South African Census 2022, 10% Sample - South Africa

Abstract
---------------------------

Censuses are principal means of collecting basic population and housing statistics required for social and economic development, policy interventions, their implementation and evaluation. The Post-Apartheid South African government has conducted four Censuses, in 1996, 2001, 2011 and 2022.

Geographic coverage
---------------------------

The South African Census 2022 has national coverage.

Analysis unit
---------------------------

Households and individuals

Universe
---------------------------

The South African Census 2022 covered every person present in South Africa on the Census reference night, midnight of 2-3 February 2022 including all de jure household members and residents of institutions.

Kind of data
---------------------------

Census/enumeration data

Sampling procedure
---------------------------

Census 2022 micro-data were sampled from the latest census data based on prescribed business requirements to ensure the generated estimates match the census population counts at local municipal level by age, sex, and population group. The Census 2022 Household record file and Person record file were used as a basis for the creation of the household and person sampling frame respectively. Households and household members in the in-scope households (records of the household questionnaire) formed part of the 10% sample. The municipal sample sizes for households were determined by taking 10% of the respective municipal measure of sizes. The household frame was implicitly stratified within each local municipality using household characteristics. Systematic samples of households were selected in each local municipality from the implicitly sorted household frame. The procedure used the allocated sample within each local municipality to produce the sample of households as well as the sampling weights, which are the inverse of the inclusion probabilities.

A Post-enumeration Survey (PES) is an independent sample survey that is conducted immediately after the completion of census enumeration to evaluate the coverage and content errors of the census. A sample of 840 sub-enumeration areas was selected across South Africa's nine provinces for the PES sampling frame. A mixed-mode data collection methodology was implemented to counteract the effects of the COVID19 pandemic. This was made possible by having integrated, digitally enabled survey processes with a geo-spatial information frame as a base.

Mode of data collection
---------------------------

Face-to-Face and Computer Assisted Personal and Telephone Interview

Research instrument
---------------------------

One questionnaire was used to capture the Census 2022 which included information on:
1. Households - particulars of the household (cover page of the household questionnaire) and household information sections of the household questionnaire.
2. Persons - particulars of the household member and person information sections of the household questionnaire.
3. Geography - geographic information associated with households and persons

A Post-Enumeration Survey was carried out after the census, which used a PES questionnaire.

Sampling error estimates
---------------------------

Coverage errors are a measure of how many persons or households were missed or counted more than once in the census.
The final net coverage error rate relative to the final true population of 61,4 million persons is thus 31,1%. The final net coverage error rate relative to the final true population of 19,3 million households is 30,5%.

Content errors indicate the quality of key characteristics in the census.
With respect to content errors, six variables were tested for consistency in terms of the responses that were recorded in the Census and the PES. The aggregated index of inconsistency was 7,5% for population group, 8,2% for sex, and 13,6% for age group, indicating a high level of agreement. The aggregated index of inconsistency for marital status was 23,0%, relationship to head of household was 34,8%, and country of birth was 42,3%, indicating moderate rates of agreement.

摘要
---------------------------
人口普查是收集基本人口和住房统计数据的根本手段，这些数据对于社会经济发展、政策干预及其实施与评估至关重要。南非后种族隔离时期的政府已进行了四次人口普查，分别于1996年、2001年、2011年和2022年进行。
地理覆盖范围
---------------------------
2022年南非人口普查具有全国性覆盖。
分析单元
---------------------------
家庭与个人
总体
---------------------------
2022年南非人口普查涵盖了2022年2月2日至3日午夜时在南非境内的每一位人员，包括所有法定家庭成员和机构居民。
数据类型
---------------------------
人口普查/普查数据
抽样程序
---------------------------
2022年人口普查的微观数据是基于规定的业务需求，从最新的人口普查数据中抽取的，以确保生成的估计值与人口普查的按年龄、性别和人口群体划分的当地市政层面的总人口计数相匹配。2022年人口普查家庭记录文件和个人记录文件分别作为家庭和人员抽样框架的基础。在范围内的家庭及其家庭成员（家庭问卷记录）构成了10%的样本。家庭样本量通过分别取相应市政规模的10%来确定。在每一个地方市政中，家庭框架根据家庭特征进行隐式分层。在每个地方市政中，从隐式排序的家庭框架中选取系统的家庭样本。该程序使用了分配给每个地方市政的样本，以产生家庭样本以及抽样权重，抽样权重是包含概率的倒数。
人口普查后的调查（PES）是一项独立的样本调查，在人口普查登记完成后立即进行，以评估人口普查的覆盖范围和内容误差。在南非的九个省份中，选出了840个次级普查区域作为PES抽样框架。为了应对COVID-19大流行的影响，实施了混合模式的数据收集方法。这是通过整合数字化调查流程，以地理空间信息框架为基础来实现的。
数据收集方式
---------------------------
面对面、计算机辅助个人和电话访谈
研究工具
---------------------------
使用一份问卷来收集2022年人口普查信息，其中包括以下内容：
1. 家庭 - 家庭细节（家庭问卷封面）和家庭问卷信息部分。
2. 个人 - 家庭成员细节和个人信息部分。
3. 地理 - 与家庭和个人相关的地理信息。
在人口普查之后进行了一次人口普查后的调查（PES），使用的是PES问卷。
抽样误差估计
---------------------------
覆盖误差是衡量在人口普查中遗漏或重复计算的人数或家庭数量的指标。
因此，相对于最终真实人口6140万人，最终净覆盖误差率为31.1%。相对于最终真实家庭人口1930万户，最终净覆盖误差率为30.5%。
内容误差表示了普查中关键特征的质量。
就内容误差而言，对六个变量进行了测试，以检查其在人口普查和PES中记录的响应的一致性。在人口群体方面，不一致的综合指数为7.5%，在性别方面为8.2%，在年龄组方面为13.6%，表明高度一致。婚姻状况的不一致综合指数为23.0%，与户主的关系为34.8%，出生国为42.3%，表明中等程度的一致性。