The Census is the official count of population and dwellings in Tonga, providing a ‘snapshot’ of the society and its most precious resource, its people, at a point in time. The official reference period of the census was midnight, the 30th of November, 2006.
The census provides a unique source of detailed demographic, social and economic data relating the entire population at a single point in time. Census information is used for policy setting and implementation, research, planning and other decision-making. The census is often the primary source of information used for the allocation of public funding, especially in areas such as health, education and social policy. The main users of this information are the government, local authorities, education facilities (such as schools and tertiary organizations), businesses, community organizations and the public in general.
The 2006 Census was taken under the authority of Section 8 of Statistical Act Chap. 53 of 1978 which empowers the Minister of Finance to make regulations necessary to conduct the population Census. This regulation was approved by the Cabinet and cited as Census Regulation 2006. The Census regulations also indicate that the Government Statistician would be responsible for the administration and completion of the Census. In addition, the regulations enabled the Statistics Department to carry out the necessary activities required to plan, manage and implement all the necessary Census activities.
Census planning and management
From a planning and management perspective, the Census had two main objectives. Firstly, it was to ensure that the process of collecting, compiling, evaluating, analyzing and disseminating of demographic, economic and social data was conducted in a timely and accurate manner. The development of procedures and processes for the 2006 Census of Population and Housing made use of the lessons learned in previous censuses, and built upon recommendations for improvements.
Secondly, it was a valuable opportunity for building the capacities of employees of the Statistics Department (SD), thus resulting in enhancing the image, credibility and reputation of the Department and at the same time, strengthening its infrastructure. Emphasis was placed on having a senior staff with a wide perspective and leadership qualities. Through the use of vision, planning, coordination, delegation of responsibility and a strong team spirit, the census work was conducted in an effective and efficient manner. Staffs at all levels were encouraged to have an innovative mindset in addressing issues. Incentives for other parties to participate, both within Statistics Department Tonga Tonga 2006 Census of Population and Housing viii and outside the government, were encouraged. As a result, the wider community including donors such as AusAID, the Secretariat of the Pacific Community (SPC) in Noumea, that provided the technical assistance and the general public, were able to support the census project.
Extensive and detailed planning is needed to conduct a successful census. Areas that required planning include: enumeration procedures and fieldwork, public communication, data processing and output systems, mapping and the design of census block boundaries, dissemination procedures, content determination and questionnaire development and training. These aspects, and how they interacted with each other, played a crucial role in determining the quality of all of the census outputs. Each phase therefore required careful, methodical planning and testing. The details of such activities, and their implementation and responsibilities were assigned to 5 subcommittees composed of staff members of the SD.
Organizational structure of the Census
A census organizational structure is designed to implement a number of interrelated activities. Each of these activities was assigned to a specific sub-committee. The census manuals provided guidelines on processes, organizational structures, controls for quality assurance and problem solving. The challenge for managers was developing a work environment that enabled census personnel to perform all these tasks with a common goal in mind. Each sub-committee was responsible for its own outputs, and specific decisions for specific situations were delegated to the lowest level possible. Problem situations beyond the scope of the sub-committee were escalated to the next higher level.
The organizational structure of the census was as follows:
a) The Steering Committee (consisting of the Head of both Government and nongovernment organizations), chaired by Secretary for Finance with the Government Statistician (GS) as secretary.
b) The Census Committee (consisted of all sub-committee leaders plus the GS, and chaired by the Assistant Government Statistician (AGS) who was the officer in charge of all management and planning of the Census 2006 operations.
c) There were five Sub-committees (each sub-committee consisted of about 5 members and were chaired by their Sub-committee leader). These committees included: Mapping, Publicity, Fieldwork, Training and Data Processing. In this way, every staff member of the SD was involved with the census operation through their participation on these committees.
The census steering committee was a high level committee that approved and endorsed the plans and activities of the census. Policy issues that needed to be addressed were submitted to the steering committee for approval prior to the census team and sub-committees designation of the activities necessary to address the tasks.
Part of the initial planning of the 2006 Census involved the establishment of a work-plan with specific time frames. This charted all activities that were to be undertaken and, their impact and dependencies on other activities. These time frames were an essential part of the overall exercise, as they provided specific guides to the progress of each area, and alerted subcommittees’ team leaders (TL) to areas where problems existed and needed to be addressed. These also provided the SD staff with a clear indication of where and how their roles impacted the overall Census process.
Monitoring of the timeframe was an essential part of the management of the Census program. Initially, weekly meetings were held which involved the GS, AGS and team leaders (TL) of the Census committee. As the Census projects progressed, the AGS and TL’s met regularly with their sub-committees to report on the progress of each area. Decisions were made on necessary actions in order to meet the designated dates. Potential risks that could negatively affect the deadlines and actions were also considered at these meetings.
For the 5 sub-committees, one of their first tasks was to verify and amend their terms of reference using the “Strengths, Weaknesses, Opportunities and Threats” (SWOT) analysis methodology, as it applied to past censuses. Each committee then prepared a work-plan and listed all activities for which that particular sub-committee was responsible. This listing included the assignment of a responsible person, together with the timeline indicating the start and end dates required to complete that particular activity. These work-plans, set up by all the 5 sub-committees, were then used by the AGS to develop a detailed operational plan for all phases of the census, the activities required to complete these phases, start and end dates, the person responsible and the dependencies, - all in a Ghant chart format. These combined work-plans were further discussed and amended in the Census team and reported to the Steering committee on regular basis as required.
Kind of Data
Census/enumeration data [cen]
Version 01 - Edited, cleaned, de-identified and labelled version of the Master file.
Population & Reproductive Health
National coverage, which includes the 5 Divisions and both Urban and Rural Areas of Tonga.
Unit of Analysis
Individual and Households.
All individuals in private and institutional households.
Producers and sponsors
Authoring entity/Primary investigators
Tonga Statistics Department
Government of Tonga
Government of the Kingdom of Tonga
Roman Catholic Diocese of Tonga
New Zealand AID
Funding for digitisation of maps
His Majesty's Cabinet
Approval of the Population Census Proposal 2006 and the Amendments of Census Regulation 2006
Minister of Finance
Managed the implementation of the Population Census Proposal 2006 and the Amendments of Census Regulation 2006
The Census Steering Committee
Management of the Census Operation
Ministry of Lands, Survey and Natural Resources.
Resources Provided the technical assistance to update and digitise census maps
Ministry of Education
Provided over 500 primary school teachers and education officers to carry out census enumeration and supervision of census fieldwork
The National Population Census was a complete enumeration census, hence no sampling procedure was employed. A Mapping Sub-committee was formed to ensure complete coverage of the country.
The Mapping Sub-committee
Led by Mr. Winston Fainga'anuku, this committee's mandate was to ensure that good quality maps were produced. The objective was to ensure that the maps provided complete coverage of the country, were designed to accommodate a reasonable workload of one census enumerator and, that geographic identifiers could be used for dissemination purposes by the PopGIS system. Collaborations with the Ministry of Land, Survey and Natural Resources (MLSNR) began in 2004 to ensure that digitized maps for Tonga could be used for 2006 Census. Mr. Fainga'anuku was attached to the MLSNR in April 2005 to assist 'Atelea Kautoke, Samuela Mailau, Lilika and others to complete the task of digitizing the maps for Tonga. In addition, frequent visits by Mr. Scott Pontifex from the Secretariat of the Pacific Community (SPC) in Noumea, assisted to ensure that quality digitized maps were prepared. SPC also assisted by lending its digitizer which was used in this mapping project. The staff of the Statistics Department (SD) visited household sites throughout Tongatapu and the main outer islands. This exercise was to redesign the Census Block boundaries by amalgamating or splitting existing census blocks to achieve an average of 50 households per census block. Various updates within the census block maps were made. These included the names of the head of household; roads and other landmarks to ensure that current and accurate information was provided to the enumerators. Reliable maps, both for enumerators and supervisors are necessary ingredients to assist in avoiding any under or over - counting during enumeration. Included in the package of maps was: PCmap1 which indicated where the census block was located; PCmap 2 indicating details about the boundaries of the census blocks, along with the numbers of the households within the block; and a list of the head of each household on the map.
Dates of Data Collection (YYYY/MM/DD)
Time periods (YYYY/MM/DD)
Mode of data collection
Supervisors were also selected from Primary School Teachers.
Type of Research Instrument
To meet the users’ needs, data quality requirements and the ability of the public to provide information, the design and wording of the census forms needed careful development and testing. The census questionnaire for 2006 consisted of two forms; A Household Form PC1 and, a Personal Form PC2.
The census team reviewed the Pacific Islands model questionnaire together with the 1996 census questionnaires, as the first step in formulating the 2006 Census questionnaire form. These two forms were distributed to all stakeholders, including Government and non-government organizations and members of the steering committee, to assist them in identifying their policy issues that could be addressed by questions on these two forms. They were then asked to bring forward their unaddressed policy issues to the SD or the steering committee meetings for consideration.
It was made clear that the content of the questionnaire form was to be discussed and finalized within the steering committee with stakeholders, whereas the questionnaire design was to be the role of the SD in collaboration with SPC. After many discussions regarding the questionnaire content in both the steering committee and census committee, the AGS then brought this proposed content to a Pacific census planning meeting in Nadi, where questionnaire design was to be addressed. At the meeting, representatives from ABS, Stats NZ, UNSD, SPC and other Pacific countries discussed the questionnaire content and design. The AGS brought back the outcome of the meetings to SD and they then, together with SPC, produced draft questionnaires for Tonga. This draft was adapted to address Tonga’s national needs for census information. As was mentioned, this draft questionnaire form was a product of the Pacific model questionnaire, Tonga’s 1996 questionnaire form (comparability), policy issues from stakeholders and the outcome of the Pacific Islands census planning workshop in Nadi. After some minor amendments, the steering committee approved the content of the questionnaire form and then asked the SD and SPC to determine any design details.
The Tonga Statistics Department staff conducted 3 informal pilot tests of the questionnaire form. At the same time, the SPC was advising on various aspects of the form to incorporate slight changes so that questions were clearer and thus permit improved data capture. Some changes resulted from the different perceptions that the respondents had regarding the interpretation of the questions that SD and SPC had not identified before the pilot test. As a result of the pilot tests, the questionnaire form was significantly improved and a number of potential problems were resolved prior to the actual enumeration. During this time, feedback from the data processing perspective also contributed to an improved questionnaire form. The final and formal pilot test was done with a sample selection of the census blocks and was also used to estimate the amount of time each household would require for enumeration.
Tonga Statistics Department
Government of Tonga
Primary School Teachers
Ministry of Education
Ministry of Education
CSPro was the software used for data editing.
Data processing sub-committee
Led by Mr. Sione F. Lolohea, this committee undertook the task of validating the responses where necessary and converting them into data for statistical analysis and dissemination. It was necessary that the various phases of the data processing were planned and implemented to ensure that the questionnaire was designed and formatted to permit accurate and complete data capture. Field checks and record forms were used for checking, verification and recording during the enumeration. Quality assurance procedures, such as those described above, were implemented to ensure that the data was recorded accurately so as to minimize any inconsistency. This was especially relevant in the interactions between Supervisors and Enumerators during the field operation. It was also used to produce a quick preliminary total population count, which was published before Christmas of 2006.
The designing of data entry screens, data dictionary, edit specifications, batch edits, tabulation plans and all related data processing systems were implemented comprehensively and according to schedule. During these processes, the SPC data processing technical advisor, Ms Leilua Taulealo worked closely with Mr. Lolohea and his team, in providing technical assistance where necessary. In addition, coding manuals were devised to code the relevant questions as part of the data processing exercise.
Once the data processing systems were ready, the coding of information began in January 2007 followed by the data entry. During these activities, recontact with households with problem forms was conducted either by phone or personal contact to ensure that the responses on the questionnaires were as accurate and reliable as possible. Following the data entry, editing and tabulations the analysis was conducted by Mr. Andreas Demmke of the SPC in collaboration with Mr. Sione Lolohea. The 2006 Census report is composed of two volumes; the Aministrative and the Basic Tables and the Analytical reports.
The software used for data entry was CSPro.
Estimates of Sampling Error
Not applicable to a full enumeration census.
Post – Enumeration Survey (PES)
For the post-enumeration survey (PES) 5 census blocks were selected: 2 blocks from the urban areas (Kolofo’ou-Pahu, Kolomotu’a - Longolongo), and 3 blocks from the rural areas (Vani, Nukunuku, and Neiafu). The PES count in these areas recorded (slightly) less people than counted in the actual census enumeration. While the discrepancy in population numbers of the PES compared to the census enumeration could be explained by either an undercount of the PES, or an over count (double count) of the census population, the most likely explanation is the high mobility (emigration) of the Tongan population during period of the census and PES count.
In the datasets, several variables contain more values than what is stipulated in the questionnaire (Person dataset: "p132": type of school attending, "p19": reason for not working). In addition, there will also be a "0" code and "9", they essentially stand for missing value, but needs to be confirmed from the NSO.
"Statistics Department of Tonga, Population and Housing Census 2006 (PHC 2006), Version 01 of the licensed datasets (February 2008), provided by Pacific Microdata Library. http://pdl.spc.int/index.php/home"
Disclaimer and copyrights
The user of the data acknowledges that the Statistics Department Tonga and the Secretariat of the Pacific Community bear no responsibility for use of the data or for interpretations or inferences based upon such uses.
(c) 2008, Statistics Department Tonga & SPC
Tonga Statistics Department
Government of Tonga
Statistics Department Tonga (SDT)
Documentation of the Census
Statistics for Development Division
Documentation of the study
Date of Production
Version 01 (July 2012): First attempt at documenting the 2006 Tongan census. Done by Statistics Department of Tonga.
Version 02 (January 2019): This is the edited version of the first documentation of Tonga Census of Population and Housing 2006.