Commission Implementing Regulation (EU) 2022/2094 of 28 October 2022 specifying the technical items of data sets, establishing the technical formats for transmission of information and specifying the detailed arrangements and content of the quality reports on the organisation of a sample survey in the consumption domain pursuant to Regulation (EU) 2019/1700 of the European Parliament and of the Council (Text with EEA relevance)

ANNEX III

DETAILED ARRANGEMENTS AND CONTENT OF QUALITY REPORTS

• (1) Member States shall transmit a quality report to the Commission (Eurostat) according to the rules set out in Article 13 of Regulation (EU) 2019/1700 and Regulation (EU) 2019/2180 specifying the detailed arrangements and content for the quality reports ¹⁶.

• (2) Member States shall include the following specific information in the quality report on the consumption domain.

STATISTICAL PRESENTATION

Data description

• — Title of the survey at national level, year of the survey, link to the survey website

Classification systems

List of the version of classification used in the data, and any deviations from European statistical standards or international standards.

Statistical concepts and definitions, including reference period

• — List of concepts and variables that deviate from the standard definitions, mentioning the national concepts used and any differences between the national concepts and the respective standard definitions

• — Consumption expenditure shall be calculated as household final consumption expenditure; any deviation from this standard definition shall be described

• — Method for calculating imputed rent (stratification/regression method specifying the type of model/regression used and variables used; self-assessment method).

• — Method for calculating/imputing income and income components, if applicable

STATISTICAL PROCESSING

Source data

• — Description of the data source used for building the sampling frame (e.g. population register, household register, dwelling register, population census, another survey, etc.), frequency of update and year of last update of the data source, and method used for obtaining or creating the sampling frame. Survey name shall be indicated where another survey has been used as a data source.

• — Regarding sample design:

• — If the ultimate sampling unit is different from the private household, it shall be reported

• — Methods for sample design (e.g. simple random sampling; systematic sampling; stratified sampling; cluster sampling, multiple stage sampling, combination of designs). If Member States use stratified sampling, the stratification and sub-stratification criteria shall be reported. If Member States use multiple stage sampling, the different stages and the corresponding probabilities (equal, unequal, proportional to size) shall be described. If Member States use a combination of designs, the designs used shall be described.

• — Sample size

Data collection

• — Description of methods used to conduct the interview (e.g. PAPI, CATI, CAPI, CASI, CAWI ¹⁷, and/or other sources such as registers, other surveys, smart ways of data collection based on mobile apps, etc.) The variables collected based on other sources shall be listed.

• — Description of methods used to compile the consumption expenditure diary (e.g. pen and pencil diary, computer-based diary, web diary, cash-register receipts, receipt scanning, loyalty cards, administrative data, mobile apps and/or other sources of collecting data in a smart way, etc.)

Data compilation

• — Household weightings: Each step of weighting should be described separately: design weights; weight adjustments for non-response, weight adjustments to external data sources (calibration techniques used, level and variables used for the adjustment), any other weight adjustments (e.g. trimming, top (bottom)-coding of the weight distribution to adjust for outliers); final weights.

• — Where imputation is used, any imputed variables shall be listed, the imputation procedure used shall be described and the reasons for imputation shall be explained. The number of imputed values as a percentage of the total number of observations shall be reported for each income and income component variable, including income in kind from non-salaried activities, both at household and personal level.

ACCURACY AND RELIABILITY

Sampling error

• — Precision requirements shall be expressed in standard errors for the indicator ‘Percentage of households whose expenditure on housing-related categories including water, electricity, gas and other fuels, are more than 50 % of the total expenditure’, and are defined as continuous functions of the actual estimate and of the size of the statistical population in a country as defined in Annex II of Regulation (EU) 2019/1700 for the consumption domain.

• — For the total consumption expenditure (HE00) and the 2-digit COICOP level consumption expenditure (HE01 to HE13) the mean, the total number of observations and the standard errors shall be reported at national level.

• — Countries shall describe the methodology for calculating the precision estimates. Where the precision requirements are reached by combining microdata concerning a maximum of three successive years of observations in line with point 8 of Annex II of Regulation (EU) 2019/1700, countries shall describe the relevant alternative methods used to estimate and assess the precision.

Non-sampling error

Measurement error

The different sources of measurement errors likely to be found in the survey shall be described. This includes:

• — Description and testing of the questionnaire and diary design,

• — Description of interview training (e.g. number of training days, skills testing before starting the fieldwork (rate of success, etc.)),

• — Information on studies to check the impact of the interviewer, such as re-interviews, record check studies, or split-sample experiments (if available), and results from models (if available).

Non-response error

Unit non-response

• — Member States shall compute household non-response rates as follows:

Household non-response rates (NRh) shall be computed as follows:

NRh = (1-(Ra * Rh)) * 100,

Where:

Ra is the address contact rate, and it is computed as the Number of addresses successfully contacted divided by the Number of valid addresses selected;

Rh is the proportion of completed household interviews accepted for the database, and it is computed as the Number of household interviews completed and accepted for database divided by the Number of eligible households at contacted addresses.

• — For those Member States where controlled substitutions are made in case of unit non-response, non-response rates shall be calculated before and after substitution.

• — Breakdowns of non-respondents by characteristics and reasons for non-response as far as available

• — Description of any measures taken to correct for non-response, and evaluation of any difficulties such measures may indirectly involve (e.g. calibration involving high and volatile weights)

• — If controlled substitutions are applied in cases of unit non-response and under the provisions of Article 6(4), (5,) (6) and (7), substitution rates shall be provided as well as a description of the method of selection of substitutes, main characteristics of substituted units compared to the original units, distribution of substituted (original) units by number of attempts for contact at address, household interview/diary acceptance, and household questionnaire/diary result.

• — Number of households successfully contacted before and after substitution

• — Number of responding households before and after substitution

• — Gross sample size, number of eligible units and net sample size, including substitution units (achieved sample size) needs to be provided

Item non-response

For variables Household Grid, Age (in completed years) of household member, Main activity status (self-defined), all income and income component variables, Main source of income, Household Tenure Status, the following information shall be provided:

• — Percentages of households (per variable collected or compiled at household level) or percentage of persons (per variable collected or compiled at personal level) having received an amount for each variable,

• — Percentage of households (per variable collected or compiled at household level) or percentage of persons (per variable collected or compiled at personal level) with missing values (before imputation) for each variable collected or compiled at household/personal level. Variables constructed/collected from several component variables where there is a missing value before imputation for at least one component at household/personal level shall be treated as variables with missing values before imputation for that household/person.

Processing error

• — Information to be provided about any errors in processing and their impact on the final data collection results arising from faulty implementation of correctly planned implementation methods. An account shall be given of the main errors detected in the post-data-collection process.

• — Description of the quality checks (e.g. data entry controls, coding controls, etc.) and the data editing process (e.g. description of the main data editing rules regarding consumption expenditure and income variables).

• — Imputation procedures shall be described. Information to be provided on the imputation rate calculated as the share of observations imputed (for each variable) out of the total number of observations.

Data revision – policy

Information shall be provided about any policy designed to ensure the transparency of disseminated data, whereby the preliminary data compiled are subsequently revised. If data are revised, this shall be reported.

Data revision -practice

• — Schedule for revisions

• — Main reasons for revisions and their nature (new source data available, new methods, etc.)

• — Impact of revisions on indicators

COHERENCE AND COMPARABILITY

Coherence – cross domain

Coherence – EU Statistics on Income and Living Conditions (EU-SILC)

A comparison of the following income variables with EU-SILC shall either be implemented at national level or be validated by the countries on the basis of calculations made by Eurostat: At-risk-of-poverty threshold (EUR), At-risk-of-poverty rate (%), Relative at-risk-of-poverty gap, Income quintile share ratio S80/S20, Gini coefficient.

Coherence – Harmonised Indices of Consumer Prices (HICP)

A comparison of the structure of consumption expenditure at 2-digit COICOP level shall either be implemented at national level or be validated by the countries on the basis of calculations made by Eurostat with the corresponding variables from HICP.

Coherence – National accounts

A comparison of the structure of Consumption Expenditure at 2-digit COICOP level shall either be implemented at national level or be validated by the countries on the basis of calculations made by Eurostat with the corresponding variables from national accounts.

Coherence – internal

Any lack of coherence in the HBS data set shall be reported, together with explanations for such inconsistencies.