The Tanzania Demographic and Health Survey (TDHS) is a national sample survey of women of reproductive ages (15-49) and men aged 15 to 60. The survey was designed to collect data on socioeconomic characteristics, marriage patterns, birth history, breastfeeding, use of contraception, immunisation of children, accessibility to health and family planning services, treatment of children during times of illness, and the nutritional status of children and their mothers.
The primary objectives of the TDHS were to:
- Collect data for the evaluation of family planning and health programmes,
- Determine the contraceptive prevalence rate, which will help in the design of future national family planning programmes, and
- Assess the demographic situation of the country.
Kind of Data
Sample survey data
Unit of Analysis
- Children under five years
- Women age 15-49
- Men age 15-60
The Tanzania Demographic and Health Survey (TDHS) is a national sample survey. This sample should allow for separate analyses in urban and rural areas, and for estimation of contraceptive use in each of the 20 regions located on the mainland and in Zanzibar.
Producers and sponsors
Tanzania. Bureau of Statistics
Macro International Inc.
Ministry of Health
Technical advice and logistical support
United Nations Population Fund -Tanzania
Advisory and logistic support
United Nations Children's Fund - Tanzania
Advisory and logistic support
United States Agency for Intemational Development
The principal objective of the Tanzania Demographic and Health Survey (TDHS) was to collect data on fertility, family planning, and health of the people. This survey involved randomly selected women aged 15-49 and men aged 15-60 in selected households.
Before the sampling frame was developed, two possibilities for the TDHS sample design were considered:
- The 1988 Population census list of Enumeration Areas (EAs)
- The National Master Sample for Tanzania created in 1986 (NMS).
The NMS was intended mainly for agricultural purposes and, at that time, only for rural areas. The NMS was based on the 1978 Census information while the urban frame was still being worked upon. Therefore, it was decided that the TDHS sample design would use the 1988 Census information as the basic sampling frame. Since the TDHS sample was to be clustered, it was necessary to have sampling units of manageable and fairly uniform size and with very well defined boundaries. The 1988 Census frame provided the list of enumeration area units (EAs) that had well defined boundaries and manageable uniform size. Therefore, EAs were used as primary sampling units (PSUs).
The target of the TDHS sample was about 7850 women age 15-49 with completed interviews. This sample should allow for separate analyses in urban and rural areas, and for estimation of contraceptive use in each of the 20 regions located on the mainland and in Zanzibar. Estimates for large domains (by combination of a group of regions) were also taken into consideration.
The TDHS used a three-stage sample. The frame was stratified by urban and rural areas. The primary sampling units in the TDHS survey were the wards/branches. The design involved the target of 350 completed interviews for each of 19 regions on the mainland and 500 in each of Dar es Salaam and Zanzibar.
In the first stage, the wards/branches were systematically selected with probability proportional to size (according to 1988 census information). In a second sampling stage, two EAs per selected rural ward/branch and one EA per selected urban ward/branch were chosen with probability proportional to size (also according to 1988 census information). In total, 357 EAs were selected for the TDHS, 95 in the urban area and 262 in the rural. A new listing of households was made shortly before the TDHS fieldwork by special teams including a total of 14 field workers. These teams visited the selected EAs all over the country to list the names of the heads of the households and obtain the population composition of each household (total number of persons in the household). In urban areas, the address of the dwelling was also recorded in order to make it easy to identify the household during the main survey. A fixed number of 30 households in each rural EA and 20 in each urban EA were selected.
About 9560 households were needed to achieve the required sample size, assuming 80 percent overall household completion rate.
See detailed sampling information in the APPENDIX B of the final 1991-1992 Tanzania Demographic and Health Survey report.
Out of the 9282 households selected for interview, 8561 households could be located and 8327 were actually interviewed. The shortfall between selected and interviewed households was largely due to the fact that many dwellings were either vacant or destroyed or no competent respondents were present at the time of the interview. A total of 9647 eligible women (i.e., women age 15-49 who spent the night before the interview in a sampled household) were identified for interview, and 9238 women were actually interviewed (96 percent response rate). The main reason for non-interview was absence from the home or incapacitation.
The Tanzania DHS male survey covered men aged between 15 and 60 years who were living in selected households (every fourth household of the female survey). The results of the survey show that 2392 eligible men were identified and 2114 men were interviewed (88 percent response rate). Men were generally not interviewed because they were either incapacitated or not at home during the time of the survey.
Dates of Data Collection
Data Collection Mode
Data Collection Notes
The training of the field staff was carried out in two classrooms at one location. The training took place at the National Bank of Commerce, Bankers' Training Institute, Iringa Municipality, Iringa Region. Two teams of two persons from the Census Office, with support of staff from DHS/Macro International and the Tanzania Food and Nutrition Centre, conducted the training. Trial interviews were done in the neighborhood of the municipality.
The four weeks of training were devoted to classroom lectures, classroom mock interviews among the trainees, and interviews with invited volunteers from outside the classrooms with everyone listening. At the end of such interviews, trainees were invited to relate their observations as to how the entire interview was conducted and describe any mistakes that were made by the interviewer under the prevailing conditions. Later, the interviewers were taken out of the classroom to the field where each was assigned a household to interview. When all the interviewers were back in the classroom, each was asked to recount some of the experiences he encountered in the administration of the questionnaire, the behavior of the respondent, and how he handled the situation.
During training, a series of assessment tests were given to the interviewers and supervisors. These tests were graded and the results were used in selecting interviewers. Those candidates who bad a better grasp of the questionnaire and were able to detect errors in completed questionnaires were chosen to be field editors.
Two guests were invited to talk to the trainees on family planning and maternal and child health. A staff person from the Tanzania Family Planning Association (UMATI) talked about family planning methods and sources where they can be obtained. Another person who came from the Maternal and Child Health Department in the Regional Hospital of Iringa discussed maternal and child health care and immunization.
Anthropometry was taught for a week and was conducted by a DHS consultant on anthropometry who was assisted by a specialist from the Tanzania Food and Nutrition Centre. Arrangements were made with the neighboring villages for practice measuring of children below age 5 years and their mothers. All trainees received the anthropometric training.
Supervisors and editors received additional training on their specific duties in the field. The main purpose was to ensure that all the teams would be following a uniform set of procedures.
The forty-eight female interviewers, 8 male interviewers, and 8 editors were selected from the larger pool of trainees at the end of the training course. Eight teams were made up and each team was allocated between two and three regions. Each field team was composed of six female interviewers, one male interviewer, one editor, a supervisor, and a driver (each team had one vehicle).
The main fieldwork began immediately after the training. Teams were dispatched to their respective regions with instructions to introduce themselves to the regional, district, ward, and village leaders before beginning the interviews. When a team got into the cluster (enumeration area), it had to identify itself to the local leaders before interviews would begin. This was necessitated by the fact that, in order to get the required maximum cooperation, the leaders bad to be aware of the project and so inform the people about it. This is the usual practice in Tanzania whenever there is anything that involves the participation of the people, be it in urban or rural areas.
The fieldwork for the main survey was conducted in the period between October 1991 and March 1992. Thirty households were selected for interview in each cluster (EA). Women and men for the individual interview were identified during the household interview. Team supervisors located the housing units and assigned the selected households to the interviewers. Completed household and individual questionnaires were handed over to the field editors who checked to ensure that all relevant questions were recorded correctly, that the skip pattern instructions were followed properly, and that responses were internally consistent. The editing work was done before the team left the EA so that the interviewer could go back to the respondent to resolve any errors.
The supervisors were required to ensure that all the selected households and eligible women and men in an EA were interviewed, and that assignment sheets for the interviewers and supervisors were duly filled. The questionnaires and other control forms were submitted to the Head Office in Dares Salaam for data entry.
The questionnaire for each DHS can be found as an appendix in the final report for each study.
The household, female, and male questionnaires were designed by following the Model Questionnaire "B" which is for low contraceptive prevalence countries. Some adaptations were made to suit the Tanzania situation, but the core questions were not changed. The original questionnaire was prepared in English and later translated into Kiswahili, the language that is widely spoken in the country. There are parts in the country where people are not very conversant with Kiswahili and would find it difficult to respond in Kiswahili but would understand when they are asked anything. The translated document was given to another translator to translate it back into English and comparisons were made to determine the differences.
A pretest to assess the viability of the survey instruments, particularly the questionnaires and the field organization, was carried out in Iringa Rural District, Iringa Region. It covered 16 enumeration areas with a total of 320 households. The pretest, which took a month to complete, was carded out in November/December, 1990, and covered both rural and urban EAs.
The pretest training took two weeks and consisted of classroom training and field practice in neighborhood areas. In all, 14 newly recruited interviewers and the Census staff were involved. The Census staffs who were to be transformed into the TDHS team handled the training for both the fieldwork management and the questionnaire. During the later fieldwork, they supervised the field exercise.
During the fieldwork, the administrative structure of the CCM Party, which involved the Party Branch Offices and the ten-cell leadership, were utilized in an effort to secure the maximum confidence and cooperation of the people in the areas where the team was working. At the end of the fieldwork, the interviewers and the supervisory team returned to the head office in Dares Salaam for debriefing and discussion of their field experiences, particularly those related to the questionnaires and the logistic problems that were encountered. All these experiences were used to improve upon the final version of the questionnaires and the overall logistic arrangements.
Estimates of Sampling Error
The results from sample surveys are affected by two types of errors, non-sampling error and sampling error. Non-sampling error is due to mistakes made in carrying out field activities, such as failure to locate and interview the correct household, errors in the way the questions are asked, misunderstanding on the part of either the interviewer or the respondent, and data entry errors. Although efforts were made to minimize this type of error during the design and implementation of the TDHS, non-sampling errors are impossible to avoid and difficult to evaluate statistically.
Sampling errors, on the other hand, can be measured statistically. The sample of women selected in the TDHS is only one of many samples that could have been selected from the same population, using the same design and expected size. Each one would have yielded results that differed somewhat from the actual sample selected. The sampling error is a measure of the variability between all possible samples; although it is not known exactly, it can be estimated from the survey results.
Sampling error is usually measured in terms of standard error of a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to calculate confidence intervals within which one can be reasonably assured that, apart from non-sampling errors, the true value of the variable for the whole population falls. For example, for any given statistic calculated from a sample survey, the value of that same statistic as measured in 95 percent of all possible samples with the same design (and expected size) will fall within a range of plus or minus two times the standard error of that statistic.
If the sample of women had been selected as a simple random sample, it would have been possible to use straightforward formulas for calculating sampling errors. However, the Tanzania DHS sample designs depended on stratification, stages, and clusters. Consequently, it was necessary to utilize more complex formulas. The computer package CLUSTERS, developed by the International Statistical Institute for the World Fertility Survey, was used to assist in computing the sampling errors with the proper statistical methodology.
Note: See detailed sampling error calculation in the APPENDIX C of the final 1991-1992 Tanzania Demographic and Health Survey report.
The following Data Quality Tables are provided in the Final Report:
- Household age distribution
- Age distribution of eligible and interviewed women
- Completeness of reporting
- Birth by calendar year since birth
- Reporting of age at death in days
- Reporting of age at death in months
Data and Data Related Resources
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
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.
DDI Document ID
University of Cape Town
DDI Document version
Version 1.0: (June 2011) Variable level metadata is not provided.