In Lithuania, the First Population Data Were Collected in the 13Th Century. After This

LITHUANIA

Introduction

In Lithuania, the first population data were collected in the 13th century. After this period historical sources mention counts based on "smokes", i.e. farmsteads with fireplaces. These were held from the beginning of the 16th to the middle of the 17th century. More precise data were attempted during the 1789 census, therefore only the 1790 census, which enumerated all existing social groups, should be considered the first overall population census in the Grand Duchy of Lithuania.

The first population census under the Russian Empire was organised in 1897. Following independence, proclaimed in 1918, a new census was conducted only in 1923 when the state borders were established. During the Soviet period four censuses took place, the last being organised in 1989, based on the programme and processing order prepared by the Central Committee of Statistics of the former Soviet Union. These censuses were held in January, when migration of the population was at its lowest, residents were interviewed where they were living.

After restoration of Lithuania’s independence, the first Population and Housing Census was carried out in 2001 having midnight of the 5 April as the enumeration reference moment. The Agricultural Census planned for 2003 will be the second in the recent history of the Country. The Project was organised and implemented by the Lithuanian Department of Statistics (Statistics Lithuania), the central statistical office of the Government of Lithuania.

The census methodology was prepared taking into account international recommendations, experience of neighbouring countries, advice from experts, and national peculiarities. Population enumeration was conducted by visiting the households for face-to-face interviews of family members and people residing in the dwellings. This traditional solution was selected owing to the limited possibility of using population registers (still under implementation and not containing information on nationality, languages spoken, religion, work place and occupation)([1]) as well as other register data for household composition and housing characteristics.

Every citizen in the Republic of Lithuania, usually resident in the territory or absent abroad for a period of a year, was enumerated as well as non-nationals and persons without citizenship with permission to reside in Lithuania temporarily or permanently, and who at the census moment had been living in the country for more than one year. Final results of the Census will facilitate revision of the structural data of the Lithuanian population during the inter-censual period.

Legislation and registers

The Census Law, passed as a Decree of the Parliament of the Government of the Republic of Lithuania on 10 July 1999, and the additional ‘Resolution on basic operations concerning the Republic of Lithuania Overall Population and Housing Census 2001’ are based on the recently developed broader legislation, and realised as the following:

-Law on Statistics (23 December 1999), provides for the definition of confidential statistical data and principles of data protection. According to this law all data collected for official statistics must be used only for statistical purposes and disseminated as aggregates;

-Law on Legal Protection of Personal Data (17 July 1999), regulates the protection of personal data in governmental authorities;

-Code of Administrative Offences (13 December 1984, 15 June 2000), is in addition to the article defining responsibility and fines for revealing confidential statistical data.

Statistics Lithuania has three levels of security and confidentiality – physical, legal and technological – each with the means and internal documents to ensure individual data protection. Regulations define principles of confidentiality, i.e. minimum number of respondents in the surveys, micro-aggregation, geographical thresholds, coding and re-coding, sampling, treatment of sensitive records. Data protection measures were largely applied to the Census. All census workers signed a deed of covenant stating that they would not reveal confidential data, and premises for census operations were prepared with a secure lock and seal mechanism. The IT equipment used for the Census was logically separated from other staff. Preparing databases for processing, name, surname and personal number were removed from micro data to avoid direct identification of persons.

Preparatory aspects and conduction of enumeration

Preparation for the Census foresaw the strategy and main objectives, as follows:

-correct census method and questionnaire contents;

-assurance of data confidentiality and security;

-timely production and release of reliable results;

-sufficient and timely financial support and economical approach taken.

The census questionnaire contained questions to which the population had to reply, it was prepared following UN and Eurostat recommendations for the 2000 census based on censuses in the ECE Region. These recommendations were analysed and adjusted to national needs in co-operation with many other Lithuanian institutions, demographers and other specialists.

In addition to this classical set of questions, useful in obtaining demographic and socio-economic characteristics of the population and households, respondents were asked specific questions on their command of foreign languages, religious beliefs and disability. Information was collected on buildings and housing and their quality for the housing component of the Project.

The 2001 Lithuanian Census did not aim to collect all information on agriculture, leaving it to a specific enumeration. However, Statistics Lithuania took the opportunity and collected data on land tenure, owned or rented, by households. It will facilitate grouping of households by land area owned, social and demographic characteristics of land users and their distribution by administrative territorial division.

To ensure the exhaustiveness of the Census urban and rural areas were divided, based on schematic plans and maps prepared by the State Land and Real Estate Cadastre and Register and other similar institutions. Lists of residential buildings and dwellings used for habitation were worked out and cartographic material repeatedly revised prior to the census day. Leaflets, newspapers and magazines as well as television and radio spots and programmes were used to raise the awareness of the public and inform respondents.

The work was organised by Statistics Lithuania, local statistical offices and census commissions established in each municipality. In rural areas much was contributed to by local administrative officers. Testing had been undertaken since the end of 1997. At the final stage, about 15 thousand people were selected and trained to carry out field operations as census managers (600 persons), supervisors (1900) and enumerators (12000).

Enumeration was completed between 8 and 10 days, depending on the census district. Persons who were unable to be enumerated were given the possibility to do so later. All possibilities for self-enumeration were offered to respondents, who in general welcomed enumerators and willingly replied to census questions. However, some problems were experienced related to the visiting of city dwellings. No sound signals had been installed in several individual private dwellings, and others had businesses set up on the premises. Overall, people who were not enumerated were able to be if they contacted the census divisions or regional statistical offices.

The census questionnaires were compared with the Population Register and lists of students. Distribution was made to conscripts and inmates in hospitals and prisons and other institutional homes based on their place of residence. The census data were received on Lithuanian citizens, as well as on family members working in Lithuania’s embassies, various institutions and international organisations abroad.

Data processing

Processing of census data began after all questionnaires were collected at the central office of Statistics Lithuania. Optical character recognition technology was used for data entry, i.e. the census questionnaires were scanned, later data recognition and correction of unrecognised symbols was conducted. Data correction and editing was performed by operators.

In order to transfer information precisely from the questionnaires into the database data correction covered entry of unrecognised characters and some logical control (full logical control was done in an Oracle environment). When an error was detected the operator updated the recognised information using data from the questionnaire. After loading data into the Oracle environment and setting the relationship between pages, logical control and adjustment was carried out. The main objective at this stage was to remove logical data errors. Special software was developed on the basis of logical control rules prepared in advance, which helped in the detection of logical data errors.

Data were coded automatically using prepared coding manuals and dictionaries (including ISCED, NACE and ISCO). Automatic procedures coded data that agreed with information available in manuals and dictionaries. If, during this process, information had no equivalent in the dictionary, it was saved into temporary storage where only unique non-coded information was stored. Operators then coded this information.

During the mixed coding stage, i.e. when the operators coded unique non-coded information, coded information was saved in the respective dictionaries, based on which all remaining information was later coded automatically. Coding software was developed for mixed data with a special user-friendly system of supply and search, i.e. the software proposed possible (similar) variants. If the proposal was correct, operators pressed a button to approve the selection. If the program had nothing to propose operators could create their requests and code information based on the results of requests.

Organised, i.e. coded and logically error-free data were transferred into the final database, eventually after comparison of various individual data, such as personal identification number, name, surname, date of birth and address, with data from the Population Register.

Post-Enumeration Survey

To measure the coverage and data quality of the census, a post-enumeration survey (PES) was carried out one week after enumeration of the population. The main objectives of the PES were to measure

-census coverage error, and

-content error in the main socio-demographic data

The cluster sample was used in the Lithuanian PES. The smallest areas for which verifiable boundaries existed on maps were the Enumeration Areas (EAs). Therefore a total of 11481 EAs were selected as a sampling frame. At this stage, EAs including institutions (hospitals, prisons, etc.) were excluded from the frame. Before selecting a sample of EAs the population was divided into three strata: large urban (4158 areas), urban (2596 areas) and rural (4727 areas). Available resources allowed questioning of approximately 1% of the entire population, that is 84 EAs were selected for the sample. The simple random sample of areas was carried out in each stratum. Allocation of the sample into strata was proportionate to the estimated population size. Thus, 24 large urban, 20 urban and 40 rural areas were selected into the sample. All people living in selected areas were asked to respond to some of the questions from the census questionnaire.

Data dissemination

Statistics Lithuania announced first provisional results on 30 October 2001, reporting a usually resident population of 3.5°million on 6 April 2001 – 2.3°million in urban areas and 1.1°million in rural areas – less than previously as a result of the negative natural population increase, emigration and temporary out migration of the permanent population.

Final, more detailed resultswill be released between 2002 and 2003. Publications will cover the following topics:

-population by sex and age

-ethno-cultural conditions (nationality, education attainment, command of foreign languages, religion, etc.)

-population by economic status, activity, occupation and source of livelihood

-households and families

-buildings, dwellings and their conditions

-disabled persons and their living conditions

-demographic behaviour (migration trends, family formation, fertility, etc.)

Data will be presented as a total for the country, by territorial administrative area (county, municipality), cities and other settlements. According to the user requests, data may also be released by certain districts, settlements, and small administrative areas to form multiple groups.

In addition to printed publications, Statistics Lithuania will release a full range of specific census products and services designed to satisfy the needs of the majority of users. Data will be provided via Internet and on CD-ROM in an appropriate format so that it may be presented in combination with other statistical data. To achieve this objective an applied system ‘Census data dissemination and analyses’ was developed. The main purpose of this system is to connect the operational data sources, staging area, data warehouse, GIS technology and OLAP Client programs. Diagram 7 shows the

Diagram 7 - Information systems architecture for data dissemination in Lithuania

Source: Statistics Lithuania

system’s organisational structure. The data warehouse is being created as a virtual union of several data marts with integrated information being shared between them. The basic structure is developed in the SQL Server environment and MS Analysis Services will be used to produce aggregated cubes for fast access via LAN. The cubes produced by the MS Analysis Services can be browsed and processed in third-party software packages.As a tool for data dissemination via Internet and on CDs, Statistics Lithuania will use PC-Axis – the family of software packages originally developed by Statistics Sweden for data dissemination.

Costs

The Lithuanian Census was based on national funding plus about 4.4% external contributions from the PHARE programme (for equipment) and the UNPFA project (support for data dissemination and analysis). In total, about €9.5°million will be used in the period 1997-2003, which is €2.7 per capita. This comes to just under two-thirds of the total. The major cost was for enumeration staff.

Conclusion

The material collected through the census provides source material for population structure and family status, structure of households, number of employed and unemployed, economic activity and occupation of population, and many other more specific topics. It will be used for many different purposes, including inter-censual revisions, annual population estimates and demographic projections. Future census plans could gain from the improvement of registers and be based on a mixed use of traditional enumeration and extraction from registers.

[1] The register system in Lithuania presents around 50 registers including the Population Register. Concerning individual data the personal identification number serves as identifier and link key between registers.