September 2020 QUALITY REPORT HOUSEHOLD BUDGET SURVEY (HBS) / PORABA V GOSPODINJSTVIH (APG) FOR 2018 Brigita Vrabič Kek
September 2020
QUALITY REPORT
HOUSEHOLD BUDGET SURVEY (HBS) / PORABA V GOSPODINJSTVIH (APG)
FOR 2018
Brigita Vrabič Kek
Quality report 2/22
CONTENT
Methodological explanations on the statistical survey ........................................ 3
1 Relevance .................................................................................................... 6
1.1 Description of administrative sources used ........................................... 6
1.2 Users of survey data ............................................................................. 6
1.3 Completeness of statistical results ........................................................ 7
2 Accuracy ...................................................................................................... 7
2.1 Standard errors ..................................................................................... 7
2.2 Coverage bias ....................................................................................... 9
2.3 Non-sampling errors ............................................................................. 9
3 Timeliness and punctuality ......................................................................... 15
3.1 Timeliness ........................................................................................... 15
3.2 Punctuality .......................................................................................... 16
4 Accessibility and clarity .............................................................................. 16
4.1 Accessibility ........................................................................................ 16
4.2 Clarity .................................................................................................. 17
5 Comparability ............................................................................................. 17
5.1 Comparability over time ...................................................................... 18
5.2 Geographical comparability................................................................. 19
6 Coherence ................................................................................................. 19
6.1 Coherence between provisional and final data ................................... 19
6.2 Coherence with the results of the reference source ............................ 20
7 Burdens ..................................................................................................... 21
7.1 Burden of reporting units ..................................................................... 21
7.2 Explanations ....................................................................................... 21
8 Evaluation of survey quality ....................................................................... 21
8.1 General quality assessment ................................................................ 21
8.2 Measures for improvement ................................................................. 22
Quality report 3/22
METHODOLOGICAL EXPLANATIONS ON THE STATISTICAL SURVEY
Purpose of the survey
The purpose of the Household Budget Survey (HBS) is to become familiar with the level and structure of personal consumption in households in whole or by individual socio economic categories. The data of the survey are used to calculate weights of the consumer price index, for national accounts and other derivative accounts related to consumption by the population.
Selection of observation units
An observation unit is a selected single person or multi person household. A household is any family or other group of persons living together and sharing their income to cover the basic costs of living (accommodation, food, other consumer goods, etc.), regardless of whether all members live all the time at the place where the household is located, or some of them live for a longer time at another place in Slovenia or abroad due to employment, schooling, or of any other reasons. Also, a person is considered a household who lives alone at the place of survey and has no other household at another place (single-person household). Such a person can live alone in a separate dwelling or together with other persons in the same dwelling (same room) if not sharing his/her income with such persons.
The following persons are also considered members of a household:
who are absent due to occupational commitments, but do not have a dwelling or household at another place (e.g. travelling salesmen, businessmen, etc.);
persons under temporary employment contracts abroad who return home each month or more frequently;
who are not family members, if they work, eat and live in the same housing community (servants and permanent workers on private agricultural holdings);
university or secondary school students who attend schools at another place, regardless of the time they spend outside the household (in schools and at study);
The following persons are not considered members of a household:
who eat in the household (boarders);
university or secondary school students who live and eat, or only live, in the household under survey (roommates, subtenants);
who are under temporary employment contracts abroad and return home only occasionally; the money and material goods intended for the household by such persons are considered as gifts from persons outside the household;
migrant workers who live permanently abroad;
Quality report 4/22
who live in institutions – collective households (boarding schools, nursing homes for children, old people’s homes, hospitals, cloisters, homes for pupils, student hostels).
Based on the statements of the persons present, and considering the definition of a household, an interviewer establishes who the members of the household are.
The Household Budget Survey is a sample survey. The sample frame is the Central Population Register (CPR). The sample stratification was made with regard to 12 statistical regions and six types of settlements. The sampling is done in two stages. In the first stage sampling units were selected (made up of one or more spatial districts). In the second stage six people were selected in each sampling unit. We do not use the method of substitution, i.e. of selecting substitute households that would replace the ones that did not co-operate. The survey does not cover collective households such as boarding schools, nursing homes for children, old people’s homes, hospitals, boarding schools, etc.
From 1997 up to 2011, we carried out surveys on small annual samples (about 1,800 households). The published data were based on a sample of three consecutive years (e.g. 2008, 2009 and 2010). These data were calculated to the middle year (2009), and this year was used as the reference year. By 2012, we changed the methodology and moved to larger samples. In 2012 the sample included 7,000 and in 2015 and 2018 7,400 households.
Sources and methods of data collection
Survey data are collected by the Household Budget Survey (HBS). The data in this survey are produced by:
A personal interview based on the Household Budget Survey (HBS) questionnaire
Diaries, in which household members record daily expenditure for 14 days
Administrative and other databases
Own resources: The structure of the population (DEM-PREB/ČL)
The household survey is distributed throughout the year; each household participates in the survey for 14 days. In the questionnaires, some basic data are entered for all members of the household; then the survey is completed for members younger than 15 years. The survey is continued by posing questions to members of at least 15 years of age which refer to data on themselves, their work habits, and also travel and personal income. Questions intended for the household as a whole concern the dwelling and any eventual other accommodations of the household, motor vehicles, furnishing, household equipment and maintenance, clothing and footwear, expenses for children, education, and help to other households, and other expenses, money transfers and gifts, the financial situation of the household, the total income of the household, and own production consumed in the household. In diaries, members of the household enter all daily expenses (description of a product or service, quantity bought, and price). The main purpose of completing the diary is to cover
Quality report 5/22
expenses which households would not remember in such detail when completing the questionnaire (food and beverages).
Administrative and other databases: the Financial Administration (income tax), the Ministry of the Interior (CPR, Records of Households), the Ministry of Labour, Family, Social Affairs and Equal Opportunities (social and family benefits, scholarships), the Employment Service of Slovenia (compensation for unemployment), and the Surveying and Mapping Authority (REN).
Key variables
The key variables are the consumption expenditure of households(such as on cold water, electricity, gas, etc.) and some other expenditure, and quantities of food and beverages purchased. Based on the data on available assets (salary of a person, pension, and other receipts), households are classified in income classes.
Key statistics
The key variables are:
Average annual allocated assets per household
Average annual allocated assets per household member
Structure of average annual allocated assets of households
Average annual quantity of food and beverages consumed per household member
Share of households with certain consumer durables
Questionnaire
Questionnaire (only in Slovene): Raziskovanje o porabi v gospodinjstvih (APG), theme: Quality of life, sub-theme: Household expenditure
https://www.stat.si/StatWeb/en/Methods/QuestionnairesMethodologicalExplanationsQualityReports
Methodological explanations
The methodological explanations are available on the website:
https://www.stat.si/StatWeb/File/DocSysFile/8350/08-112-ME.pdf
Quality report 6/22
1 RELEVANCE
Relevance describes to what extent statistical data satisfy the needs of users. The rate is determined by whether all statistical data the users need are available and to what extent the published data and concepts used (definitions, classifications) meet the needs of users.
1.1 DESCRIPTION OF ADMINISTRATIVE SOURCES USED
1.1.1 Origin and original purpose of the data
Administrative and other databases: the Financial Administration (income tax), the Ministry of the Interior (CPR, Records of Households), the Ministry of Labour, Family, Social Affairs and Equal Opportunities (social and family benefits, scholarships), the Employment Service of Slovenia (compensation for unemployment), and the Surveying and Mapping Authority (REN).
1.1.2 Method of takeover
Data are collected from databases or files, which are obtained and edited by the content administrators of each administrative source at SURS.
1.2 USERS OF SURVEY DATA
1.2.1 Key users of survey data
Table 1.1: Key users of survey data
Public sector Ministries
Business entities legal entities
Science, research and education
National Institus of Public Health, Institute for Economic Research, libraries, faculties, students
General public yes
Media radio and television houses, printed media, the Slovene Press Agency
Foreign users
Eurostat, OECD, DG Energy, statistical offices of other countries, UNICEF, Luxembourg Income Study, researchers
Internal users national accounts, price statistics
1.2.2 User needs and satisfaction
The main institute for communication between users and the Statistical Office is the e-mail or phone and the Statistical Advisory Committee on Level of Living. The Advisory Committee includes, in addition to representatives of the Office, representatives of the Ministry of Labour, Family and Social Affairs, the Institute of Macroeconomic Analysis and Development of the Republic of Slovenia, the
Quality report 7/22
Institute of Public Health of the Republic of Slovenia, the Health Insurance Institute of Slovenia, the Employment Service of Slovenia, the Institute for Economic Research, the Institute for Pension and Disability Insurance, the Bank of Slovenia, the Association of Social Institutions of Slovenia, the Slovene Institute of Social Protection, Ministry of Agriculture, Forestry and Food and Government Office for Development and European Affairs
SURS measured general user satisfaction for the last time in 2018. Respondents assessed general satisfaction with SURS with the average score of 7.5 (on a scale from 1 – disagree completely to 10 – agree completely).
1.3 COMPLETENESS OF STATISTICAL RESULTS
1.3.1 Completeness of statistical results
Completeness of statistical results is the ratio between the number of statistical results that were disseminated (within a specific field) and the number of statistical results that were demanded (e.g. with regulations, in agreements). Statistical results that are not appropriate for a Member State or derogation is in force for them are not taken into account in the calculation.
Completeness of statistical results is 100 %.
1.3.2 Explanations
All results are calculated.
2 ACCURACY
Accuracy is defined as the degree to which the value at the end of statistical processing matches the true but unknown population value.
2.1 STANDARD ERRORS
2.1.1 Procedure for calculating standard errors
Standard error was estimated with the use of Taylor linearisation.
2.1.2 Standard error
Standard error is the square root of the variance of the statistical estimate. To a large extent it is determined by the sampling error. It is also affected by other random errors in the survey implementation. The standard error can be shown also in the form of the coefficient of variation or the confidence interval.
Quality report 8/22
Table 2.1: Standard error
Reference period
Domain name
Domain value Statistics Coefficient of variation %
2015 Household expenditure
Total allocated assets average per household (EUR)
1.89
2018 Household expenditure
Total allocated assets average per household (EUR)
1.33
2015 Household expenditure
00_Consumption expenditure average per household (EUR)
1.19
2018 Household expenditure
00_Consumption expenditure average per household (EUR)
1.22
2015 Household expenditure
01_Food and non-alcoholic beverages
average per household (EUR)
1.05
2018 Household expenditure
01_Food and non-alcoholic beverages
average per household (EUR)
1.10
2015 Household expenditure
02_Alcoholic beverages and tobacco average per household (EUR)
3.26
2018 Household expenditure
02_Alcoholic beverages and tobacco average per household (EUR)
2.99
2015 Household expenditure
03_Clothing and footwear average per household (EUR)
2.51
2018 Household expenditure
03_Clothing and footwear average per household (EUR)
2.22
2015 Household expenditure
04_Housing, water, electricity, gas and other fuels
average per household (EUR)
1.06
2018 Household expenditure
04_Housing, water, electricity, gas and other fuels
average per household (EUR)
1.22
2015 Household expenditure
05_Furnishings, household equipment and routine household maintenance
average per household (EUR)
3.85
2018 Household expenditure
05_Furnishings, household equipment and routine household maintenance
average per household (EUR)
3.39
2015 Household expenditure
06_Health average per household (EUR)
3.73
2018 Household expenditure
06_Health average per household (EUR)
3.70
2015 Household expenditure
07_Transport average per household (EUR)
2.62
2018 Household expenditure
07_Transport average per household (EUR)
2.96
2015 Household expenditure
08_Communication average per household (EUR)
1.18
2018 Household expenditure
08_Communication average per household (EUR)
1.07
2015 Household expenditure
09_Recreation and culture average per household (EUR)
2.02
2018 Household expenditure
09_Recreation and culture average per household (EUR)
2.29
2015 Household expenditure
10_Education average per household (EUR)
5.55
2018 Household expenditure
10_Education average per household (EUR)
5.20
2015 Household expenditure
11_Restaurants and hotels average per household (EUR)
3.30
2018 Household expenditure
11_Restaurants and hotels average per household (EUR)
3.00
2015 Household expenditure
12_Miscellaneous goods and services
average per household (EUR)
1.59
2018 Household expenditure
12_Miscellaneous goods and services
average per household (EUR)
1.29
2015 Household expenditure
20_Other expenditure (not part of consumption expenditure)
average per household (EUR)
12.35
2018 Household expenditure
20_Other expenditure (not part of consumption expenditure)
average per household (EUR)
6.20
Quality report 9/22
2.1.3 Explanations
In calculating average expenditures per household, all households are always taken into account, not just those that had particular expenditures. If the table contains estimated population totals of (continuous) variables (average allocated assets / quantities per household/household member), publishing limitations are determined by the relative standard errors or the coefficients of variation (CV). In such cases it holds: If the coefficient of variation (CV) of the estimate is - 10% or below (CV <= 10%) the estimate is of acceptable precision and is published without limitations; - in the interval from 10% and up to 30% (10% < CV <= 30%) the estimate is less precise and is flagged for caution with letter M; - over 30% (CV > 30%), the estimate is too imprecise to be published and therefore suppressed for use by letter N.
2.2 COVERAGE BIAS
2.2.1 Procedure for calculating the bias
2.2.2 Coverage bias
Coverage bias measures the error in a statistical estimate caused by the fact that a part of the population was left out of the observation on purpose. Bias due to other factors, such as non-response, measurement errors, processing errors, etc., is not estimated.
2.2.3 Explanations
The survey is not conducted on the basis of threshold sampling.
2.3 NON-SAMPLING ERRORS
2.3.1 Non-response errors
2.3.1.1 Unit non-response rate
The unit non-response rate is the proportion of eligible units for which we were not able to obtain any desired data or the obtained data were not useful. Unweighted and weighted values of the indicator can be calculated.
Table 2.2: Unit non-response rate
Reference period
Domain name
Domain value
Number of non-responses
Number of eligible units
Non-response rate (in %)
2012 SKUPAJ - 3102 6765 45.85
2015 SKUPAJ - 3487 7237 48.18
2018 SKUPAJ - 3711 7243 51.24
Quality report 10/22
Table 2.3: Weighted unit non-response rate
Reference period Domain name Domain value Non-response rate (in %)
2012 SKUPAJ - 45.17
2015 SKUPAJ - 47.05
2018 SKUPAJ - 49.02
2.3.1.2 Explanations
The survey is carried out by personal interviews with the help of interviewers. Unit (households) non-response may occur due to poor work of interviewers, but the respondent may refuse to participate in the survey, despite the motivation to participate. In these cases, missing household responses are not imputed. By weighting, we want to achieve the representativeness of the sample, so that the weighted data is the best possible estimate of the observed population at a certain time point. The weighting procedure was determined based on the sample design, unit non-response, and the availability of auxiliary population variables used for calibration. The following auxiliary variables are used in weighting: statistical regions, settlement types, gender, age, survey quarter. The final weight o is the product of the weights for the probability of unit selection, the weights for non-response and the calibration factor.
2.3.1.3 Item non-response rate
The item non-response rate is the proportion of units for which we were not able to obtain data for a specific variable, even though the unit was eligible for this variable. In using administrative data the value of the indicator is the rate of unsuccessful linking of the variable showing the proportion of units for which we were not able to link units using one or more administrative sources and thus determine the variable value.
Table 2.4: Item non-response rate
Reference period
Domain name Domain value
Variable Item non-response rate (in %)
2012 SKUPAJ - Expenditure for petrol 2.66
2015 SKUPAJ - Expenditure for petrol 1.83
2018 SKUPAJ - Expenditure for petrol 2.37
2012 SKUPAJ - Tax for land 11.74
2015 SKUPAJ - Tax for land 10.33
2018 SKUPAJ - Tax for land 9.08
2012 SKUPAJ - Expenditure for cold water 2.10
2015 SKUPAJ - Expenditure for cold water 2.90
2018 SKUPAJ - Expenditure for cold water 4.46
2012 SKUPAJ - Annual expenditure for registration of the vehicle
5.96
2015 SKUPAJ - Annual expenditure for registration of the vehicle
4.85
2018 SKUPAJ - Annual expenditure for registration of the vehicle
4.92
Table 2.5: Weighted item non-response rate
Reference period
Domain name Domain value
Variable Item non-response rate (in %)
2012 SKUPAJ - Expenditure for petrol 2.45
2015 SKUPAJ - Expenditure for petrol 1.78
2018 SKUPAJ - Expenditure for petrol 2.17
2012 SKUPAJ - Tax for land 10.80
2015 SKUPAJ - Tax for land 9.85
2018 SKUPAJ - Tax for land 9.62
Quality report 11/22
Reference period
Domain name Domain value
Variable Item non-response rate (in %)
2012 SKUPAJ - Expenditure for cold water 2.20
2015 SKUPAJ - Expenditure for cold water 3.52
2018 SKUPAJ - Expenditure for cold water 4.89
2012 SKUPAJ - Annual expenditure for registration of the vehicle
5.21
2015 SKUPAJ - Annual expenditure for registration of the vehicle
4.63
2018 SKUPAJ - Annual expenditure for registration of the vehicle
5.19
2.3.1.4 Explanations
The survey is carried out with the help of a computer-assisted personal interviews, and diaries are kept by household members themselves. In some cases, questionnaires or diaries are poorly completed. This can happen due to the poor work of the interviewers or the respondent may not know or do not want to answer the question. In these cases, we need to impute the missing answers.
2.3.1.5 Imputation rate
The imputation rate is the ratio between the number of units for which the data for the key variable were imputed (for any reason) and the number of units for which we have at least a datum for the variable (after statistical processing). Unweighted and weighted values of the indicator can be calculated. The indicator is calculated only for key variables.
Table 2.6: Rate of unsuccessful integration of variables
Reference period
Domain name
Domain value
Variable Imputation rate (in %)
2012 SKUPAJ - Expenditure for petrol 3.06
2015 SKUPAJ - Expenditure for petrol 1.12
2018 SKUPAJ - Expenditure for petrol 1.39
2012 SKUPAJ - Tax for land 10.16
2015 SKUPAJ - Tax for land 8.59
2018 SKUPAJ - Tax for land 7.67
2012 SKUPAJ - Expenditure for white bread 12.83
2015 SKUPAJ - Expenditure for white bread 29.07
2018 SKUPAJ - Expenditure for white bread 28.53
2012 SKUPAJ - Expenditure for cold water 9.31
2015 SKUPAJ - Expenditure for cold water 5.28
2018 SKUPAJ - Expenditure for cold water 66.11
2012 SKUPAJ - Expenditure for chicken meat 14.02
2015 SKUPAJ - Expenditure for chicken meat 29.61
2018 SKUPAJ - Expenditure for chicken meat 30.78
2012 SKUPAJ - Annual expenditure for registration of the vehicle
5.62
2015 SKUPAJ - Annual expenditure for registration of the vehicle
4.72
2018 SKUPAJ - Annual expenditure for registration of the vehicle
4.98
Table 2.7: Weighted rate of unsuccessful integration of variables
Reference period
Domain name
Domain value
Variable Imputation rate (in %)
2012 SKUPAJ - Expenditure for petrol 5.02
2015 SKUPAJ - Expenditure for petrol 1.76
2018 SKUPAJ - Expenditure for petrol 2.54
2012 SKUPAJ - Tax for land 10.95
2015 SKUPAJ - Tax for land 9.91
2018 SKUPAJ - Tax for land 9.62
Quality report 12/22
Reference period
Domain name
Domain value
Variable Imputation rate (in %)
2012 SKUPAJ - Expenditure for white bread 12.27
2015 SKUPAJ - Expenditure for white bread 32.08
2018 SKUPAJ - Expenditure for white bread 29.80
2012 SKUPAJ - Expenditure for cold water 10.83
2015 SKUPAJ - Expenditure for cold water 6.47
2018 SKUPAJ - Expenditure for cold water 62.69
2012 SKUPAJ - Expenditure for chicken meat 14.05
2015 SKUPAJ - Expenditure for chicken meat 31.28
2018 SKUPAJ - Expenditure for chicken meat 33.41
2012 SKUPAJ - Annual expenditure for registration of the vehicle
5.75
2015 SKUPAJ - Annual expenditure for registration of the vehicle
5.36
2018 SKUPAJ - Annual expenditure for registration of the vehicle
6.11
2.3.1.6 Explanations
For the imputation of the missing data in the questionnaires, we use different methods: the hot-deck method for discrete variables and the average value method for continuous variables. Households surveyed are divided into classes according to different variables (most often by household type and income). Within these classes we calculate the average values of the variables used for missing values. The high proportion of imputed data is usually the consequence of the nature of the question with which we collect this data. Example: households pay for water, refuse collection and sewage together and cannot divide this expenditure separately into individual expenditures. Thus, in the data processing phase, the total amount is divided into individual expenditures by imputation.
2.3.2 Coverage errors
2.3.2.1 Overcoverage rate
The overcoverage rate is the proportion of units in the sample frame that are not part of the target population (ineligible units). If the survey is conducted based on the sample, the overcoverage rate is estimated based on information from the sample, meaning that the weighted value of the indicator is calculated and disseminated.
Table 2.8: Overcoverage rate
Reference period
Domain name
Domain value
Number of ineligible units in the frame
Number of all units in the frame
Overcoverage rate (in %)
2012 SKUPAJ - 35317 1610802 2.19
2015 SKUPAJ - 35517 1672151 2.12
2018 SKUPAJ - 32211 1654600 1.95
2.3.2.2 Explanations
Units that represent over-coverage are units that are ineligible for research and were sampled for incorrect or missing information. If we have selected a person who lives abroad or has died, our household is ineligible for the survey. Thus, we have about 2% of non-compliant units in the survey.
Quality report 13/22
2.3.2.3 Undercoverage errors
Under coverage errors occur when certain units are not included in the sample, but should have been. Dwellings and households that live in such dwellings, in which no person has his/her permanent residence registered, are thus not included in the sampling frame. Homeless people are not included in the sample. An undercoverage error also occurs if we cannot determine the address of the selected person or if the selected person does not live at the address we have in the sample frame.
2.3.3 Measurement errors
2.3.3.1 Error detection controls
In the survey, the questionnaire as the source of measuring the consumption in households is prioritised over diaries. When the same code appears in the questionnaire and in the diary, the code in the questionnaire should therefore be considered, while the code in the diary should be eliminated.
The survey is conducted by means of a laptop, which enables the regular control of the entered data. The programme has logical data controls embedded so that certain data are checked on the basis of preliminary set rules. Incorrect data entry is prevented by means of so called light and heavy logical controls. Upon entry of a value which is not very probable, a warning for the interviewer appears on the monitor that the value he/she has entered is not likely, although possible. In the case of such a mistake, the interviewer corrects the value entered, and if the value is correct, he/she need only confirm it (light control). When an illogical error occurs (e.g. if the interviewer tries to enter an inappropriate answer), a warning appears on the monitor that it is a serious error. The interviewer cannot continue data entry as long as he/she does not eliminate the error (heavy control).
Error code Course - control Comments
Heavy control A5. MESEC (month of birth)
Possible entry: 01-12
Light control 10<=cost of electricity<=250
Did you realy pay less then 10 EUR or more than 250 EUR in the last month?
2.3.3.2 Reasons for measurement errors
Major errors appear when data are collected in the field. They result from the interviewer's incomprehension or misinterpretation of questions. Errors may also occur due to a negative, disinterested attitude, or poor knowledge on the part of the persons interviewed. The selected persons participate in the survey on a voluntary basis and are therefore not obliged by law to provide credible data.
Immediately after receipt, we examine the questionnaires and warn the interviewer of the errors which occurred during interviewing. When in individual cases the questionnaire was poorly completed or the answers were unclear, we call the interviewer and ask him/her to pay the selected person another visit. If
Quality report 14/22
logical errors in the entry procedure were found (light of heavy), we correct or supplement the questionnaire in accordance with the answers which the selected person provided to other questions.
2.3.3.3 Editing rate
The editing rate is the ratio between the number of units for which we corrected the values of a certain variable in the editing phase and the number of all units that reported the data for the variable. Unweighted and weighted values of the indicator can be calculated. The indicator is calculated only for key variables.
Table 2.9: Editing rate
Reference period
Domain name
Domain value
Variable Editing rate (in %)
2012 SKUPAJ - Expenditure for petrol 0.00
2015 SKUPAJ - Expenditure for petrol 0.00
2018 SKUPAJ - Expenditure for petrol 0.00
2012 SKUPAJ - Tax for land 0.00
2015 SKUPAJ - Tax for land 0.00
2018 SKUPAJ - Tax for land 0.00
2012 SKUPAJ - Expenditure for white bread 0.00
2015 SKUPAJ - Expenditure for white bread 0.00
2018 SKUPAJ - Expenditure for white bread 0.00
2012 SKUPAJ - Expenditure for chicken meat 0.00
2015 SKUPAJ - Expenditure for chicken meat 0.00
2018 SKUPAJ - Expenditure for chicken meat 0.00
2012 SKUPAJ - Annual expenditure for registration of the vehicle
0.00
2015 SKUPAJ - Annual expenditure for registration of the vehicle
0.32
2018 SKUPAJ - Annual expenditure for registration of the vehicle
0.62
Table 2.10: Weighted editing rate
Reference period
Domain name
Domain value
Variable Editing rate (in %)
2012 SKUPAJ - Expenditure for petrol 0.00
2015 SKUPAJ - Expenditure for petrol 0.00
2018 SKUPAJ - Expenditure for petrol 0.00
2012 SKUPAJ - Tax for land 0.00
2015 SKUPAJ - Tax for land 0.00
2018 SKUPAJ - Tax for land 0.00
2012 SKUPAJ - Expenditure for white bread 0.00
2015 SKUPAJ - Expenditure for white bread 0.00
2018 SKUPAJ - Expenditure for white bread 0.00
2012 SKUPAJ - Expenditure for chicken meat 0.00
2015 SKUPAJ - Expenditure for chicken meat 0.00
2018 SKUPAJ - Expenditure for chicken meat 0.00
2012 SKUPAJ - Annual expenditure for registration of the vehicle
0.00
2015 SKUPAJ - Annual expenditure for registration of the vehicle
0.37
2018 SKUPAJ - Annual expenditure for registration of the vehicle
0.42
2.3.3.4 Explanations
The share of data editing is small because the controls are already built-in CAPI questionnaire.
Quality report 15/22
2.3.3.5 Coherence of data sources
The indicator is calculated when we have the data for at least some units of the selected variable from two sources and both sources are used as data sources for this variable. The value of the indicator shows coherence of data from these two sources.
2.3.3.6 Explanations
The index calculation is not applicable.
3 TIMELINESS AND PUNCTUALITY
Timeliness of the release measures the time lag between the period to which the data refer and the release date. Punctuality of the release measures the time lag between the actual and previously announced date of data release, which is determined in the release calendar. If the mentioned dates coincide, the release is punctual.
3.1 TIMELINESS
3.1.1 Timeliness of the first release
Timeliness of the first release measures the time lag between the date of the first release of the data and the reference date of the observed (reference) period. The reference date is usually the last day of the (reference) period to which the published data refer, but it can also be another key date within this period (e.g. start of the school year).
Table 3.1: Timeliness of the first results
Reference period End of the reference period Date of publishing Time lag (in months)
2010 31.12.2010 25.7.2012 19
2012 31.12.2012 28.10.2013 10
2015 31.12.2015 6.10.2016 9
2018 31.12.2018 8.10.2019 9
3.1.2 Timeliness of the final results
Timeliness of final results measures the time lag between the date of the release of final data and the reference date of the observed (reference) period. The reference date is usually the last day of the (reference) period to which the published data refer, but it can also be another key date within this period (e.g. start of the school year).
Quality report 16/22
3.1.3 Explanations
The index calculation is not applicable (timeliness of the final results). Timeliness of the final results is equal to timeliness of the first results.
3.2 PUNCTUALITY
3.2.1 Punctuality of the first release
Punctuality of the first release measures the time lag between the announced and actual date of the first release.
Table 3.2: Punctuality of the first results
Reference period Announced date Date of publishing Time lag (in days)
2010 25.7.2012 25.7.2012 0
2012 27.12.2013 28.10.2013 -60
2015 6.10.2016 6.10.2016 0
2018 8.10.2019 8.10.2019 0
3.2.2 Reasons for deviations in punctuality
In 2012 we revised the survey and that was the reason for deviation in the date of publishing the results.
4 ACCESSIBILITY AND CLARITY
Accessibility of statistical data describes the possibilities available to users for simple access to statistical data. It refers to physical circumstances in which the data are available to users: where and how the data can be obtained, when they will be available, how much does the service cost (clear price list of services), conditions for using the data (copyright), availability of microdata and metadata, availability in various formats. Clarity of statistical data describes how simple it is for users to understand the data. It refers to the information environment in which the data are presented: are the data equipped with appropriate methodological explanations and are they properly presented with graphical presentations or other material, is information on punctuality of data and on limits to use available to the users, do the users have access to additional information should they need it.
4.1 ACCESSIBILITY
4.1.1 Frequency of publication
Data are publishend monthly till 2010 and afterwards every three years (2012, 2015, 2018).
Quality report 17/22
4.1.2 Means used for dissemination
Table 4.1: Means used for dissemination
Means Used
Web release yes
Tables in the SiStat Database yes
Publication in interactive tools and applications no
Printed publications yes
Publication on social media yes
International databases (Eurostat database, OECD database) yes
Microdata of the statistical survey yes
Metadata yes
4.2 CLARITY
4.2.1 Disseminated results
The results of the survey are displayed in the form of absolute numbers, structures and shares. In addition to tables, they are displayed in the form of graphs and infographics.
4.2.2 Level (detail) of dissemination
Data are shown for Slovenia as a whole. The sample is too small to enable the calculation and publication of results for lower spatial units.
Data on consumption expenditure are published according to the European Classification of Individual Consumption by Purpose (ECOICOP) and according to:
household income class,
type of household,
number of household members,
main source of funds available.
5 COMPARABILITY
Comparability of statistics measures the differences due to the use of different statistical concepts (classifications, definitions, target population) or different statistical methods in calculating statistics in different geographical areas, other domains of the population or different time periods.
Quality report 18/22
5.1 COMPARABILITY OVER TIME
5.1.1 Time coverage
Data are collected with the Household Budget Survey, between 1954 and 1997with the results that are more or less comparable.
In 1997 there were some bigger changes in survey methodology. The surveywas harmonised with Eurostat’s recommendations and the concept ofexpenditure classified by the COICOP (Classification of Individual Consumptionby Purpose) was implemented. This enables international comparison of theresults. Comparable data series is available from 1997 onwards.
5.1.2 Length of comparable time series
The indicator shows the length of the time series since the last break in the time series, i.e. the number of points in the time series since the last break.
5.1.3 Explanations
5.1.4 Factors influencing comparability over time
Data are collected with the Household Budget Survey, which was between 1983 and 1997 conducted according to the unified methodology and with the same questionnaires as regards the contents.
In 1997 there were some changes in survey contents and implementation. The survey was harmonised with Eurostat’s recommendations and became a continuous one. By combining data of three consecutive years we obtained more accurate estimates. Data from three years were calculated to the middle year, which was used as the reference year for the interpretation of results. We no longer use the balance approach but the concept of expenditure classified by the COICOP-HBS (Classification of Individual Consumption by Purpose). Comparison of results with previous years is only possible at the level of current income and consumption expenditure, since certain questions are no longer part of the survey (e.g. on decrease of savings, paying back consumer credits and investment loans, etc.). Therefore, there is no balance between available assets and allocated assets. The latest available data according to this methodology are for 2010.
In 2012, the survey was again revised. The content of the questionnaire was expanded (maintenance of housing, education, household equipment, etc.). Questions on household income were excluded from the questionnaire and the data were obtained from administrative and registry data. At the same time also the data processing was revised. Consumption expenditure is classified by the newer version of the harmonized classification COICOP (5-digit), which ensures comparability of data between different users of the 5-digit code. The survey consists of two parts or sources: the survey and the register and administrative
Quality report 19/22
data. The data that refer to income and some other information are gathered from the existing registers and administrative records and we link them to the survey results. The use of registers and administrative sources not only facilitates the diminishing of the burden on the respondents, but it also helps cut down the survey expenses.
In 2015 there were no major changes in the survey.
In 2018, monitoring of consumption expenditure remained the same. In the part where we monitor other expenditures (group 20 Other expenditures), i.e. purchase of a dwelling, major works and renovations, other expenditure, we excluded monitoring of the purchase and sale of dwelling or land.
5.1.5 Seasonal adjustment
Seasonal adjustment was not performed.
5.2 GEOGRAPHICAL COMPARABILITY
5.2.1 Comparability with other international organisations
Carrying out HBS is not governed by the regulations of the European Commission. It is conducted in all EU Member States, although according to different methodologies. The differences concern mostly the extent of data collected and the method of carrying out the survey. Therefore, the results are not fully comparable between different countries. Eurostat collects data on the basis of the Gentlemen’s Agreement (for every 5 years); therefore, different countries gradually harmonise the survey with Eurostat requirements.
The results are available in Eurostat database: http://epp.eurostat.ec.europa.eu/portal/page/portal/statistics/search_database -> Population and social conditions -> Living conditions and welfare (livcon) -> Consumption expenditure of private households (hbs)
6 COHERENCE
Coherence in statistics is the adequacy of statistical data to be reliably combined in different ways and for various users. It describes limitations in combining statistics from different sources that are the result of using different statistical procedures.
6.1 COHERENCE BETWEEN PROVISIONAL AND FINAL DATA
6.1.1 Policy of releasing provisional data
Provisional data are not disseminated. Only final data are published.
Quality report 20/22
6.1.2 Coherence between provisional and final data
Coherence between provisional and final data shows absolute or relative difference between the disseminated value at first release and the value at the release of final data. Revisions are only corrections that are part of the regular procedure of publishing statistical results. Corrections due to errors are not revisions and are not taken into account in calculating the indicator. Even though the revision policy stipulates several versions of (provisional) data for the same reference period, for the sake of simplicity only the difference between the values of provisional data in the first release and the values in the final data release is calculated.
6.1.3 Reasons for larger differences between provisional and final data
Provisional data are not disseminated. Only final data are released.
6.2 COHERENCE WITH THE RESULTS OF THE REFERENCE SOURCE
6.2.1 Brief description of the reference source
The main reference source for comparing data on income is the Statistics on Income and Living Condition (EU-SILC). For comparison data on some durable consumer goods are also available from this survey.
6.2.2 Coherence with reference sources
Coherence with reference sources shows absolute or relative difference in view of the results of the selected reference survey in which the same or at least a related and therefore a comparable phenomenon is observed but which can be observed with a different methodology or a different periodicity. Results of short-term surveys can be compared with results of structural surveys, and vice versa. Results of surveys can also be compared with results disseminated by national accounts or various institutions in the country or even abroad (e.g. mirror statistics on tourism).
Table 6.1: Coherence with reference data
Reference period
Domain name
Domain value
Statistics Difference
2018 SKUPAJ - Total disposable household income (EUR) 1366.00
2018 SKUPAJ - Value of goods produced by own consumption (EUR) 103.00
2018 SKUPAJ - Regular inter – household cash transfer received – net (EUR) 34.00
2018 SKUPAJ - Property income (EUR) 17.00
2018 SKUPAJ - Regular taxes on wealth net (EUR) 3.00
2018 SKUPAJ - Family/Children related allowances net (EUR) 9.00
2018 SKUPAJ - Regular inter – household cash transfer paid – net (EUR) 5.00
2018 SKUPAJ - Share of households according to how they cope with their disposable income (%)
2.00
2018 SKUPAJ - Personal computer (%) 3.00
2018 SKUPAJ - Washing machine (%) 2.00
2018 SKUPAJ - Car (%) 2.00
2018 SKUPAJ - TV (%) 1.00
Quality report 21/22
6.2.3 Reasons for larger differences between sources
The data source for income in the HBS and EU-SILC surveys are different (households and administrative sources), which is the main difference in the data. In HBS all income is monitored as net amounts, so comparable net income is also taken from EU-SILC.
7 BURDENS
Burden of interviewed persons and business entities is not a separate quality component but an important factor in assessing the quality, since it usually has an impact on all other quality components.
7.1 BURDEN OF REPORTING UNITS
Table 7.1: Burden of the reporting units
Reference period
Total time spent on reporting data (in minutes)
Number of completed questionnaires
Time spent on completing a questionnaire (in minutes)
2018 135234 3532 38.00
7.2 EXPLANATIONS
Measures to reduce the burden on respondents are expressed in the concern for the shortest and most comprehensible questionnaire, because in this way the time of the survey is shortened.
8 EVALUATION OF SURVEY QUALITY
At the end of survey implementation, quality indicators and other important information resulting from the survey implementation are reviewed. On this basis survey implementation is evaluated, weak points regarding survey implementation are identified and improvements for the next implementation are planned.
8.1 GENERAL QUALITY ASSESSMENT
The survey is carried out on the field, and the data collected cover the total household expenditure. Due to the heavy burden to the respondents, about half of them decide to participate, which affects the quality of the data. Therefore, in 2012, we introduced precision of the statistical estimates to alert users of statistics to draw attention to less precise estimates by flagging them with a special sign or by not publishing them at all.
Quality report 22/22
8.2 MEASURES FOR IMPROVEMENT
The Statistical Office of the Republic of Slovenia endeavours to inform households on the importance of participating in the survey. We provide information letters to the addresses of all selected persons, informing them that the household at their address was selected for the survey, and providing an explanation of the purpose of the survey. The households also receive an information leaflet with the presentation of the survey and examples of published results from the survey. If the household did not receive the letter, this can be delivered by interviewers, or they can inform the Statistical Office of the name and address of the person to whom the letter should be sent. Measures to reduce the burdens of respondents are indicated as a concern for the questionnaire, which should be as short and clear as possible, in order to significantly reduce interview time. Therefore in the renewed survey we plan to exclude the questions about household income for different categories and use the administrative sources instead.
Coverage errors are established by well-trained interviewers, so as to avoid problems distinguishing ineligible units from non-contacted units or units which are absent. Prior to interviewing, each interviewer must attend a training session where he/she becomes acquainted with the basic methodological characteristics of the survey and with practical examples, and trained to work with a laptop. Interviewers are provided with exhaustive instructions which contain methodological instructions and practical examples for completing the survey. Since the survey is continuous, it requires permanent interviewers who are well acquainted with the survey and know how to approach households. They also have a positive attitude towards the survey. Meetings with the interviewers resulted in a decreased trend of nonresponses; at the same time we call their attention to errors we have noticed when analysing data.
We believe that the majority of measurement errors occur during the phase of data collection and entry; therefore we implement numerous measures to limit them. In the programme for data entry we prepare the text, controls and skips, so as to enable as fluent and correct interviewing as possible. The selection of a wrong answer is also considered as an error of the interviewer, thus this is not very likely since questions are inter-related and an error may easily be noticed. Influencing the answers of the persons interviewed also constitutes interviewer error; solving this problem is not easy.
In 2012 the survey was revised in order to get even better results and special effort was put to clarity with extended methodological explanation, detailed comment of the results, precision of published estimates , new tables in data portal.
In the future, we want to modernize our data collection and introduce modern modes of collecting the data (mobile applications, web diaries, etc.) and motivate them even more for the participation in the survey.