Login
Login
    Home / Central Data Catalog / DDI-UGA-UBOS-UDHS-1995-V01
central

UGANDA DEMOGRAPHIC AND HEALTH SURVEY 1995

Uganda
Get Microdata
Reference ID
DDI-UGA-UBOS-UDHS-1995-v01
Producer(s)
UGANDA BUREAU OF STATISTICS
Metadata
DDI/XML JSON
Created on
Aug 17, 2026
Last modified
Aug 17, 2026
Page views
111511
Downloads
108
  • Study Description
  • Data Description
  • Downloads
  • Identification
  • Version
  • Scope
  • Coverage
  • Producers and sponsors
  • Sampling
  • Data Collection
  • Metadata production

Identification

Survey ID Number
DDI-UGA-UBOS-UDHS-1995-v01
Title
UGANDA DEMOGRAPHIC AND HEALTH SURVEY 1995
Country
Name Country code
UGANDA UGA
Abstract
The 1995 Uganda Demographic and Health Survey (UDHS) was implemented by the Statistics Department, Ministry of Finance and Economic Planning, with technical and material support from Macro International Inc. and funding from USAID; UNICEF contributed to questionnaire design. Fieldwork ran from late March to mid-August 1995 and, unlike the 1988/89 round, covered all of Uganda's districts except Kitgum — a major improvement in national coverage. The survey collected data from a nationally representative sample of 7,070 women aged 15–49 and, for the first time, 1,996 men aged 15–54, and substantially expanded the 1988/89 survey's scope by adding modules on AIDS/STD awareness and behaviour, maternal mortality, and health service availability.

The survey aimed to measure levels and trends in fertility, family planning knowledge and use, infant and child mortality, and maternal and child health, allowing direct comparison with the 1988/89 UDHS. Findings indicate that fertility had begun a modest decline (from 7.1 to 6.9 births per woman nationally, and from 7.3 to 6.8 births per woman in the area comparable to the 1988/89 survey), while contraceptive prevalence nearly tripled among currently married women (from about 5 percent to 15 percent) over the same period. Childhood mortality also declined, though it remained high (an infant mortality rate of 81 per 1,000 live births and under-five mortality of 147 per 1,000), and full childhood vaccination coverage rose from 31 percent to 47 percent. Substantial regional and urban-rural disparities persisted in contraceptive use, fertility, and access to health services. Awareness of AIDS was found to be almost universal, though many respondents underestimated their personal risk of infection. Roughly 38 percent of children under four were classified as stunted, indicating continued high levels of chronic child undernutrition.

As with the 1988/89 round, the 1995 UDHS was conducted as part of the international DHS Program, using standardized methodology to enable both national trend analysis and cross-country comparison.
Unit of Analysis
Unit of Analysis

Individuals — the primary unit for most substantive content. Two sub-populations:
Women aged 15–49 (all women in selected households), covering fertility, family planning, maternal and child health, AIDS knowledge, etc. — 7,070 women interviewed.
Men aged 15–54, interviewed in a subsample of one-third of selected households, using a shorter questionnaire covering much of the same content minus full reproductive history — 1,996 men interviewed. This is new relative to 1988/89, which had no men's questionnaire.

Households — the Household Questionnaire enumerated all usual members and visitors of each sampled household, collecting characteristics of housing (used to identify individuals eligible for interview and to describe household-level living conditions).

Communities/clusters (service delivery areas) — via the Service Availability (SA) Questionnaire, administered by a separate interviewer team to assess family planning and health service infrastructure at the community/cluster level. Unlike the 1988/89 round (where SA data were collected but not published), this round's SA results are reported, in Chapter 12 — 417 community questionnaires were completed across 296 clusters (since some clusters spanned multiple Local Council 1 (LC1) areas). This makes "community/cluster" a genuine, published unit of analysis for this round.

Version

Version Description
VERSION 1

Scope

Keywords
Keyword
Link

Coverage

Geographic Coverage
Geographic scope: National, with near-complete coverage — a substantial improvement over the 1988/89 round. The 1995 UDHS was conducted in all districts of Uganda except Kitgum District, which was excluded entirely due to insecurity. Additionally, isolated enumeration areas (EAs) in a few other districts could not be surveyed for the same reason — six EAs in Kitgum, one in Apac District, and one in Moyo District were dropped, while two further EAs (one in Arua, one in Moroto) were replaced with substitute EAs rather than dropped. Overall, only ~80% of the geographic area covered by this round overlaps directly with the 1988/89 round's coverage for trend-comparison purposes, since 1988/89 excluded nine full districts while 1995 excludes essentially just Kitgum.

Administrative/geographic units referenced:

Country: Uganda
Sampling frame: 1991 Population and Housing Census
303 primary sampling units (enumeration areas, EAs)

Geographic coding levels available in the data/variables:

Region (4-category variable, a change from the six-region scheme used in 1988/89): Central, Eastern, Northern, Western — the standard four administrative regions replace the six-region language-based grouping used in the prior round. Note this is an important discontinuity for cross-round comparison — 1988/89 used West Nile, East, Central, West, South West, Kampala; 1995 collapses to the four-region national standard, meaning region codes are not directly comparable across the two rounds without a mapping/crosswalk.
Residence: Urban / Rural
DISH project domains (5 reporting domains, grouped by district-pairs, used for oversampled contraceptive prevalence estimates under the USAID-funded Delivery of Improved Services for Health (DISH) project):
Domain I: Kasese and Mbarara Districts
Domain II: Masaka and Rakai Districts
Domain III: Luwero and Masindi Districts
Domain IV: Jinja and Kamuli Districts
Domain V: Kampala District

Sampling frame notes relevant to geographic coding:

Institutional populations (army barracks, hospitals, police camps, etc.) were excluded from the household listing and thus from the geographic universe entirely.
303 EAs selected with probability proportional to size, listed/mapped in December 1994.
Universe
Individual-level universe (Women's Questionnaire):
All women aged 15–49 who were usual residents or visitors present in a sampled household, across all districts of Uganda except Kitgum District (and a handful of individually-excluded EAs in Apac and Moyo Districts due to insecurity). No restriction by marital status — both ever-married and never-married women were eligible.

Individual-level universe (Men's Questionnaire) — new in this round:
All men aged 15–54 who were usual residents or visitors in the subset of households selected for the male subsample (every third household in the overall sample). Note the differing upper age bound relative to women (54 vs. 49) and the narrower sampling fraction (1-in-3 households) — both important distinctions from the women's universe when defining eligibility rules in the DDI record.

Geographic restriction on the universe: The surveyed population excludes residents of Kitgum District entirely, plus residents of the specific excluded EAs in Apac and Moyo Districts. The universe is therefore "women aged 15–49 and men aged 15–54 residing in Uganda, excluding Kitgum District (1995)" — a much closer approximation to the full national population than the 1988/89 round, but still not 100% nationally exhaustive.

Household-level universe (Household Questionnaire):
All households within the 303 sampled enumeration areas (EAs), including their usual members and visitors of any age — used to enumerate household composition/housing characteristics and to identify individuals eligible for individual interview.

Community-level universe (Service Availability Questionnaire) — new/published in this round:
Health and family planning service infrastructure within the community/cluster areas corresponding to the sampled EAs — 296 clusters, yielding 417 completed community questionnaires (some clusters contained multiple Local Council 1 areas, each separately assessed).

Explicitly excluded from the universe:

Residents of Kitgum District and the specific excluded EAs (Apac, Moyo)
Institutional populations — army barracks, hospitals, police camps, etc. — explicitly excluded from the household listing exercise, and therefore from the sampling universe entirely (this is stated explicitly in this round's methodology, unlike 1988/89 where it was only implicit)
Men outside the 1-in-3 household subsample (for the men's survey specifically)

Producers and sponsors

Primary investigators
Name Affiliation
UGANDA BUREAU OF STATISTICS MOFPED
Producers
Name Affiliation Role
MINISTRY OF HEALTH MOFPED FUNDING
UGANDA BUREAU OF STATISTICS MOFPED DATA COLLECTION AND ANALYSIS

Sampling

Sampling Procedure
Sampling Procedure

Sample type: Stratified, two-stage probability sample, not self-weighting at the national level — weights are explicitly required for national-level estimates, due to deliberate oversampling of specific districts.

Sampling frame: The 1991 Population and Housing Census, from which 303 primary sampling units (PSUs), i.e., enumeration areas (EAs), were drawn.

Domains (strata) used:

Uganda as a whole
Urban and rural areas, separately
Four regions: Central, Eastern, Northern, Western
DISH project areas (USAID-funded Delivery of Improved Services for Health project districts), for calculating contraceptive prevalence rates at that level — grouped into 5 reporting domains:
I. Kasese and Mbarara Districts
II. Masaka and Rakai Districts
III. Luwero and Masindi Districts
IV. Jinja and Kamuli Districts
V. Kampala District

Two-stage design:

Stage 1: 303 EAs selected with probability proportional to size (PPS) from the census frame.
Between stages: A complete household listing and mapping exercise was conducted in each selected EA in December 1994, carried out by 11 trained listers from the Statistics Department. Institutional populations (army barracks, hospitals, police camps, etc.) were excluded from listing.
Stage 2: Households selected from the listing with probability inversely proportional to household size.

Eligibility within selected households:

All women aged 15–49 in selected households were eligible for individual interview.
In a 1-in-3 subsample of selected households, all men aged 15–54 were additionally eligible for individual interview.

Target sample size: 6,000 women and 2,000 men (achieved: 7,070 women and 1,996 men interviewed — women exceeded target, likely due to household eligibility yielding more than anticipated; men came in essentially at target).

Fieldwork losses: 8 EAs could not be surveyed due to insecurity (6 in Kitgum District, 1 in Apac District, 1 in Moyo District) and were dropped entirely; 2 additional EAs (1 in Arua, 1 in Moroto) could not be surveyed but were replaced with substitute EAs.

Oversampling / disproportionate allocation: Households in the DISH project districts were oversampled relative to their actual population share, to allow independent domain-level contraceptive-prevalence estimates for those areas. As a direct consequence, the report explicitly states: "the 1995 UDHS sample is not self-weighting at the national level; weights are required to estimate national-level indicators."

Note on rounding: Due to the weighting factor and rounding of estimates, published percentages/totals in tables may not sum exactly to 100.0 or to stated totals.

Sampling error estimation: Standard errors are presented in Appendix B, computed separately for women's and men's samples, and broken down by urban/rural sample and by each of the four regions (Central, Eastern, Northern, Western) — a considerably more granular sampling-error presentation than the 1988/89 round.
Weighting
Weighting

Why weighting is required: The report is explicit and direct about this — a two-stage stratified sample was used with deliberate oversampling of households in the DISH (Delivery of Improved Services for Health) project districts relative to their actual population share, in order to produce independent estimates for those areas. As a result, the report states outright: "the 1995 UDHS sample is not self-weighting at the national level; weights are required to estimate national-level indicators." Sample weights must therefore be applied to any national-level (or cross-domain) analysis to correct for the unequal probabilities of selection introduced by the DISH oversampling and by the household-selection procedure (which selected households with probability inversely proportional to size, from the listing).

Data Collection

Data Collection Mode
Data Collection Mode / Instrumentation

Primary mode: Face-to-face structured interviews using paper questionnaires, supplemented by direct anthropometric measurement of children and mothers — same core mode as 1988/89, but with an expanded instrument set (four questionnaires instead of three) and, notably, mixed-sex interviewer teams.

Instruments used (four questionnaires — one more than 1988/89):

Household Questionnaire (Household Schedule) — listed names and basic characteristics (age, sex, education, relationship to household head) of all usual members and overnight visitors. Also collected dwelling characteristics: water source, toilet facilities, floor materials, ownership of consumer/durable goods. Main purpose: identify individuals eligible for interview.
Women's Questionnaire — administered to all women aged 15–49, covering: background characteristics, reproductive history, family planning knowledge/use, fertility preferences, antenatal and delivery care, breastfeeding/weaning, child vaccination and health status, marriage and sexual activity, husband's occupation/education, woman's employment, AIDS/STD awareness and behaviour, adult mortality (including maternal mortality), and height/weight of children under four and their mothers.
Men's Questionnaire (new in this round) — administered to men aged 15–54 in a 1-in-3 subsample of households, covering much of the same ground as the women's instrument but shorter, omitting reproductive history and maternal/child health questions.
Service Availability (SA) Questionnaire — administered by a separate team at community level, near each selected Local Council 1 (LC1) area, capturing health and family planning service infrastructure. Where an EA spanned multiple LC1s, one SA questionnaire was completed per LC1 (yielding 417 questionnaires across 296 clusters).

Instrument development: Questionnaires were developed in English by a Steering Committee chaired by the Population Secretariat.

Language/translation: All questionnaires except the Service Availability instrument were translated into and printed in six major languages — Ateso, Luganda, Lugbara, Luo, Runyankole/Rukiga, and Runyoro/Rutoro (two more languages than the four used in 1988/89: Ateso and Luo were added, reflecting the near-full national coverage achieved this round).

Anthropometric measurement: Height and weight measured for children under four and their mothers (mothers' anthropometry is new relative to 1988/89, enabling the maternal-nutrition indicators reported in the World Summit for Children indicator table, e.g., "percent of mothers with low BMI").

Pretesting: Conducted in November 1994; 14 interviewers (7 mixed-sex teams of one female and one male interviewer each) trained for two weeks to implement the pretest — a deliberate design choice to pretest with mixed-sex teams given the new Men's Questionnaire.

Quality control: As with 1988/89, field editing and supervisor oversight were built into the fieldwork structure; the Population Secretariat convened Steering and Technical Committees to oversee survey design and implementation.

Data processing/instrumentation for entry: (Let me pull the exact data-entry/processing details from the report if you'd like the same level of specificity as I gave for 1988/89 — e.g., software used, number of data entry staff, location of processing — just say so and I'll dig into that section.)

Metadata production

DDI Document ID
DDI-UGA-UBOS-UDHS-1995-v01
Back to Catalog
Uganda Bureau of Statistics (UBOS)

© Uganda Bureau of Statistics (UBOS), All Rights Reserved.