Uganda National Household Survey 2002/2003

Uganda, 2002 - 2003

Uganda Bureau of Statistics (UBOS)

Identification

IDNO

UGA-UBOS-UNHS-2002-v1.0

Title

Uganda National Household Survey 2002/2003

Country

Name	Country code
Uganda	UGA

Abstract

The Uganda National Household Survey 2002/03 was the eighth in a series of household surveys that started in 1988. The UNHS 2002/2003 collected information on the economic characteristics of the population and its activity status at the household level. The main objective of the survey was to collect high quality and timely data on population and socio-economic characteristics of households for monitoring development performance. The UNHS 2002/2003 focused on four modules namely the Socio-economic, Labour force, the Informal Sector, and Community modules. The survey covered 55 districts of Uganda, with some parts of Gulu and Kitgum districts not fully covered due to insecurity. Pader District was not covered at all. Indicators on population characteristics, labourforce participation rates, education, health, household expenditure and poverty among others have been presented at national, regional and rural-urban levels. The UNHS 2002/03 survey findings estimate the population of Uganda at around 25 million. The average household size is estimated at 5 persons per household. Like in the previous surveys, a large proportion of the population is below 15 years of age, with the majority of household members being children of the household head, which trend has been the same over years. The Poverty Monitoring and Evaluation Strategy targets 98 percent primary school enrollment by the year 2003. The results of the survey reveal that in spite of efforts made so far, Net Primary Enrollment for children aged 6-12 is below the target at 86 percent. This is partly caused by the fact that some children enroll late for primary school. The results also show that many children continue to attend primary school after the official age of 12. For example, more than half of all children aged 13-18 years attend primary school. In addition, households report that the monetary costs related to schooling deter participation to a certain extent. The results show that the percentage enrolled increases with increased household wealth. There are consistent differences in educational attainment and in literacy, and these differences are consistent across regions, both by sex and income bracket. The northern region consistently emerges worse-off in almost every education indicator. About twenty eight percent of the country’s population fell sick in the 30 days preceding the survey with malaria/fever reported as the major cause of ill health. Of those who fell sick, many practiced self-treatment while others preferred to go to private clinics. Usage of mosquito nets remains low with only 11 percent of the population using them. Awareness of HIV/AIDS is almost universal. However it is not matched by knowledge of specific ways to avoid HIV/AIDS. The condom however is most mentioned as the specific method one can use to avoid HIV/AIDS. The radio is reported to be the main medium through which people acquire information on HIV/AIDS. Most of the housing and household conditions have improved especially the housing structure i.e. wall, roof and floor. Households are still dependent on “tadoba” for lighting and worse still, the majority of the households depend on wood as fuel for cooking. The 2002/03 survey has shown an increase in Per-household and Per-capita expenditure. Foods, Beverages and Tobacco still dominate the household budget share, despite a drop of 8 percent observed over the same period. However, these changes have not been high enough to over turn the observed increases in poverty levels. The percentage of the population living below the poverty line rose from 34 percent to 38 percent. This rise is statistically significant. The main finding is that, despite some very modest economic growth, poverty increased. This is in contrast to trends in the 1990s, where growth was stronger and appeared to be broadly shared. There has been a general downward trend in the welfare indicators between 1999/00 and 2002/03 periods. Ownership of clothes declined between the 1999/00 and 2002/03 periods while ownership of bicycles and radios has improved over the same period. One in every 5 children aged 0 – 5 years, in the eastern and northern regions does without breakfast. About 36 percent of the households in Uganda own non-crop enterprises. The major enterprises being in the manufacturing and trade and services broad industries. These two categories employ 1.8 million persons while livestock, poultry, bee-keeping, and fishing industry employs another 0.5 million persons. Most household based enterprises are sole proprietorship, and similarly there are mainly started by owners. Nearly 90 percent of the persons aged 10 years and above were usually active during the 12 months prior to the survey. About 60 percent of these were own account workers followed by unpaid family workers (26 percent). The distribution of usually active persons by Industry show that the agricultural sector is still dominant accounting for 68 percent of the employed persons . Considering the last seven days, a higher proportion of persons aged 10 years and above were own account workers (54 percent). The Northern and Eastern Regions recorded the highest proportion of persons employed in agriculture. It is noted that most of the urban dwellers are employed in the sales and service sector. Occupational categories of household members show that 2 in every 3 persons were engaged in agriculture, only 4 percent were involved in Market Oriented Agriculture Production. A higher proportion of women than men was recorded for those who were engaged in domestic duties. Those who did not participate in economic activities during the last 7 days, stated being ill as the dominant reason. Among reasons for not being usually active during the last 12 months, attending school featured as the prominent reason followed by attending to domestic duties. Nearly 25 percent of the employed population were engaged in secondary activities and the service workers were more likely to engage in secondary activities than any other occupational category. The current labourforce participation rate is 67 percent. Participation levels by selected background characteristics show that persons without education had higher participation levels than those with primary education. The findings further show that rural women had higher participation rates than their urban counterparts. Twenty percent of the persons in paid employment earn at most shillings 20,000. Of those persons earning more than shillings 60,000, 63 percent were men while only 37 percent were women. Of the currently economically active persons, 3 percent are unemployed. Youth unemployment (5.3 percent) rate was higher than the national rate (3.2 percent). Central region had unemployment rate higher than other regions. Sixty five percent of the unemployed had attempted to look for work. This was mainly through friends and relatives. Most of the unemployed depended on relatives or spouses for survival with females depending more on spouses. The underemployment rate is highest among youth but steadily declines among those aged 50 years and over. The national underemployment rate stood at 15 percent. The survey results show that the underemployed are mainly in the agricultural sector. The underemployed were willing to do any job to earn more money. Findings show that although one in every five working children is an orphan, three out of every four children were non-orphans. Children whose parent survival status is not known are more likely to work. Children who were not attending school engaged in paid employment than those attending school. Furthermore, those who engaged in paid domestic services were more likely to work for more hours in a day than those engaged in other activities.

Kind of Data

Sample survey data [ssd]

Version

Version number

- v1.2: Edited data, second version, for internal use only.

Version Date

2003-10-31

Scope

Topic Classification

Topic	Vocabulary	URI
ECONOMICS [1]	CESSDA	http://www.nesstar.org/rdf/common
LABOUR AND EMPLOYMENT [3]	CESSDA	http://www.nesstar.org/rdf/common
TRADE, INDUSTRY AND MARKETS [2]	CESSDA	http://www.nesstar.org/rdf/common

Keywords

Keyword	Vocabulary	URI
ECONOMICS [1]	CESSDA	http://www.nesstar.org/rdf/common
LABOUR AND EMPLOYMENT [3]	CESSDA	http://www.nesstar.org/rdf/common
TRADE, INDUSTRY AND MARKETS [2]	CESSDA	http://www.nesstar.org/rdf/common

Coverage

Geographic Coverage

The Uganda National Household Survey 2002/03 was conducted in all districts except Pader. Some parts of Kitgum and Gulu districts were also not covered due to insecurity.

Unit of Analysis

The following are the units of analysis;
- Individual
- Household
- Community

Universe

The survey covered all resident population.

Producers and sponsors

Authoring entity/Primary investigators

Agency Name	Affiliation
Uganda Bureau of Statistics (UBOS)	Ministry of Finance, Planning and Economic Development

Funding Agency/Sponsor

Name	Abbreviation	Role
Government of Uganda	GoU	Funding of operational cost
World Bank	WB	Funding of operational cost

Other Identifications/Acknowledgments

Name	Affiliation	Role
Respondents		Provided the required data
Local Community leaders		Mobilising the communities
Economic Policy Research Centre	Makerere University	Research Support
Institute of Statistics and Applied Economics of Makerere University	Makerere University
Ministry of Gender, Labour and Social Development	Government of Uganda
Population Secretariat	Ministry of Finance, Planning and Economic Development
Uganda Manufacturer’s Association

Sampling

Sampling Procedure

The sampling design was chosen to fit the purpose of the survey. Stratified two stage sampling was adopted, but with a few refinements such as over-sampling of urban areas, and possibly of some rural areas with concentrated informal sector activity. The sampling frame for selection of first stage units (FSUs) was the list of EAs with the number of households based on cartographic work for the 2002 Population and Housing Census. For selection of the second stage units, which were the households, listing exercise through listing schedules was done in selected EAs. Each district was a stratum and was divided into rural and urban sub-strata. The urban area was further sub-divided into district town and other urban areas. This deep stratification enabled a better spread and representation of the sample, thereby increasing the efficiency of the estimates. Additionally, the continuity over rounds was maintained to enable pooling of results over rounds, if ever considered necessary. The total number of about 1,000 FSUs was firstly allocated between urban and rural in the proportion of 40:60. Thereafter, the urban and rural sample was generally allocated between the strata in proportion to the number of households with certain adjustments. The allocated sample was selected with probability proportional to number of households. A suitable plan for sub-stratification and selection of households at the listing stage, was introduced to ensure adequate representation of households with at least one unemployed person and an informal sector enterprise activity. The households were at first divided into 2 groups namely; households with at least one unemployed person and households with no unemployed person. The total 10 sample households in an EA were allocated between the unemployed and employed groups. Half of the sample from the unemployed was selected from households having one or more household enterprises, while the other was selected from the households having no enterprise activity. In case any one of the sub-groups did not exist, the total sample would be allocated to the existing sub-group. For odd sample sizes, the group with household enterprises got preference. The households to be selected from the group that did not have any unemployed person were sub-stratified by kind of informal sector activity. The allocation between the sub-groups was in proportion to the number of households, with a minimum of 1 from each group. As explained earlier, the allocation of the total sample between strata, was not strictly proportional to the number of households. Firstly, the urban areas were over-sampled and secondly some areas both in rural and urban were over or under sampled on the basis of degree of concentration of informal sector activity. Another refinement in the design made was to have a balanced independent inter-penetrating network of sub-samples (IPNS), on a quarterly basis to enable studying seasonality of some survey variables, to provide independent quarterly estimates and to eliminate seasonal effects while taking the average over four quarters. The UNHS sample was drawn through a stratified two-stage sampling design. The Enumeration Area (EA) was used as the first stage sampling unit and the household as the second stage-sampling unit. The sampling frame used for selection of first stage units (fsus) was the list of EAs with the number of households based on the cartographic work of the 2002 Population and Housing Census. A total of 972 EAs (565 in rural and 407 in urban areas) were covered. In order to select the second stage units, which are the households, a listing exercise using listing schedules was done in all selected EAs. Details of the sampling design are given in Appendix III. The sample size was determined by taking into consideration several factors, the three most important being: the degree of precision (reliability) desired for the survey estimates, the cost and operational limitations, and the efficiency of the design. UNHS 2002/03 covered a sample of 9,711 households.

Response Rate

The response rate for the Uganda National Household Survey 2002/2003 was approximately 97%. A total of 9711 households were interviewed out of the 10,000 households initially targeted. A total of 289 households could not be interviewed mainly due to insecurity.

Weighting

Details of the weighting are available in section 3.2 of the Uganda National Household Survey 2002/2003 - Data Documentation. This document is accessible from the Technical Documents link of this website.

Data Collection

Dates of Data Collection (YYYY/MM/DD)

Start date	End date
2002-05-03	2003-04-04

Mode of data collection

Face-to-face [f2f]

Supervision

The Survey staff comprised of a total of 15 field teams.Four teams were recruited for each region except for Northern region which had three. Each field team was composed of one supervisor, four enumerators, and one driver. There was no gender bias and these supervisors and interviewers could be either male or female. For instance, out of 75 enumerators and supervisors, 30% were female. Each field supervisor was responsible for one team of Interviewers. During each round of fieldwork, each team supervisor had to write a report detailing what transpired on the respective trip. It included technical issues concerning the questionnaires where he/she could not take outright decision, coverage in terms of the EAs allocated, and the administrative issues. During each round of fieldwork, each team supervisor had to write a report detailing what transpired on the respective trip. It included technical issues concerning the questionnaires where he/she could not take outright decision, coverage in terms of the EAs allocated, and the administrative issues. On the other hand, the team supervisor was tasked to make appointments with Local Council 1 (LC1) officials for the enumeration of their village, equitably distribute work amongst team members, check through the sampling process done by the enumerator, and check through questionnaires administered by the enumerators. He/she was also supposed to conduct an interview using the community questionnaire with LC1 officials and opinion leaders as respondents, interview the institution heads i.e. schools and health units most used by the community. Where applicable, he/she would also at times assist the enumerators in administering the household questionnaires. Repeated visits were made to those households where no 'well informed' member to be interviewed was found. In addition to the field teams, the survey management team as well as Bureau Staff at the level of officers and above undertook field supervision. On average, each team was assigned two days to complete an enumeration area. An allowance of one day was allocated to teams that had to travel very long distances from the UBoS head office. For details on the number of field teams dispatched and the number of field days allocated for each survey month refer to the UBoS, Administrative Report, 2003. In the central office at the UBoS headquarters there were editing clerks who performed the data scrutiny before the data were sent for data entry. The field staff who did not go for field work during the respective field trips acted as the edit clerks.

Type of Research Instrument

Nine types of questionnaires were used during the survey namely; Household Listing questionnaire, the Socio-Economic questionnaire, the Labourforce questionnaire, the Community questionnaire, Forestry Enterprise questionnaire, Trade and Services Enterprise questionnaire, Manufacturing, Mining and Quarrying Enterprise questionnaire, Livestock Enterprise questionnaire and Hotel Enterprise questionnaire. The last five questionnaires were administered to small-scale establishments and household enterprises. These were developed in consultation with various stakeholders. The household listing questionnaire was used to list all houses and households in the selected Enumeration Areas (EAs). Finally, the community questionnaire was administered at community level (Local Council level I).

Data Collectors

Name	Abbreviation	Affiliation
Uganda Bureau of Statistics	UBOS	Ministry of Finance, Planning and Economic Development

Data Processing

Cleaning Operations

A manual system of editing questionnaires was set up and a set of scrutiny notes to guide in manual checking was developed. In addition, range and consistency checks were included in the data-entry program. More intensive and thorough checks were carried out using MS-ACCESS by the processing team. Besides the editing done before data entry, the validation checks inbuilt in the program and double data entry, additional in-depth data cleaning on sections relevant for basic poverty analysis was done. For instance, individual level files were linked together to ensure that the same individual code reported in different sections of the questionnaire and in other modules corresponded to the same individual. Data cleaning on the other sections was also done . Any inconsistencies, data entry errors etc found were corrected after checking the original questionnaires.

Other Processing

CSPro was used to develop the data entry applications; Consistency and range checks were implemented in the CSPro data processing applications. Visual Basic was used to develop data Management Applications. These could automatically convert CSPro flat data files into Ms-Access, which was the final data storage tool. The system developed for data management could coordinate the sampled EAs, interviewed households, and entered households. It could also coordinate entered, compared, cleaned and converted data files. This left the data with very minimal completeness, consistence and data value errors. Other measures were put in place to minimize errors at the data entry stage. Data entry was done twice, that is, main entry and verification; there was 100 percent verification. A questionnaire could be considered clean after both datasets were perfectly comparable. When an entrant encountered a field problem, a field officer would be called upon to rectify the error.

Data Appraisal

Estimates of Sampling Error

There are two types of errors possible in any estimate based on a sample survey – sampling and non-sampling errors. Non-sampling errors can be attributed to many sources which include: definitional difficulties, differences in the interpretation of questions by the interviewers, inability or unwillingness to provide correct responses on part of the respondents, mistakes in coding or recording the data, et cetera. Nonsampling errors would also occur in a complete census. On the other hand, sampling errors occur because observations are made only on a sample, and not the entire population. Thus the accuracy of survey results is determined by the joint effects of the sampling and nonsampling errors. For a given indicator, the sampling error is usually measured by the standard error. The standard error of a survey estimate is a measure of the variation among the estimates from all possible samples, and is a measure of the precision with which an estimate from a particular sample approximates the results from all possible samples. The accuracy of a survey result de pends on both the sampling and nonsampling error measured by the standard error and the bias; and other types of nonsampling errors not measured by the standard error. The standard errors of the rates presented in this appendix were computed using the SASÓ PROC SURVEYMEANS procedure. This procedure does not assume that the data was taken from a simple random sample, but rather from a more complex design. The SurveyMeans Procedure takes into account the effect of clustering and stratifying in the calculation of the variances and standard errors, using the Taylor expansion method to estimate these sampling errors. The sampling errors are computed for selected variables considered to be of interest, but can be computed for all variables in the dataset. The sampling errors are presented for the country as a whole, for women and men where relevant, and for rural and urban areas and for each of the four regions: Central, East, West and North. For each variable the type of statistic (mean, sum, rate) are given as well as the standard error, the 95% confidence limits, and the coefficient of variation. Generally the standard errors of most national estimates are small and within acceptable limits, but there is wider variability for the estimates of the subpopulations. For example for the Net Attendance Ration (NER), the standard error for the whole country is 6.5 percent, while for urban and rural areas it is 7.6 and 7.3 percent respectively. For more details about the estimates of sampling error can be found in Appendix IV of "UNHS 2002/2003 Report on the Socio-Economic Survey"

Data access

Contact

Name	Affiliation	Email	URI
Uganda Bureau of Statistics (UBOS)	Ministry of Finance, Planning and Economic Development	ubos@ubos.org	www.ubos.org

Confidentiality Declaration

Confidentiality of respondents is guaranteed by Article 19 of The Uganda Bureau of Statistics Act, 1998. Before being granted access to the dataset, all users have to formally agree: 1. To make no copies of any files or portions of files to which s/he is granted access except those authorized by the data depositor. 2. Not to use any technique in an attempt to learn the identity of any person, establishment, or sampling unit not identified on public use data files. 3. To hold in strictest confidence the identification of any establishment or individual that may be inadvertently revealed in any documents or discussion, or analysis. Such inadvertent identification revealed in her/his analysis will be immediately brought to the attention of the data depositor.

Conditions

Metadata is provided to you for exclusive use. The metadata may not be transferred to any other user without prior written permission from UBOS.

Citation requirement

The following statement must be used as citation " Source of data: Uganda Bureau of Statistics, Uganda National Household Survey 2002/2003, Version 1.0 of the data set (November 2006), provided by UBOS."

Disclaimer and copyrights

Disclaimer

The Uganda Bureau of Statistics (UBOS) provides these metadata to external users without any warranty or responsiblity implied. UBOS accepts no responsibility for the results and/or implications of any actoins resulting from the use of these data.

Contacts

Contact

Name	Affiliation	Email	URI
Uganda Bureau of Statistics (UBOS)	Ministry of Finance, Planning and Economic Development	ubos@ubos.org	www.ubos.org

Metadata production

Document ID

DDI-UBOS-UNHS2002-2003/02 - v1.0

Producers

Name	Abbreviation	Affiliation	Role
Uganda Bureau of Statistics	UBOS	Ministry of Finance, Planning & Economic Development	Documentation of the survey

Date of Production

2009-11-30

Document version

Version 1.1 (November 2009). This is the second version of the survey documentation.