Household budget surveys have quite a long history in Estonia. The first survey of the kind was conducted in Estonia in autumn 1925. The survey, called the budget survey at that time, was conducted by the State Central Bureau of Statistics among the workers’ households of Tallinn. The similar survey of a smaller scope was conducted also in 1926. The later surveys were conducted in 1937-1938, 1952, 1960-1978, 1995-1999.
The aim of the 2010 Estonia Household Budget Survey is to get reliable information on the expenditures and consumption of households. Besides obtaining data about the household composition, the survey also provides information on household members’ main demographic and social indicators (marital status, employment, education), as well as on living conditions and owning of durable goods. The data of the survey are used a lot by ministries and research institutions.
Since 2000 the HBS consisting of four parts has been rather voluminous. The Household Picture concerns general data about the household’s background data such as sex, age, marital status, education, coping, employment, etc. of household members. Post-Interview is intended for registering the changes entered during the survey. The Diary Book for Food Expenditure reflects the expenditure made by the household during half a month. The Diary Book for Income, Taxes and Expenditure contains data about monetary and non-monetary income received by the household as well as the expenditures on all commodities and services.
Kind of Data
Sample survey data [ssd]
Unit of Analysis
- v01: Edited, anonymous dataset for public distribution.
The scope of the study includes: Sex, age, marital status, education, coping, employment, etc. of household members; income, taxes and expenditure.
Producers and sponsors
The population of the Household Budget Survey was made up of all permanent residents of the Republic of Estonia aged 15 or older as of 1 January 2010, who live in private households, excl. those residing in institutions on a long-term basis (at least for a year). The Estonian Population Register, administered by the Ministry of Internal Affairs, was used as a sampling frame representing the survey population.
The HBS is a sample survey i.e. the population is evaluated on the basis of the data collected from the sample. The survey sample was drawn from among the persons registered in the Population Register who were 15 years of age or older as at 1 January 2009. The person included in the sample (address person) brought his/her household into the sample.
Sample persons were drawn from the Population Register by the stratified unproportional systematic sampling procedure. In case of this sampling procedure, the population is divided into non-overlapping subpopulations or strata, and independent subsamples are drawn separately from every subpopulation following the systematic sampling procedure and by applying different inclusion probabilities. The population was stratified by the county in which the address person's place of residence was. In the stratification procedure, the stratification principles worked out for and applied to the Estonian Social Survey, which has been carried out on an annual basis since 2004, were used, and thus three strata were formed by the number of inhabitants in the respective county. Hiiu county being smaller than other counties comprised a separate stratum, the remaining counties were distributed into two strata - the larger and smaller ones. Counties with the population less than 60,000 belonged to the stratum of smaller counties (as at 1 January of the survey year).
To ensure an even distribution of the sample and preclude several address persons living at the same address from falling into the sample, records in the strata were sorted by address: first by the county code; within the county, by the rural municipality code; within the rural municipality, by the name of village; next, by the street name; and finally, by the house number.
The original sample included 8,100 persons. In order not to put an excessive burden on the respondents, those who had participated in Statistics Esonia's surveys before were excluded. The final size of the sample was 7,803 persons.
Although the inclusion probability is smaller in the stratum of larger counties than in other strata, the result gives a relatively large sample for Tallinn. This is necessary for the purpose of analysis, because in Tallinn the response probability is the lowest, but the diversity of households is the largest. Thus, a larger sample size from other (more homogenous) regions guarantees a required accuracy of estimates.
In order to expand the survey results to the whole population, a weight was assigned to every household. The calculation of weights consisted of the following stages:
- calculation of design weights;
- compensation for non-response;
Household weights were calculated regarding how many ways there were to reach that household. Thereby, two factors had to be taken into account. First, the stratified unproportional systematic sampling of address persons was used for drawing samples. Second, sampling based on address persons granted a bigger household a bigger probability to be included in the sample – a household could be included in the sample through any of its member who was at least 15 years of age. Thus, the design weight of a household and its members is proportional to the inclusion probability of address person in his/her stratum and to the size of household.
The design weight is in inverse proportion to the inclusion probability. The design weight was calculated for all households: the responded and non-responded households. In case of non-responded households, the number of persons registered at the sample person’s residence address was considered to be the number of the at least 15-year-old persons in the respective household. If the number of residents registered at the address of a non-responded household exceeded 9 persons, the number of the at least 15-year-old persons in the household was imputed.
Dates of Data Collection
Data Collection Mode
Estimates of Sampling Error
Only a part of the population can be surveyed by sample survey. Because of that, the indicators calculated on the basis of sample data are always somewhat different from the actual value of the estimated population parameter. Such a difference is called the random error or sampling error of estimation. It is not possible to specify the sampling error exactly, but it can be estimated statistically by taking the variability or dispersion of the statistic that is used for parameter estimation as the basis for the sample design used in the survey. In addition to the sample design, the sampling error depends on the sample size. A smaller sampling error can be expected in case of larger sample sizes.
An important group of quality indicators consists of the accuracy estimations of parameters calculated on the basis of the survey. The accuracy estimations provided by Statistics Estonia are estimates of the sampling error i.e. these estimations do not reflect other
possible error sources. Estimates of sampling errors are calculated for more important indicators.
Standard error is the main sampling error estimate. Standard error is a mathematical value that describes the variance of parameter estimates given on the basis of the sample. As the sample is selected randomly, the parameter estimate is also a random variable and variance can be calculated for it. The smaller the variance, the more exact is the parameter estimate. The variance of estimate depends on the sample size and sample design.
Relative standard error shows the proportion that the estimate’s standard error forms of the estimated value. As a rule, it is presented as a percentage. Relative standard error is independent from measuring units, due to that it allows for comparing of different parameter estimations with each other irrespective of measurement units. Relative standard error is an operative tool in order to receive a quick overview of the accuracy of estimates.
Use of the dataset must be acknowledged using a citation which would include:
- the Identification of the Primary Investigator
- the title of the survey (including country, acronym and year of implementation)
- the survey reference number
- the source and date of download
Statistics Estonia. Household Budget Survey 2010, Ref. EST_2010_HBS_v01_M. Dataset downloaded from [URL] on [date].
Disclaimer and copyrights
The user of the data acknowledges that the original collector of the data, the authorized distributor of the data, and the relevant funding agency bear no responsibility for use of the data or for interpretations or inferences based upon such uses.