National Institute of Statistics of Rwanda

IPUMS-International Country Report: Rwanda

DRAFT VERSION

Prepared by Emmanuel GATERA

Management Information Systems Directorate

Building ex-Miniplan

Muhima, KIGALI

History of census taking

In Rwanda, the first census after the independence (1962) was held in 1978 with support from the UN Statistics Division. The enumerations were carried out by the authorities of National Office of Statistics (NOS) the provinces, the communes and the sectors officers for control purposes by the MNRD party (in power). A National Census Commission (NCC) and a Central Census Bureau (CCB) were created for management purposes.

The censusshould not only serve one purpose (until now mostly demographic) but should serve the needs ofthe administration of the state on ethnic basis. Three majors ethnic groups were identified since the first population census for a development planning based on ethnic quota.

The characteristics were covered for the total of the population (i.e. age, sex, nationality, ethnic group, religion, marital status, occupation by activity). For each administrative entity the present and the absent population was determined as well as the native population andforeigners.Dwelling and individuals have been enumerated and housing characteristics were also collected in 1978.

The next census was carried out in 1991. At the beginning only printouts of the requested tables were available.

Since 2002 (RGPH3) aggregatedfiles have been available as ASCII-delimited files on CD Rom. Theamount of tables from the censuses stored in the database has been steadily growing and inmost cases users find their needs satisfied. A website was designed for data dissemination thru internet (tables and thematic analysis mainly) but due to the low bandwidth connection, it became very soon unavailable.

2. Census documentation (table 1)

2.1 Enumeration forms

Since 1978 individual forms for each person in the household have been used. Persons living in the household have to be listed in the census list giving their residence status (present orabsent). The census list serves as an instrument for controlling completeness of enumeration.All questionnaires and explanation forms were part of the documentation translated in Kinywarwanda (the sole national tongue) since the first national census in 1978.

2.2 Instructions to enumerators

Instructions to enumerators are part of the documentation. The size of the booklets wasgrowing from census to census. In 2002, also a handbook for the local census manager was

Issued (controller and supervisors manuals).

2.3 Technical reports, data processing documentation

Reports were issued before, during and after the 2002 census mostly elaborated by UNFPA consultants in demography and data processing containing instructions for thedata processing, data editing and imputations. Suchdocumentation is not available for the 1978 and 1991 census. There are only single reports for the RGPH2 (1991) such as data editing programmes, a report on the final data checks, instructions for manualprocessing, for coding, but they have not been published.

2.4 Other documentation

1991: disparate documentation can be find out of the country as all the public archiving system was completely destroyed during the genocide (1994).

2002:For the last census (RGPH2002), a huge documentation is available at Rwanda National Institute of Statistics included printouts reports, thematic analysis reports, geographic atlas (shape files), etc.

RNIS will soon disseminate all the surveys publications since the ten past years via Internet using Microdata Management Toolkit (sponsorized by the World Bank).

3. History of census microdata (table 2)

3.1 Census characteristics

The 1978and1991 population and housing censuses were carried out by the Central Census Bureau (CCB) acting under National Census Commission (NCC). The 2002 Population and Housing Census was carried out under National Census Service (NSC) created for management purposes with financial support from European Union, UNFPA and Rwanda Gov. The scope of theenumeration was 100 per cent. People were counted in the territory where they have their usualresidence. Units of enumeration were compounds, dwellings andhouseholds (composed or individuals).

4. Contents of microdata files

4.1 Population census

Microdata files contain nearly all variables available from the population censuses. Variables

which are not mentioned in the following section were not part of the Rwanda censusprogramme.

Geography and internal migration: only place of usual residence provided.
Place ofprevious residence was asked in the 2002 census.
Household and family structure: Relationship to the head of household.
Demographic and social variables: Sex, age, marital status, de-facto marital status,
citizenship and religion provided.
Ethnic group for the 1978 and 1991 census asked : (forbidden in 2002).
Fertility and mortality: Number of children born alive asked in 1978, 1991 and 2002
Education: School attendance, instruction level, educational attainment and field of study provided.
Economic variables: Current activity status, status in employment,
Disability: asked in all census

Household variables: type of household, household status, size of household; ID-number ofhousehold
Other household variables: number of persons of a certain age in the household, number ofeconomically active persons in the household; characteristics of household reference person (head).
Other family variables: age of children; number of children in the family, characteristics of the wife/female partner (1978and1991);

4.2 Housing census

Dwelling variables, facilities: location of dwelling (ground floor, first floor etc.), number ofoccupants, number of rooms, useful floor space (1978to 2002), standard of equipment,water, toilet (only 2001), type of energy for lighting, type of energy for cooking,

5. Harmonization (1978 to 2002)

5.1 General remarks

International recommendations for Population and Housing Censuses play an important rolewhen developing questionnaires and determining variables and their categories, for examplethe UNSD and UNFPA Recommendations on serial 1970, 1980, 1990, 2000 and the 2010 rounds population censuses.

Rwanda administrative Map changed many times since the independence in 1962. The last one occurred in 2005 for the decentralization process. The country is now divided in five provinces(5), thirty districts (30), four hundred sixteen sectors (416) and three thousands cells (3000). NIS has updated the 2002 Population Census database to the new system.

Coding was done manually by codifiers directly on the forms in 1978, 1991 and 2002 mainly for geographical codes (provinces, districts, sectors, enumeration zones, cells, etc.).

NIS will work on how to make the 1991 population microdata comparable to 2002 classificationsand definitions. Up to now the “new” 1991 census data is available.

5.2 Variables

a)Social variables

Marital status: is defined as the legal (de jure) marital status (single [never married], married,divorced, widowed). The definition and the number of categories are the same for the threecensuses in discussion and meet international standards.

Relationship to head of household (HoH):

In all three censuses the household reference person (“head of household”) was determinedby the household. In 1991 and 2002, the question on status in household was placed on theindividual form with code to be ticked, for example “head of household”, in 1978to 2002 the relationshiphad to be stated in the household list and was coded.

b)Administrative geography and migration

Province comparability is influenced by the fact that the number of municipalities changesbetween the censuses or/and that boundaries between them included districts have beenredefined. NIS has established equivalence tables which allowed comparison between provinces before the last administrative reform. Many surveys like DHS 2005, QUIB, E-ICV 2001, E-ICV 2005 and the RGPH3 (2002) census are drawn in the previsious administrative framework.

c)Educational variables

“Type of school attended” has a lot more categories in 2002 than before. Full comparability

can be reached only if some 1991 categories are summarized according to the formerclassification.

d)Economic variables

Activity status

Between 1978 and 2002 a major change occurred in the field of current activity status. “Employed” was defined as having performedwork on regular or temporally basis.

“Unemployed”persons comprise all persons who during the reference period are without

Work. ” and “actively seeking work” are not considered.

Status in employment

It is not only possible to tabulate economically active personsby economic variables but also the whole population.Persons in the household who are not economically active (children, homemakers, elders) are giventhe characteristics of the active person in the household.

Occupation

Occupational classification was changed from census to census. In 1978, a complex classification was used regarding specialization by area. In 1991 and 2002, they used international standards promoted by ILO (international Labor Organization)

e)Demographic variables

Age is defined as completed years of age at census day and comparable over time.Comparability of definition also applies to sex and citizenship. 1978 to 2002 the variable“country of citizenship” only consisted of 12 categories. No citizenship was considered as “apatride”

Ethnic group has been asked in every census except in 2002 for Rwandan and this characteristic was a very sensible one; there was an obligation to answer the question. The ethnic group was referred to the father’s in case of mixed couple.

Religion has been asked in every census. From 1978 to 2002 there were only afew categories to be ticked on the questionnaire. The characteristic is one of very few with a category “unkown” which can beinterpreted as animist.

f)Dwelling variables

The 1991 and 2002 microdata on dwellings will be drawn from the harmonized files. So thevariables will be comparable over time for standard of equipment, variables on energy for cooking, for lighting, sanitation, origin of water, housing materials, etc…

6. Publications

The results of the population and housing censuses were published in a series of publications.Main results are at national level and provinces level, for example education, households and individuals characteristics. Unfortunately, it is hard to find printouts publications for the 1978 and 1991 Population and Housing Census as the national archive system was destroyed during the genocide in 1994.

2004: Many publications and reports were done on the 2002 Population and Housing Census. CD-ROM are available at National Institute of Statistics of Rwanda (NISR). The NISR portal is under construction and will provide very soon all the publication and microdata files.