National Population and Housing Census 2009 - Vanuatu

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

The key objective of every census is to count every person (man, woman, child) resident in the country on census night, and also collect information on assorted demographic (sex, age, marital status, citizenship) and socio-economic (education/qualifications; labour force and economic activity) information, as well as data pertinent to household and housing characteristics. This count provides a complete picture of the population make-up in each village and town, of each island and region, thus allowing for an assessment of demographic change over time.

With Vanuatu, as many of her Pacific island neighbours increasingly embracing a culture of informed, or evidence-based policy development and decision-making, national census databases, and the possibility to extract complex cross-tabulations as well as a host of important sub-regional and small-area relevant information, are essential to feed a growing demand for data and information in both public and private sectors.

Educational, health and manpower planning, for example, including assessments of future demands for staffing, facilities, and programmed budgets, would not be possible without periodic censuses, and Government efforts to monitor development progress, such as in the context of its Millennium Development Goal (MDG) commitments, would also suffer greatly, if not be outright impossible, without reliable data provided by regular national population counts and updates.

While regular national-level surveys, such as Household Income and Expenditure Surveys, Labour force surveys, agriculture surveys and demographic and health surveys - to name but just a few - provide important data and information across specific sectors, these surveys could not be sustained or managed without a national sampling frame (which a census data provides). And the calculation and measurement of all population-based development indicators, such as most MDG indicators, would not be possible without up-to-date population statistics, which usually come from a census or from projections and estimates that are based on census data.

With most of this information now already 9 years old (and thus quite outdated), and in the absence of reliable population-register type databases, such as those provided from well-functional civil registration (births and deaths) and migration-recording systems, the 2009 Vanuatu census of population and housing, will provide much needed demographic, social and economic statistics that are essential for policy development, national development planning, and the regular monitoring of development progress.

Apart from achieving its general aims and objectives in delivering updated population, social and economic statistics, the 2009 census also represented a major national capacity building exercise, with most Vanuatu National Statistics Office (VNSO) staff who were involved with the census, having no prior census experience. Having been carefully planned and resourced, all 2009 census activities have potentially provided very useful (and desired) on-the-job-training for VNSO staff, right across the spectrum of professional rank and responsibilities. It also provided for short-term overseas training and professional attachments (at SPC or ABS, or elsewhere) for a limited number of professional staff, who subsequently mentored other staff in the Vanuatu National Statistics Office (VNSO).

With some key senior VNSO members involved with the 1999 census, they provided a wealth of experience that was available in-house and not to mention the ongoing surveys such HIES and Agriculture Census that the office has conducted before the census proper. The VNSO has also professional officers who have qualified in the fields of Population and Demography who had manned the project, and with this type of resources, we managed to conduct yet another successful project of the 2009 census.

While some short-term census advisory missions were fielded from SPC Demography/ Population programme staff, standard SPC technical assistance policy arrangements could not cater for long-term, or repeated in-country assignments. However, other relevant donors were invited for the longer-term attachments of TA expertise to the VNSO.

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

The 2009 Population and Housing Census Geographical Coverage included:

- National (Vanuatu)
- Provinces (Torba, Sanma, Penama, Malampa, Shefa, tafea)
- Inhabited Islands (From Hiu, Torres Islands to Aneityum, Southern Islands)
- Ennumeration Areas (EA assigned to each enumerator)
- Villages / Towns
- Household or Dwelling

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

The Unit Analysis of the 2009 Population and Housing Census included:
- Household
- Person (Population)

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

The census covered all households and individuals throguhout Vanuatu

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

Census/enumeration data [cen]

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

Face-to-face [f2f]

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

The questionnaire basically has 5 sections; the geographical identifiers, the general population questions and education, labour force questions, the women and fertility questions and the housing questions.The geographical identifiers include the Village name, GPS code, EA number, household number and the Enumerator ID.The Person questions contain the person demographics including the education level and labour force status. A section on fertility for women in the reproductive age is also included. All have been guided by 'skip patterns' to guide the flow of questions asked.Household questions contained the basic description of the house materials, tenure, access to water and sanitation, energy, durables, use of treated mosquito nest and internet access.

Cleaning operations
---------------------------

In the Census proper, the Optical Character Recognition (OCR) system (ReadSoft Application System) was used to capture information from the completed forms. The captured data were then exported to MS Access database system for further editing and cleaning before the final data is transferred to CSPro for more editing and quality checks before the data was finalised. All system files and data files were stored in the server under 2009PopCensus folder. Three temporary data operators were hired to do the job, under the supervision of Rara Soro, the system analyst for VNSO. No data was stored in work stations, because all data were directly written to the DATA folder in the server.

Range checks and basic checks (online edits) were built in the manual data entry system, while the complex edits were written in a separate batch edit program. If the system encounter and error during data entry, an error message will be displayed and the data operator cannot proceed unless the error displayed is fixed. e.g Males + Females = Total Persons. Please re-enter. It was strongly recommended to the data operators not to make up answers but consult the supervisor if he/she cannot fix it. Listed below are the checks that were built into the data entry system.

01 Person 1 must be the head of household
02 Sex against relationship
03 Age against date of birth
04 Marital status - Married people should be age 15+
05 Spouse should be married
06 P9, P10, P11 against village enumerated
07 Never been to school but can use internet - Is this possible
08 Check for multiple head or spouse in the household
09 Husband and wife of same sex
10 Total persons match total people in personal form
11 Total children born and live in household (F2a) against total persons total
12 Age difference of head and child is less than 13
13 Total children born (F4) against total alive(F2) + total died(F3)

A separate batch edit program was developed for further data cleaning. All online edits were also re-written in this program to make sure that all errors flagged out during data entry were fixed. Some of the errors detected are not really errors, but still requires double checking, and if the answer recorded is the correct answer, don't change it. The batch edit was performed on each batch, and also on the concatenated batch. Below is the summary list of errors generated from manual data entry data before batch editing.

MDE Error message summary
Age does not match date of birth 272
Total children born and living in household (F2a) > total in 1
Attend school full-time in P12 but also working 16
Too young for highest education recorded 14
Highest education completed does not match with grade currently attending 80

Age had the highest errors rate, and this is due to an error in the logic statement, otherwise all ages that do not match their date of birth are corrected during data entry.

The Data capturing (Scanning) and Editing process took about 6 months to be completed but then more checks were made after that to finalise the dataset before publishing the results.

During re-coding of zero's and blanks, a couple of batch edit statement written in the batch edit program were wrong, and it created errors in the scanned data. The batch edit was suppose to recode only those people that didn't answer questions P19, P23 - P25, but instead it recoded valid codes as well to blanks. This was only picked up when tables were generated and numbers were found to be so much different in manual data entry and scanned data. Another batch edit program was developed to recode and fix this problem.

Data appraisal
---------------------------

Household characteristics and basic demographic variables for the census data was used in comparision with the 1999 census data to determine the accuracy of the pilot data. Some of the key indicators used for comparision are the household size, sex ratio, educational attainment, employment status. A pyramid was also used for the comparisions.

摘要
---------------------------

每一项人口普查的关键目标均为在普查之夜统计国家内所有居民（男性、女性、儿童）的人数，并收集各类人口统计学（性别、年龄、婚姻状况、国籍）和社会经济（教育/资格；劳动力与经济活动）信息，以及与家庭和住房特征相关的数据。此统计提供了每个村庄和城镇、每个岛屿和地区的人口构成的全貌，从而允许对人口变化的趋势进行评估。

随着瓦努阿图及其太平洋岛屿邻国越来越多地拥抱基于信息的或基于证据的政策制定和决策文化，国家人口普查数据库以及提取复杂交叉表和众多重要次区域及小区域相关信息的可能性，对于满足公共和私营部门对数据和信息的日益增长的需求至关重要。

例如，在教育、卫生和人力规划方面，包括对未来人员配备、设施和计划预算的需求评估，如果没有定期的普查，将无法实现；在没有可靠的国家人口计数和更新的数据的情况下，政府在千年发展目标（MDG）承诺的背景下监测发展进程的努力也将受到极大影响，甚至可能完全无法实现。

尽管像家庭收入和支出调查、劳动力调查、农业调查和人口与健康调查等国家层面的常规调查提供了跨特定领域的重要数据和信息，但这些调查若没有国家抽样框架（普查数据提供）的支持，则无法维持或管理。此外，没有最新的统计数据（通常来自普查或基于普查数据的预测和估计），也无法计算和测量所有基于人口的发展指标，如大多数MDG指标。

现在，大部分此类信息已经过时9年（因此相当陈旧），而且在缺乏可靠的人口登记类型数据库的情况下，例如那些由功能健全的民事登记（出生和死亡）和移民记录系统提供的数据，2009年瓦努阿图人口和住房普查将提供迫切需要的、对于政策制定、国家发展规划和定期监测发展进程至关重要的人口、社会和经济统计数据。

除了实现提供更新的人口、社会和经济统计数据的一般目标和目的外，2009年的人口普查还代表了一次重大的国家能力建设实践，因为大多数参与普查的瓦努阿图国家统计局办公室（VNSO）工作人员此前均无普查经验。经过精心规划和资源配置，所有2009年普查活动都可能为VNSO工作人员提供了非常有用（且期望的）在职培训，覆盖了专业等级和责任的各个方面。它还为有限数量的专业人员提供了短期海外培训和职业实习（在SPC或ABS或其他地方），他们随后在瓦努阿图国家统计局办公室（VNSO）中指导其他工作人员。

一些VNSO的高级成员参与了1999年的普查，他们提供了内部可用的丰富经验，以及办公室在普查正式进行之前已经进行的持续调查，如家庭收入和支出调查和农业普查。VNSO还拥有在人口和人口统计学领域获得资格的专业官员，他们参与了项目，凭借这类资源，我们成功实施了2009年普查的另一项成功项目。

虽然来自SPC人口/人口计划工作人员的一些短期普查顾问任务被部署，但标准的SPC技术援助政策安排无法满足长期或重复在国内的任务。然而，其他相关捐助者被邀请进行更长期的技术援助专家的VNSO实习。

地理覆盖范围
---------------------------

2009年人口和住房普查的地理覆盖范围包括：

- 国家（瓦努阿图）
- 省份（托尔巴、桑马、彭纳玛、马拉帕、谢法、塔法亚）
- 居住岛屿（从希乌岛、托雷斯岛到安内蒂乌姆、南方岛屿）
- 统计区域（分配给每个统计员的EA）
- 村庄/城镇
- 家庭或住宅

分析单位
---------------------------

2009年人口和住房普查的分析单位包括：
- 家庭
- 个人（人口）

总体
---------------------------

普查覆盖了瓦努阿图的所有家庭和个人

数据类型
---------------------------

普查/统计数据 [cen]

数据收集方式
---------------------------

面对面 [f2f]

研究工具
---------------------------

问卷基本上分为5个部分；地理标识符、一般人口和教育问题、劳动力问题、妇女和生育问题以及住房问题。地理标识符包括村庄名称、GPS代码、EA编号、家庭编号和统计员ID。个人问题包含个人人口统计学，包括教育水平和劳动力状况。还包括关于生育年龄妇女的生育部分。所有这些问题都由'跳转模式'引导，以指导问题的提问流程。家庭问题包含了房屋材料的基本描述、租约、水源和卫生设施、能源、耐用消费品、使用处理过的蚊帐和互联网接入。

清理操作
---------------------------

在普查过程中，使用光学字符识别（OCR）系统（ReadSoft应用程序系统）从完成的表格中捕获信息。然后，捕获的数据被导出到MS Access数据库系统进行进一步编辑和清理，最后在数据最终确定之前将其传输到CSPro进行更多编辑和质量检查。所有系统文件和数据文件都存储在服务器上的2009PopCensus文件夹下。为此工作，聘请了三名临时数据操作员，在VNSO的系统分析师Rara Soro的监督下工作。没有数据存储在工作站上，因为所有数据都是直接写入服务器上的DATA文件夹。

范围检查和基本检查（在线编辑）内置在手动数据输入系统中，而复杂的编辑则编写在一个单独的批量编辑程序中。如果系统在数据输入过程中遇到错误，将显示错误消息，数据操作员无法继续，除非显示的错误得到修复。例如：男性+女性=总人数。请重新输入。强烈建议数据操作员不要捏造答案，如果无法修复显示的错误，应咨询主管。以下列出了内置在数据输入系统中的检查。

01 第一个人必须是户主
02 性别与关系相矛盾
03 年龄与出生日期相矛盾
04 婚姻状况 - 已婚人士应年龄15岁以上
05 配偶应该是已婚的
06 P9、P10、P11与已统计的村庄相矛盾
07 永远没有上过学但可以使用互联网 - 这可能吗
08 检查家庭中的多头或配偶
09 同性婚姻的夫妻
10 总人数与个人表格中的总人数相符
11 家庭中出生并居住的总孩子数（F2a）与总人数相符
12 户主与孩子之间的年龄差异小于13岁
13 总出生人数（F4）与总存活人数（F2）+ 总死亡人数（F3）相符

为进一步数据清理开发了单独的批量编辑程序。所有在线编辑也重写在程序中，以确保在数据输入过程中标记的所有错误都得到修复。一些检测到的错误并非真正的错误，但仍需进行双重检查；如果记录的答案是正确的答案，则不要更改它。批量编辑在每个批次上执行，也在连接的批次上执行。以下是手动数据输入数据在批量编辑之前的错误摘要列表。

MDE 错误消息摘要
年龄与出生日期不符 272
家庭中出生并居住的总孩子数（F2a）> 总人数 1
在P12中全日制上学但也在工作 16
对于记录的最高教育水平来说太年轻 14
最高完成的教育水平与目前就读的年级不符 80
年龄的错误率最高，这归因于逻辑语句中的错误，否则所有与出生日期不符的年龄在数据输入过程中都得到了纠正。

数据捕获（扫描）和编辑过程大约用了6个月时间完成，但在那之后还进行了更多的检查，以在发布结果之前最终确定数据集。

在重新编码零和空白的过程中，批量编辑程序中编写的几个批量编辑语句是错误的，这导致了扫描数据的错误。批量编辑本应重新编码未回答问题P19、P23 - P25的人，但反而将有效的代码也重新编码为空白。这个问题只有在生成表格并发现数字在手动数据输入和扫描数据中差异很大时才被发现。为此开发了另一个批量编辑程序来重新编码和修复这个问题。

数据评估
---------------------------

将普查数据的家庭特征和基本人口统计学变量与1999年普查数据进行了比较，以确定试点数据的准确性。用于比较的一些关键指标包括家庭规模、性别比、教育成就和就业状况。还使用金字塔进行了比较。