Office for National Statistics
Universe
The reference population for the Family Resources Survey is all private households in the UK.
Source Of Sampling Frame
for Great Britain: Royal Mail’s small-users Postcode Address File (PAF); for Nothern Irland: POINTER address database
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Postecode sectors
Secondary Sampling Units
Addresses
Stratification Criteria
The PSUs are stratified by 27 regions on the FRS and 30 regions on the GLF. The PSUs are then sorted by three other variables (different for the two surveys), which provides an implicit stratification:
For the FRS, within each region the postcode sectors are ordered and grouped into eight equal bands using the proportion of households where the household reference person (HRP) is in National Statistics Socio-Economic Classification (NS-SEC) 1 to 3. Within each of these eight bands, the PSUs are ordered by the proportion of economically active adults aged 16-74 and formed into two further bands, resulting in sixteen bands for each region. These are then ordered according to the proportion of economically active males aged 16-74 who are unemployed.
For the GLF, postcode sectors were stratified according to selected indicators taken from the 2001 Census. Sectors were initially ordered according to the proportion of households with no car, then divided into three bands containing approximately the same number of households. Within each band, sectors were re-ordered according to the proportion of households with a household reference person in socio-economic groups 1 to 5 and 13, and these bands were then subdivided into three further bands of approximately equal size. Finally, within each of these bands, sectors were re-ordered according to the proportion of people who were pensioners.
Sampling Method
Prior to 2012, all households for cross-sectional and longitudinal EU-SILC originated from the GLF sample. Since 2012, cross-sectional and wave 1 respondents have been selected from the FRS. The FRS in Great Britain is a stratified clustered probability sample design, as was the first wave of the GLF. The FRS in Northern Ireland is a systematic stratified sample design, while for the GLF it was a simple random sample.
Start Date
04-2014
End Date
09-2014
Type Of Data Source
Survey
Interview Mode
- 76.6 % CAPI
- 23.4 % Websurvey
Percentage Of Proxy Interviews
0 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
There is no correction for individual non-response, because only fully cooperating households are included in the dataset. Household non-response is corrected for in the grossing procedure, which sets grossing factors so that the population estimate is representative of the population in several dimensions.
Coverage Adjustment Weight: Method
The grossing regime for Great Britain aims to calibrate population estimates derived from EU SILC data against seven different variables: Individuals (Age, sex and region), Dependants aged 16-19, Benefit units, Benefit units (Lone Parents), Households (Tenure type), Households (Council Tax Band), Households (Region).
The grossing regime for Northern Ireland is slightly different from that for the rest of the UK: Local taxes in Northern Ireland are collected through the rates system, so Council Tax Band as a control variable is not applicable. Northern Ireland housing data are based largely on small sample surveys. It is not desirabe to introduce the variance of one survey into another by using it to compute control totals; therefore tenure type has not been used as a control variable.
-
IT - Italy
Producer
Italian National Institute of Statistics
Universe
All persons living in private households.
Source Of Sampling Frame
Registers of the municipalities
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Municipality
Secondary Sampling Units
Households
Stratification Criteria
Stratification of primary sampling units by the number of inhabitants so that the total number of inhabitants in each stratum is approximately constant (this guarantees selfweighting design in each region). Municipalities which sizes are higher than a threshold are self-representing units i.e. are strata themselves and included with certainty in the sample of PSU. Secondary sampling units are not stratified.
Number Of Rotational Groups
4
Start Date
04-2014
End Date
10-2014
Type Of Data Source
Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
30.6 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The design weights were adjusted by a correction factor for total non-response worked out as the reciprocal of the response probability identified by the information given on the extracted sample (for the households at wave 1) or gathered from the most recent year of survey (for the households at wave 2, 3, 4).
Coverage Adjustment Weight: Method
After the non-response adjustments, a further adjustment of design weights has been achieved with a first calibration stage to constrain the sample to the same structure of the Labour Force Survey with regard to the educational and professional condition by sex.
-
MT - Malta
Producer
National Statistics Office Malta
Universe
All private households and their current members residing in Malta at the time of data collection. Persons living in institutions are excluded from the target population.
Source Of Sampling Frame
2011 Census of Population & Housing
Sampling Design
Probability
- random
- not stratified
- one stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Stratification Criteria
Not applicable
Number Of Rotational Groups
4
Start Date
05-2014
End Date
08-2014
Type Of Data Source
Survey
Interview Mode
- 91.9 % CAPI
- 8.1 % CATI
Percentage Of Proxy Interviews
31.1 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Correction for non-response was carried out separately for each panel. For new households, the adjustment for non-response at individual level was incorporated in the calculation of design weights.
For old households, i.e. for the remaining three panels, the adjustment for attrition was carried out through post-stratification at personal level. The values of the variables for SILC 2014 used in the post-stratification were as at 2013 (since these may not be necessarily the same as in the current situation). Specifically, the variables used were age-group, sex and district (NUTS 4), also as at 2013.
Coverage Adjustment Weight: Method
A temporary weight was created as the product of the design weight (or previous year’s cross-sectional weight for old households) and non-response adjustment. This weight was normalised to reflect the total population as at December 2013 and trimmed so as to lie within the lowest and highest deciles, thus reducing the range of the weights. The logit method was applied and the calibrating variables used were household size, districts (NUTS 4), household type and number of persons in a household.
-
PL - Poland
Producer
Central Statistical Office of Poland
Universe
The survey unit was a household and all the household members who have reached the age of 16 years by December 31, 2013.
The survey did not cover collective accomodation households
Source Of Sampling Frame
TERYT system
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Census enumeration units
Secondary Sampling Units
Dwellings
Stratification Criteria
The strata were the voivodships (NUTS2) and within the voivodships primary sampling units were classified by class of locality. In urban areas census areas were grouped by size of town. Big cities formed independent strata, but in the five largest cities districts were treated as strata. In rural areas strata were represented by rural gminas (NUTS5) of a subregion (NUTS3) or of a few neighbouring powiats (NUTS4). Altogether, 250 strata were distinguished.
Number Of Rotational Groups
4
Start Date
05-2014
End Date
07-2014
Type Of Data Source
Survey
Interview Mode
- 100 % PAPI
Percentage Of Proxy Interviews
23 %
Design Weight: Target
Dwelling
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Weights were adjusted with the use of household non-response rates estimated for each class of locality separately.
Coverage Adjustment Weight: Method
With the use of the integrated calibration method weights were calculated for individuals and for households simultaneously. The following information about households was used: 4 size categories and number of persons by age and gender. The data at the level of NUTS2, additionally classified by urban/rural areas, were derived from the 2011 Census and current demographic estimates.
-
PT - Portugal
Corresponding National Study
ICOR (Income and living Conditions)
Producer
Statistics Portugal
Universe
All private households and their current members living in Portugal (Mainland, Açores and Madeira).
Sampling Design
Probability
Number Of Rotational Groups
4
Type Of Data Source
Survey
Interview Mode
- 100 % CAPI
Percentage Of Proxy Interviews
36.1 %
-
RS - Serbia
-
CH - Switzerland
Producer
Federal Statistics Office
Source Of Sampling Frame
The first sample (wave 1) was drawn from the SRPH survey base in 2014. Previously, the sample had been drawn from the CASTEM (v2-3-4) survey base. (Eurostat (2014). National Reference Metadata in ESS Standard for Quality Reports Structure (ESQRSSI), CH, p.3.)
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Stratification Criteria
The seven major geographical regions (level NUTS2)
Number Of Rotational Groups
4
Start Date
03-2014
End Date
07-2014
Type Of Data Source
Register,Survey
Interview Mode
- 0.1 % CAPI
- 99.9 % CATI
Weighting: Method
see Eurostat (2014). National Reference Metadata in ESS Standard for Quality Reports Structure (ESQRSSI), CH, Appendix 16.
-
ES - Spain
Producer
National Statistics Institute (INE-Spain)
Universe
The target population was members of private households residing at main family addresses, and the households themselves.
Source Of Sampling Frame
Municipal Register
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Census enumeration units
Secondary Sampling Units
Addresses
Stratification Criteria
In each Autonomous Community [self-ruling region], first-stage units were stratified by the size of the municipality to which the census section belonged.
Number Of Rotational Groups
4
Start Date
03-2014
End Date
07-2014
Type Of Data Source
Survey
Interview Mode
- 82.6 % CAPI
- 17.4 % CATI
Percentage Of Proxy Interviews
25.5 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Adjustment for non-response is carried out by multiplying a certain factor to provide an estimate of the inverse probability of response in the stratum.
Coverage Adjustment Weight: Method
The factor is weighted to adjust estimated distribution to the population distribution by Autonomous Community, age group and gender provided by the Demographic Projections Unit.
-
NL - Netherlands
Producer
Statistics Netherlands
Universe
The reference population of EU-SILC is all private households and their current members residing in the Netherlands at the time of data collection. The West Frisian Islands with the exception of Texel were excluded from the target population. This is also true for persons living in collective households and in institutions.
Source Of Sampling Frame
Dutch municipality administration (Gemeentelijke Basisadministratie or GBA)
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Municipality
Secondary Sampling Units
Addresses
Stratification Criteria
The stratification variables used in this design are the 40 COROP-regions (NUTS3). These are regional areas within the Netherlands and are used for analytical purposes by, among others, Statistics Netherlands.
Number Of Rotational Groups
4
Start Date
06-2014
End Date
10-2014
Type Of Data Source
Register,Survey
Interview Mode
- 100 % CATI
Percentage Of Proxy Interviews
1.5 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
The method of logistic regression was adopted to approximate the response probabilities for the new rotational group. These response probabilities were modelled by the explanatory variables age, degree of urbanisation, type of household, and labour status.
Coverage Adjustment Weight: Method
The non-response adjustments and the adjustments to external data are combined into a single procedure. The following variables were included in the calibration scheme: Household size, Sex, Age class.
-
EE - Estonia
Producer
Statistics Estonia
Universe
All Estonian citizens and those having a living permission. Persons living in collective households are included in the reference population.
Source Of Sampling Frame
Sampling frame for selection of new part of the sample was the 2011 Population and Housing Census. Census involves the data of all permanent residents, households and dwellings in Estonia.
Sampling Design
Probability
- systematic
- stratified
- one stage
Primary Sampling Units
Households
Secondary Sampling Units
Not used
Stratification Criteria
Geographical stratification was used. The counties (and capital Tallinn) were grouped into three strata by the population size.
Number Of Rotational Groups
4
Start Date
02-2014
End Date
05-2014
Type Of Data Source
Survey
Interview Mode
- 99 % CAPI
- 1 % CATI
Percentage Of Proxy Interviews
29.1 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Correction for non-response was done with logistic regression model predicting the response probability of the household. Auxiliary variables were: sex and age of address-person, county group and urbanization status (rural/urban).
Coverage Adjustment Weight: Method
Non-response corrected weights were calibrated to reproduce known population totals.
Calibration totals originated from demographic statistics published by Statistics Estonia with institutionalized persons deducted. Calibration was made on the following auxiliaries: County, Sex x age group; urbanization status (rural/urban).
-
EL - Greece
Producer
Hellenic Statistical Authority
Universe
The reference population is all citizens officially living at Greek territory (population de facto).
Source Of Sampling Frame
2011 population census
Sampling Design
Probability
- systematic
- stratified
- multi-stage
Primary Sampling Units
Area units
Secondary Sampling Units
Households
Stratification Criteria
There are two levels of area stratification in the sampling design. The first level is the geographical stratification based on the partition of the total country area into thirteen (13) Regions corresponding to the European NUTS2 level. The two former major city agglomerations of Greater Athens and Greater Thessalonica constitute separate major geographical strata. The second level of stratification entails grouping Municipal/Local communes within Region by degree of urbanization, i.e., according to their population size. The scaling of urbanization was finally designed in four groups.
Number Of Rotational Groups
4
Start Date
06-2014
End Date
09-2014
Type Of Data Source
Survey
Interview Mode
- 2 % CAPI
- 8.7 % CATI
- 89 % PAPI
- 0.3 % Self-administered
Percentage Of Proxy Interviews
3.4 %
Design Weight: Target
Household
Design Weight: Method
Inverse of Selection Probability
Non-response Weight: Method
Within each design stratum, the non-response adjustment of the responding households is carried out by the inverse of the response rate, so as to “make up” for non-responding cases in that stratum. Target variable DB080 was adjusted for non-response for the variables DB 120 (record of contact at address) and DB 130 (household questionnaire result). The corrections were conducted at subsequent steps. The multiplication of DB080 with each one of the two corrections, results in a corrected DB080 weight that is used as initial weight in the calibration procedure.
Coverage Adjustment Weight: Method
Calibration of the household and personal weights in conjunction is done with external data. The auxiliary variables used at household level are the household size, the tenure status and the region (NUTS2). At personal level the auxiliary variable used is the distribution of population by age and sex. The weights obtained after this procedure of calibration are the household cross-sectional weights. As all the household members reply to the household questionnaire, it is also the weight of each member of the household. The last step involves the calculation of the personal cross sectional weights for household members aged of 16+. The calibration procedure was applied again using as initial weights the weights of each member of the household and as auxiliary variable the distribution of population aged 16+ by age and sex.