Asian Population Association

Workshop: Hopes for the 2010 Census Round in Asia

Welcoming Speech
Opening Session of the Workshop on Hopes for the 2010 Census Round in Asia:
National Challenges and Regional Collaboration

Professor Bhassorn Limanonda
Executive Secretary
Asian Population Association (APA)

Establishment Committee of Asian Population Association, Distinguished Participants of the Workshop

I have the pleasure to welcoming you to the one-day Workshop on the 2010 Census Round in Asia: National Challenges and Regional Collaboration organized here in Bangkok by the Asian Population Association.

As you are aware, the Asian Population Association, in short APA, has been officially established in December 2007, and gained its legal status in April 2008, through the registration in Thailand. The House of the Association is situated at College of Population Studies, Chulalongkorn University. The first meeting of the Establishment Committee of the APA followed by the first one-day Workshop on Emerging Population Issues in Asia was held at the University of Tehran, Iran during 1-3 December 2007.

The Second meeting of the Establishment Committee was held during the past two days (1-2 June 2008) in Bangkok. Today, Tuesday 3rd of June, the Second one-day Workshop on The 2010 Census Round in Asia: National Challenges and Regional Collaboration is held in Bangkok, Thailand.

The objective of the Workshop is to bring together key population professionals from countries across Asian region for a day of discussion and debate between users and producers of census data. This will be different from most technical meetings held to date in that it will take an avowedly users? perspectives and focus on issues of timeliness, clarity, accessibility and relevance of census data.

Today, there are approximately 30-35 population professionals from overseas and Thailand to participate and discuss extensively on the Workshop issues. We hope to get your expert views and experiences as data producers and users, as important contribution to improve quality of census. At the end of the day, we expect to be able to come up with important source of knowledge and wisdoms vital to all parties involved. I would like to sincerely thank all the experts who gather here, who have willingly and spontaneously accepted our invitation to Bangkok, especially those who received the invitation from the association within a very short notice.

The APA is just ?a new born baby? who needs to be well nurtured. The Establishment Committee and the international advisors have contributed in a big and small way to raise this baby to grow and strong to be on its own feet. Since the beginning of this year, Professor Terry Hull has been working very hard looking for funding from various sources to support for APA Activities including the meeting and Workshop held in Bangkok, Dr.? Mohammad Jalal Abbasi-Shavazi, the EC Chair works vigorously on recruiting founding members of Association, other EC members including Professors Baochang Gu, Monneer Alam and Mehtab Karim? have continuously shared ideas, advice, comments as well as to provide moral supports. Under the simple but elegant APA logo designed by the designer Saeid Azimi from Tehran-e Parantez Company under the scrutiny of the EC Members, I am sure that the Association will strongly move forward to reach the goals set since its inception in December 2007.

Last but not least, on behalf of the Asian Population Association, (APA), I would like to express most sincere thanks to the funding agencies, UNFPA, New York, Ford Foundation, through the kind assistance of the Demographic and Social Research Institute, Australian National University, and MacArthur Foundation which provide financial support, and College of Population Studies which provide necessary support on logistic arrangements to make the meeting and Workshop organized in Bangkok possible. Thank you.

In the folder, besides the information about the Workshop, there is piece of important information that the association needs your kind help disseminating to a wider circle for membership drive. We have two pages of information about the Association Website and method of applying to be a member. The Association currently has over 300 members who applied during the past three months when the website was launched. With your kind cooperation, we expected to see much larger number of members including yourself.

Thank you once again for your participation.

Workshop
Hopes for the 2010 Census Round in Asia:
National Challenges and Regional Collaboration

Sponsored by the Asian Population Association

3 June 2008 (Tuesday)
Bangkok, Thailand

With the establishment of the Asian Population Association this year professional population analysts in Asia now have a forum to hold discussions of demographic trends, developmental impacts, and social analysis with a regional perspective. National censuses are widely regarded as the most important source of demographic data available to governments and individual researchers.

Census operations face financial, logistic and methodological constraints that limit the scope and depth of data that can be made available. For this reason users of census data need to make clear their priority data needs, and justify their choices with clear reasons. This workshop will bring together researchers and data producers to discuss priority data needs and work through alternatives strategies during the 2010 Census round across Asia.

Format: There will be an introductory panel, followed by four roundtable discussions and a concluding session. Each will be chaired by a member of the APA Establishment Committee. The named participants in each session will open discussion with brief presentations to pose the key questions relevant to the 2010 Census round. The Chairs will then encourage contributions from the floor.

0830 ? 0900 Opening Session. (Chair: Bhassorn Limanonda). Introductions of the participants and acknowledgment of the sponsors. Overview of the purposes of the day?s discussions.

Roundtable? Discussions

0900 ? 1000 Global and Asian Census Innovations and Challenges in 2010 (Gavin Jones) Panel members will have 12 minutes each:
? Haishan Fu, Chief, Statistics Development Section, ESCAP (Innovations to improve data quality).
? Uzair Suhaimi (From forms to data sets ? using computer technologies for censuses)
? Rajana Netsaen (Topic on the Thai census preparations)

1015 ? 1100 Fertility (Chair: Jalal Abbasi-Shavazi). The countries around the region rely on a variety of methods to monitor fertility levels and trends. Most common among these are census estimates using ?Own Child? techniques and DHS style direct questions on pregnancy histories. Both produce fertility estimates relevant for periods from one year to a decade before the enumeration. What can we do to develop better estimates of fertility from the 2010 Census? Roundtable discussion to be initiated by three five minute comments:
? Jalal Abbasi-Shavazi ? Own child reliability in Iran
? Zhigang Guo ? China?s difficulties in resolving differences in TFR estimates
? Apichat Chamratrithirong? ? Census versus surveys in Thai fertility estimates

1100 ? 1200 Mortality (Chair: Mehtab Karim). Censuses in developing countries produce a variety of data for the calculation of estimates of infant and child mortality, from which expectations of life are often inferred. In countries with good coverage of vital registration, the census is used to generate denominators for the calculation of mortality rates. In recent years there have been calls for questions on maternal mortality to be included in census enumerations. Are current approaches to mortality estimation robust? Are they adequate to the demands of monitoring MDGs? Roundtable discussion to be initiated by seven minute comments:
? Wendy Hartanto ? Maternal mortality estimates from censuses
? Sri Moertiningsih Adioetomo ? Maternal mortality data from the Lao census

1330 ? 1500 Labour Force and Migration (Chair: Moneer Alam). While censuses are notionally meant to take a detailed snapshot of the entire population of a nation, we know that there are often problems of finding members of the defined population, and even when found, some do not respond to the enumerators. Problems of coverage and response are particularly worrisome for analysts interested in labour force and residential characteristics of the population. Will the 2010 round improve the coverage and accuracy of the enumerations?
? Uzair Suhaimi ? Coordinating census results with national survey results; Problems with work and occupational categories
? Sri Moertiningsih Adioetomo ? Who is working and who is employed? Does it make a difference to dependency rates and the Demographic Window?

1515 ? 1645 Education, Marriage and Family (Chair: Baochang Gu). While the census is collection of individual level data, the process of enumeration focuses on households and families. In recent years it has been found that one of the most powerful drivers of social change is the level of schooling attained by individuals. Across Asia governments and researchers have found education to be a key variable for both policy analysis and social programs. How well are we capturing information about individuals, families, households and that important dimension of education?
? Preeya Mitranon ? issues of measuring experience of schooling among adults
? Josie Cabigon ? local and international definitions of marriage
? Sri Moertiningsih Adioetomo ? Will education levels reported in the Census match the schooling enrolments in the Ministries of Education?

1645 ? 1700 Conclusions and Recommendations (Chair: Terence Hull).

1830 All invited for dinner at Chulanaruemit Hall, Chulalongkorn University

Participants depart for their home countries on Wednesday June 4.

Name List of Participants
of
Workshop on The 2010 Census Round in Asia:
National Challenges and Regional Collaboration

Organized by the Asian Population Association (APA)

June 3, 2008
At Prakai-Petch Room, Asia Hotel, Bangkok

Report of the Workshop on

Hopes for the 2010 Census Round in Asia:
National Challenges and Regional Collaboration

Organized by the Asian Population Association

3 June 2008 (Tuesday)
Asia Hotel, Bangkok, Thailand

National censuses are widely regarded as the most important source of demographic data available to governments and individual researchers. Census operations face financial, logistic and methodological constraints that limit the scope and depth of data that can be made available. For this reason users of census data need to make clear their priority data needs, and justify their choices with clear reasons. This workshop brought together researchers and data producers to identify problems faced in past censuses and discuss priority data needs and for the 2010 Census round across Asia.

Opening Session. (Chair: Bhassorn Limanonda). Introductions of the participants and acknowledgment of the sponsors.

Jalal Abbasi-Shavazi (Chair of the APA Establishment Committee) reported on the results of the EC meetings, and encouraged all participants to visit the APA website to submit applications for membership (http://www.apa.cps.chula.ac.th).

Roundtable? Discussions

Session 1 Global and Asian Census Innovations and Challenges in 2010 (Chair: Mehtab Karim):
? Haishan Fu, Chief, Statistics Development Section, ESCAP (Innovations to improve data quality).
? Uzair Suhaimi (From forms to data sets ? using computer technologies for censuses)
? Rajana Netsaen (Thai census preparations)

ESCAP sponsored meetings have identified a number of serious difficulties in the 2000 Census round, including: insufficient planning, limited dissemination of results, technical problems with scanning of forms, and overall a widespread shortage of budgets at all levels, in part due to the Asian Financial Crisis of 1997-2000.

One approach to dissemination of individual records from census enumerations is the Microdata Management Toolkit found on: http://www.surveynetwork.org/home/ website. It was noted that CSPro is the most popular program for processing census data. It is free and can be downloaded easily (see http://www.cspro.org ).

Uzair Suhaimi told of the difficulties of machine readable questionnaires for data entry in Indonesia. In 2000 major problems arose from using optical reading technologies. As a result Statistics Indonesia has experimented with new ?intelligent? software that adapts to different types of handwriting. They will also invest in better systems of supervision and data analysis to ensure that 2010 results will be more accurate and timely.

Inadequate legislative frameworks for census operations continue to hamper statistical offices across Asia. Parliaments tend to lack an awareness and understanding of technical issues associated with large scale household enumerations. At the same time politicians frequently make demands for particular questions, and even for particular results. This can undermine the preparation of high quality data. In addition, national security systems and population registration systems run by local governments can lead to serious biases in population census counts in urban areas where unregistered migration is frowned upon by officials. There needs to be a greater effort to achieve independent estimates of census populations through the application of de facto definitions of residence.

Session 2 Fertility (Chair: Jalal Abbasi-Shavazi). The countries around the region rely on a variety of methods to monitor fertility levels and trends. Most common among these are census estimates using ?Own Child? techniques and DHS style direct questions on pregnancy histories. Both produce fertility estimates relevant for periods from one year to a decade before the enumeration. What can we do to develop better estimates of fertility from the 2010 Census? Roundtable discussion was initiated by three presentations:
? Zhigang Guo ? China?s difficulties in resolving differences in TFR estimates
? Apichat Chamratrithirong? ? Census versus surveys in Thai fertility estimates
? Jalal Abbasi-Shavazi ? Own child estimates in Iran

East Asia has increasingly suffered from governmental dissatisfaction with fertility levels recorded in the census. In China national fertility rates have been consistently overestimated since the early 1990s. Arbitrary ?adjustments? and inappropriate data collection procedures (sometimes using family planning fieldworkers to collect information) have been promoted by officials who wanted to ensure continued support for a mass fertility control campaign. Ironically they have continued to insist that fertility was ?around replacement level? when true rates were closer to 1.5 births per woman ? levels similar to Korea, Taiwan, and Hong Kong. While this should have been a sign of success in achieving fertility reduction, instead it was interpreted as a threat to the agencies responsible for family planning. If fertility was 1.5 they feared that government would cease supporting their efforts.

Use of a variety of indirect fertility measures such as the Own Child (OC) Method can help to verify the true levels and trends of fertility, both nationally and sub-nationally.? This requires the use of individual household records to link all women of childbearing age with their own living children, usually being co-resident.? The data are then adjusted for deaths to obtain the total number of births in the recent past, thus allowing the calculation of estimates of age specific fertility and total fertility for the years prior to the census. The OC method was used to generate estimates for Iran?s fertility decline in very small geographic units. It is also the most common census based method used across Asia.

Toening Adioetomo and Josie Cabigon noted that the OC estimates in Indonesia and the Philippines yielded very different levels and trends than those coming from the pregnancy histories of the Demographic and Health Surveys (DHS). In Indonesia the DHS estimates have been consistently higher than OC (2.6 versus 2.3 in the most recent enumerations), while in the Philippines the OC census estimates are higher than the DHS. In Iran the two types of measure are consistent. This stands as a warning to analysts that trend analysis should be restricted to a single measure.

Session 3 Mortality (Chair: Mehtab Karim). Censuses in developing countries produce a variety of data for the calculation of estimates of maternal, infant and child mortality, from which expectations of life are often inferred. In countries with good coverage of vital registration, the census is used to generate denominators for the calculation of mortality rates. In recent years there have been strong calls for questions on maternal mortality to be included in census enumerations. Are current approaches to mortality estimation robust? Are they adequate to the demands of monitoring MDGs? Roundtable discussion was initiated by comments from two speakers:
? Wendy Hartanto ? Maternal mortality estimates from censuses
? Sri Moertiningsih Adioetomo ? Maternal mortality data from the Lao census

Demographers have long been under pressure to produce robust measures of maternal mortality. Since 2000 this demand has intensified as world leaders chose to consider Maternal Mortality Ratios (MMR) as the key indicator for Millennium Development Goal number 5 ? to improve maternal health (http://www.un.org/millenniumgoals/) which is defined as ?reduce by three quarters the maternal mortality ratio? by the year 2015.

The most recent efforts to monitor MMR involve direct questions about the deaths of members of a household in the last year or two. By further asking about the situation surrounding deaths of adult women, these questions are meant to allow identification of all maternal deaths, which can then be compared with estimates of births to give the MMR. Wendy Hartanto reported that plans under way in Indonesia to apply this procedure appear to have major problems due to tendencies to underreport both deaths and births. In addition, deaths can result in the total disappearance of single person households and reconstitution of multi-person households, producing confusion when people are asked about deaths in the current household during the census enumeration.

Participants noted the importance of improving civil registration systems, but this is a long-term goal rather than an immediate answer to the need to improve the 2010 census round. Three suggestions were tabled to improve data quality in the census:

1. Multiple methods should be used to generate mortality and fertility estimates from census data. Comparisons should be made between the full-count and sample census based estimates, and inconsistencies should be discussed publicly so political debate will be better informed and more realistic.
2. Efforts to measure MMR should not be made at small geographic levels since the difference of only one or two deaths in the data collection can have a large impact on the estimate of MMR. All estimates should be presented with standard errors of the estimates.
3. The census form should record the identity of the person who answers the questions on the household form. This would assist in the analysis of sources of incorrect or inconsistent information.

Session 4 Labour Force and Migration (Chair: Moneer Alam). While censuses are notionally meant to take a detailed snapshot of the entire population of a nation, we know that there are often problems of finding members of the defined population, and even when found, some do not respond to the enumerators. Problems of coverage and response are particularly worrisome for analysts interested in labour force and residential characteristics of the population. Will the 2010 round improve the coverage and accuracy of the enumerations?
? Uzair Suhaimi ? Coordinating census results with national survey results; Problems with work and occupational categories
? Sri Moertiningsih Adioetomo ? Who is working and who is employed? Does it make a difference to dependency rates and the so-called Demographic Window?

Contradictions between detailed labour force surveys and broad based or full count census data are inevitable given the complexity of the concept of work, occupation and industry. The concept of labour force goes back to the 1930s when national economies were grappling with global depression. Since then there have been major changes in the nature of the economy and it is increasingly difficult to make sense of concepts of unemployment and poverty without a lot of very clear questions that are sufficient to capture the full meaning of economic relations. Full count censuses cannot fit so many questions.

Uzair Suhaimi argued forcefully that the 2010 census round should not attempt to capture all the nuances that are needed for planning. Instead it should stick to simple questions that ensure an accurate count of the population (through full coverage, including all migrants as well as regular residents) and sufficient descriptive information to give rough position of individuals in different sectors of the economy and school system. Toening Adioetomo reminded participants that the definition of work is often very weak in any case, with employment in Indonesia requiring only one hour of work in the preceding week.

At a time when millions of Asians are working outside their home province or even outside their country of citizenship, the census suffers greatly from definitions of residence used to list members of a household. The reliance on de jure definitions or even mixtures of de jure and de facto leads to serious risks of large under-enumerations and mis-enumerations of individuals, particularly those in the young working ages. Movement to more purely de facto counts would help to overcome these problems.

Despite the complexity of these residential and economic concepts, government officials and local leaders are particularly insistent that the census or very large surveys attempt to capture this information for districts and sub-districts.

Session 5 Education, Marriage and Family (Chair: Baochang Gu). While the census is collection of individual level data, the process of enumeration focuses on households and families. In recent years it has been found that one of the most powerful drivers of social change is the level of schooling attained by individuals. Across Asia governments and researchers have found education to be a key variable for both policy analysis and social programs. How well are we capturing information about individuals, families, households and the important dimension of education?
? Chintana Pejaranonda ? issues of measuring experience of schooling among adults in Thailand
? Josie Cabigon ? Local and international definitions of marriage
? Terry Hull ? Will education levels reported in the Census match the schooling enrolments in the Ministries of Education?

Social variables on the census are subject to great variation and complexity.

In many countries very simple but misleading questions are asked about marriage. Women are categorized as single (meaning never married), married, divorced or widowed. However in Southeast Asian cultures there are different types of formal and informal marriage and some women who are not in such sustained relationships may still be mothers. Josie Cabigon pointed out that marriages may reflect very different types of motivations, including procreation, companionship, cohabitation, and partnership for economic purposes. Unions can also take the form of ?living apart together? as has been discussed in the Western sociological literature. While the demographic literature places great emphasis on the age at first marriage in studies of fertility and household formation, the realities of daily lives indicate that this is not always easy to measure.

The experiences of East and Southeast Asian countries in censusing households points to increasing importance of dormitories and other institutional and commercial forms of residence for young students and workers in both cities and the countryside. As these become more common the distinction between ?family? and ?household? becomes all the more important, yet it is very common to see reference to families in newspaper articles, government reports and daily conversation talking about census enumerations.

Education is a key variable in census collections. Adults are categorized by educational achievements and young people are asked about current school enrolment. At times questions are asked about literacy and language skills, both of which are crucial in matters of occupational planning and definition of welfare needs of a population. Unfortunately there are often large inconsistencies between the information reported by the school authorities and the data reported on the census. In Indonesia this was partially related to the fact that the census is conducted during school holidays (in part to take advantage of using teachers as interviewers) so when asked if young children are attending school parents sometimes answer no, a truthful but misleading response.

Session 6 General Discussion (Chair: Baochang Gu).
Baochang Gu set out four principles that arose from the workshop:

1. The purpose of the census is not just to carry out an accounting exercise, but rather needs to lead to valid interpretations of statistics that are acknowledged to inevitably contain errors. The accuracy of the census should be judged to see if it helps or misleads our understanding of trends, patterns, and situations. Part of the purpose is to provide the information needed to hold government agencies accountable. For this reason the census authorities need to be autonomous, independent and adequately funded.
2. In serving the people?s interests the census needs to be transparent and accessible. Too often the results are withheld on spurious grounds. This needs to be subject to international standards and monitoring.
3. The variables covered by the census need to target issues of importance that can be measured reliably and validly. There are many important issues, like maternal mortality, that may not be reliably measured by full-count censuses.
4. In the census more can mean less. There are real conflicts between the total number of variables included and the quality of the answers provided.