Government Statistics Research Paper

This research paper provides a general review of government statistics largely in the context of the USA, but it also addresses broader issues relevant to other countries. The focus is on the national government and, therefore, mostly on national statistics, but much is also relevant to governments at subnational levels, which also produce statistics.

The paper begins with the purposes of government statistics and the rationale for why governments produce them. Next it takes up the principles by which the statistics should be produced to serve these purposes effectively; this is followed by descriptions of various types of data collection, and of how the responsibilities for collection, analysis, and reporting of statistics are organized among statistical agencies. Effective practices of these agencies are identified. The paper concludes with a discussion of some challenges faced by statistical agencies.

1. What Are Government Statistics?

‘Government statistics’ means factual information that is reported upon for statistical purposes by a unit of government. Some countries also use the term ‘official statistics.’ In this research papers, ‘statistics’ and ‘data’ are used interchangeably to mean factual information, and the terms are not restricted to quantitative information. Usually, the unit or agency of government that issues the statistics has collected or compiled the data to produce them. In most cases, the agency is a distinct entity of government entrusted with reporting the statistics. In some cases, the agency may analyze its data to provide more information about the statistics in its reports, especially information that makes the statistics more relevant to public policy concerns. For the agency to be effective, the focus of any analysis must be on ways to improve the relevance and accuracy for use in public policy analysis and decision making, independently of any particular policy.

The sine qua non of the definition of government statistics is that they are reported upon only for statistical purposes:

‘Statistical purposes include description, evaluation, analysis, inference, and research. For these purposes, a statistical agency may collect data from individuals, establishments, or other organizations directly, or it may obtain data from administrative records, but it does not do so for administrative, regulatory, or enforcement purposes. Statistical purposes exclude any interest in an individual person or economic unit. The data are used solely to describe and analyze statistical patterns, trends, and relationships involving groups of persons or other units’ (Martin et al. 2000)

2. Why Do Governments Produce National Statistics?

A government provides statistics to inform its citizens, satisfying an essential requirement of a democracy (see, e.g., United Nations Economic and Social Council 1994). The government, however, is most often the primary user of the statistics it produces. The statistics are used in the formulation of, and decision making on, public policies, as well as in their implementation and evaluation. And the statistics are used for much private-sector and individual decision making. Information from statistics is, for example, vital to inform economic markets.

In most cases, a government collects and reports the statistics itself, either directly or through the services of a private organization. In some cases, the private sector satisfies the demand for national statistics. Examples are the Conference Board’s compilation of US leading economic indicators, abortion statistics for many countries compiled by the Alan Guttmacher Institute, and many private market-research surveys and reports of economic forecasting firms. But even when the private sector can do so, there are reasons for a government to collect the statistics. Alternative sources may not exist. Timeliness may be a critical factor. Statistics collected by government agencies are often deemed to be of greater quality. And the government may collect statistics because they contain confidential information.

A more compelling reason may be trust and credibility. Potential subjects of government data collection may be more willing to volunteer their time and effort, and divulge confidential information, for a government survey rather than for a private one. And the public and other important users may find data collected by the government to be more credible.

A fundamental reason, however, for a government to collect data has to do with the very nature of data. If a commodity can be consumed by many and consumption by one does not reduce the consumption of others, then the commodity is a ‘public good.’ Data are manifestly public goods: use of data by one party in no way diminishes their usefulness to other parties. Individuals benefit from a public good but, unless each individual is charged for the consumption of the good, they are unlikely to pay for them voluntarily. Consequently, the goods may not be produced, unless the government intervenes. In the case of data, the government may intervene directly by collecting the data, or it may intervene indirectly, for example, by defining rights for a license or copyright to provide the means for the private sector to charge for the value added in producing secondary data. For these reasons, a government typically provides primary data and the information infrastructure. The private sector typically provides secondary data and may build on the infrastructure.

3. Principles Of Government Statistics

The US National Academy of Sciences–National Research Council report (Martin et al. 2000) defines a ‘statistical agency’ as a unit of government whose principal function is the compilation and analysis of data, and the dissemination of information for statistical purposes. The unit is recognized generally as a distinct entity, which may be located in a cabinet department or independent agency. The mission of a distinct entity or agency of government that issues government statistics may or may not be solely statistical. The USA has a highly decentralized statistical system, with over 70 agencies having direct funding of over $500,000 per year for statistical activities. Only some of these agencies are statistical agencies.

Statistical agencies are established to be a credible source of accurate, timely, and relevant statistics in one or more subject areas. Accuracy requires proper concern for consistency in spatial and temporal dimensions, as well as statistical measures of error. Timeliness requires concern for the frequency of issuing data, as well as disseminating data to users when they need them. Relevance requires concern for anticipating future data needs, as well as meeting current ones. And credibility requires concern for both the reality and appearance of impartiality, as well as the reality and appearance of independence from political control. It is essential to the mission of statistical agencies to work to ensure the goals of accuracy, timeliness, relevance, and credibility of statistical information.

Martin et al. (2000) identify three principles for a statistical agency. First is that the agency is in a position to provide information relevant to issues of public policy. That requires the agency to be knowledgeable about public policy issues and be able to provide objective information that is relevant to the needs of formulating, implementing, and evaluating public policies. Second is that the agency has a relationship of mutual respect and trust with those who use its data and information. To develop that relationship, an agency must strive to maintain credibility for itself and for its data. In particular, the agency must be perceived to be free of political interference and policy advocacy. Other aspects that contribute to credibility are a commitment to quality and professional standards, and openness about all aspects of the data, especially the quality of the data.

The third principle is that the agency have a relationship of mutual respect and trust with respondents who provide data, and with all data subjects whose information it obtains. Statistical agencies rely in large part on information supplied by individuals and/organizations outside the federal government, such as state and local governments, businesses, and other organizations. Some of this information is required by law, such as employers’ wage reports, and some of it is compiled from records of government programs. But much of it is obtained through the voluntary cooperation of respondents to surveys. Even when the response to a survey is required by law, cooperation by respondents promotes accuracy (Edmonston and Schultze 1995), which, in turn, promotes the credibility of the agency’s data. In encouraging cooperation, it is important to statistical agencies that respondents perceive that the information they are being asked for is important; that it is being collected in an impartial, professional manner; and that their personal information will remain confidential.

Confidentiality of individual responses means that the information will be used only for statistical purposes and that, in particular, it will not be used for any administrative, regulatory, or enforcement purpose regarding the individual. Statistical agencies go to great lengths to maintain the confidentiality of individual responses (see Duncan et al. 1993, Mackie and Bradburn 2000). Some, such as the US Bureau of the Census, have legislation to protect confidentiality, and some, such as the US Bureau of Labor Statistics, have a history of successful reliance on executive order, custom, and case law. A Federal Statistical Confidentiality Order (US Office of Management and Budget 1997) directs 12 statistical agencies not to disclose or use data acquired for exclusively statistical purposes in identifiable form for any other purpose unless otherwise compelled by law. In Europe, the 1995 European Community Directive on Data Protection governs the transfer of individually identified or identifiable data.

The respect and trust of data providers are also essential for statistical agencies to achieve credibility for their data. Protecting the confidentiality of individual responses engenders trust, but, as was noted, agencies may be limited in their abilities to do so. In that case, agencies need to provide respondents with realistic promises of confidentiality that the agency can reasonably be expected to honor.

Respect for data providers also means valuing their time and effort to respond—in particular, designing data collection so that it is not unduly burdensome— and valuing their responses so that they are put to important uses. It is important too that respondents perceive that the agency collecting their information is trustworthy and credible; in particular, that the information will not be used for a particular partisan interest. A statistical agency must be perceived, by its data providers as well as its data users, to be working in the national interest.

4. How Are Government Statistics Collected And Reported?

Statistical agencies collect data through censuses, surveys, and compilations from administrative records. Many countries conduct a census of their population. In the UK and USA, one is conducted every 10 years; in Canada, every five years. Other examples of US censuses are the quinquennial census of agriculture, which covers agricultural production on farms that are defined as places from which $1,000 or more of agricultural products were produced and sold in the census year; the quinquennial economic census, which covers the economic activity of nonfarm establishments; and the quinquennial census of governments, which covers employment and finances of state and local governments.

Examples of surveys in the USA include the Current Population Survey, from which monthly employment and unemployment data are derived; the Consumer Expenditure Survey, from which data on expenditure patterns are used periodically to adjust the weights of goods and services in producing the Consumer Price Index; and the National Health Interview Survey, which provides annual estimates of health conditions and use of healthcare services. Data in government surveys may be collected by field staff of an agency (e.g., the Census Bureau conducts the Survey of Income and Program Participation with its field staff); by state or local governments under a cooperative program (the Current Employment Statistics program of the Bureau of Labor Statistics is a monthly survey, based on payroll records of business establishments, conducted by state employment security agencies); by a private organization under a contract or grant from the agency (the University of Michigan conducts the Health and Retirement Survey under a grant from the National Institute on Aging); or by another federal agency (the Census Bureau conducts the Current Population Survey and the Consumer Expenditure Survey for the Bureau of Labor Statistics and the National Health Interview Survey for the National Center for Health Statistics). Many of these surveys have counterparts in other countries: for example, the Labour Force Survey of the Office of National Statistics in the UK and the Survey on Employment of the National Institute for Statistics and Economic Studies in France.

The units of observation in surveys may be individuals, households, families, schools, establishments, or other economic units. Surveys may be cross-sectional to provide an estimate at a point or interval of time, or they may be longitudinal in design in that they follow the same units of observation over time. Longitudinal surveys can provide information on changes in a unit over time: for example, transitions of families in an out-of-poverty status from the Survey of Income and Program Participation. Some surveys, however, may have a longitudinal component in their design to reduce error or increase efficiency, such as the American Housing Survey.

Statistical agencies also compile information from administrative records. For example, vital statistics are compiled by the National Center for Health Statistics from state agencies, and administrative records are the source of basic statistics for very small firms in the quinquennial economic census, thus reducing or eliminating the need to send these firms census questionnaires.

Statistical agencies may be dependent on the censuses and surveys of other agencies as well as their own data collection in order to produce national indicators. The Bureau of Economic Analysis relies heavily on economic statistics from other sources in compiling the national income and product accounts. And most agencies rely on the Census Bureau to produce the population denominators for reported rates.

Since the US statistical system is so decentralized, important information on a population of interest, such as children or the elderly, is often collected by many agencies and in different ways that may make relating the data difficult, if not impossible. Sometimes, agencies collaborate to l important information gaps. Two examples are key indicators of wellbeing for older Americans (Federal Interagency Forum on Aging-related Statistics 2000) and for children (Federal Interagency Forum on Child and Family Statistics 2000).

Whatever the source of data, one of the most important activities of a statistical agency is measuring, controlling, and reporting upon error. Statistical agencies have been pre-eminent in the development of survey designs to reduce sampling error. And many require, as a standard for some forms of publication, that statistical errors be reported. The agencies too are concerned about measurement and other forms of nonsampling errors, and employ various means to study them, including reinterview studies, record checks, multiple indicators, and cognitive studies, which focus on how errors arise in the process of answering questions. Three statistical agencies, the Census Bureau, the Bureau of Labor Statistics, and the National Center for Health Statistics, have cognitive laboratories to evaluate questionnaires and conduct other research to better understand and reduce error.

5. How Are Government Statistics Organized?

Government statistical activities may be centralized into a single agency or not. The organization in the USA, which is described first, is highly decentralized with activities dispersed among many agencies. Some of the trade-offs between centralized and decentralized systems are described at the end of this section.

5.1 The US Statistical System

The highly decentralized statistical system in the USA arose historically because, as new subject matter areas emerged as a national priority, new statistical agencies were created to meet the needs for new data. Little attention was paid to the relevant data-collection activities of other agencies. Efforts at collaboration have met with some successes, such as the interagency forums on aging-related statistics, and on child and family statistics, but efforts to assure systemic collaboration have failed. Attempts to achieve collaboration through centralization, or at least a major consolidation, have run into stakeholders with a special interest in retaining strong, relatively independent agencies within a particular department of government (see Norwood 1995).

Responsibility for coordinating the myriad statistical activities of the US government rests with a small unit in the Office of Management and Budget, the Statistical Policy Branch in the Office of Information and Regulatory Affairs. Authorization of that responsibility derives from legislation principally concerned with paperwork reduction. Despite its size, however, the Statistical Policy Branch has brought statistical agencies together to address many matters of common concern.

The 1995 reauthorization of the Paperwork Reduction Act created an Interagency Council on Statistical Policy, formalizing an arrangement whereby statistical agency heads participate with the Statistical Policy Branch to coordinate their activities. Ten agencies are recognized as having statistical activities as their principal mission: the Bureau of Economic Analysis and the Bureau of the Census in the Department of Commerce; the Bureau of Justice Statistics in the Department of Justice; the Bureau of Labor Statistics in the Department of Labor; the Bureau of Transportation Statistics in the Department of Transportation; the Economic Research Service and the National Agricultural Statistics Service in the Department of Agriculture; the Energy Information Administration in the Department of Energy; the National Center for Education Statistics in the Department of Education; and the National Center for Health Statistics in the Department of Health and Human Services. The heads of these principal statistical agencies, together with the head of the Office of Management and Budget Statistical Policy Branch and the heads of statistical units in the Environmental Protection Agency, the National Science Foundation, the Internal Revenue Service, and the Social Security Administration, are the members of the Interagency Council.

Through this council, its Federal Committee on Statistical Methodology, and other interagency committees it convenes, and its encouragement of the Committee on National Statistics of the National Academy of Sciences–National Research Council, the Statistical Policy Branch has addressed many methodological and other issues that cut across the interests of most statistical agencies. These include issues of poverty measurement, and the definitions of race and ethnicity and of metropolitan areas. It is also concerned with international comparability of statistics, and it represents the USA at the United Nations Statistical Commission and other international organizations. Moreover, the Statistical Policy Branch has striven to extend by regulation and law the protection of confidential data in ways that would allow statistical agencies to share needed information under that protection.

The Statistical Policy Branch is the only unit of government charged to be concerned with the government statistical system as a whole. It conducts an annual review of all major statistical programs for budget directors in the Office of Management and Budget, and publishes a special analysis of funds proposed for statistical activities in the President’s budget (US Office of Management and Budget 2000).

5.2 Centralized vs. Decentralized Systems

Most large countries of the world have centralized most or all of their statistical activities into a single agency with broad authority for collecting and compiling government statistics. Coordination of the different activities occurs within the agency itself. Examples are Statistics Canada, the statistical agency of Canada, and the Office of National Statistics in the UK.

A central statistical agency is usually seen to be more efficient in resource allocation; in reaping the benefits of scale; in sharing information across different subject matter areas in order to enhance data and to avoid duplicating activities; in coordinating data activities, such as developing a single longitudinal survey of youth that serves the purposes of both health and education policy makers; and in arranging for the effective use of administrative data. Advantages of a decentralized system are that priorities for budget, relevance, and timeliness, and the urgency to meet needs for new data are set within the department with the most concern for the information and with policies that are most affected by it, and that different perspectives and ways of conducting statistical activities can bring strength to the system as a whole.

6. Practices Of Effective Statistical Agencies

The operation of an effective statistical agency begins, first, with a clearly defined and well-accepted mission and, second, with a strong standing of independence. With these prerequisites, a number of practices contribute to the agency’s effectiveness: continual improvement of and development of more useful data, cooperation with data users, openness about its data, commitment to quality and professional standards, wide dissemination of data, advancing the effective use of statistics, analysis and interpretation, an active research program, a highly professional staff, and coordination and cooperation with other statistical agencies.

6.1 Mission

An agency’s mission should include responsibility for assessing and meeting needs for information in its purview, determining the sources of the data and the measurement and collection methods, and ensuring that its data are available to the public. An agency mission statement often goes beyond its charge in legislation. In particular, it should articulate the principles that will govern all of its work. The mission should be well accepted externally by the agency’s department and others who oversee the agency, and especially by data users. It should also be well accepted internally, so that all agency employees understand how their work furthers the mission.

6.2 Independence

A strong and widely acknowledged standing of independence is necessary for a statistical agency to have the credibility to carry out its mission effectively. In particular, the agency must be independent of other parts of its department engaged in policy making, enforcement, regulatory, or administrative activities. It must be impartial and avoid even the appearance that its data might be manipulated for political purposes, or that information on an individual might be used for enforcement, regulatory, or administrative purposes.

Forms of independence differ among statistical agencies. Agency heads may be required by law to be professionally qualified in the collection and analysis of the relevant statistics. They may be appointed by the highest levels of government for a specific term not coincident with the administration. And they may have direct access to the head of the cabinet department in which the agency is located.

Other characteristics of independence include authority for the following: to decide the scope, content, frequency, and manner of data collected, compiled, or reported; to select and promote professional staff; to release statistical information without prior clearance and with a clear separation from policy interpretations by others; to represent the agency’s statistical program before legislatures and the public; and to establish and follow predetermined schedules for the release of important economic and other indicators.

6.3 Continually De Eloping More Useful Data

In order to make their data more useful and to meet changing public policy needs, statistical agencies must continually look to develop data systems that can serve broader purposes. Integrating data from surveys conducted by different agencies or integrating data from surveys with data from administrative records, such as tax or social welfare program records, are ways to achieve this goal. When separate data sets are collected and analyzed in such a way that they may be used together, the value of each may be greatly enhanced. Data sharing and statistical uses of administrative records, with appropriate safeguards for confidentiality, also make a statistical agency more efficient as well as more effective.

6.4 Cooperation With Data Users

An effective statistical agency consults with a broad range of users of its data in order to make its products more useful. It seeks advice on data concepts, methods, and products in a variety of formal and informal ways, from data users as well as subject-matter and other experts. It endeavors to meet the needs for access to data while safeguarding the confidentiality of individual responses. And it strives to provide equal access to all potential users.

6.5 FAIR TREATMENT OF DATA PROVIDERS

To maintain a relationship of mutual respect and trust with data subjects and other data providers, an effective statistical agency observes fair information practices, which include: protecting the confidentiality of data identified with an individual, whether the data are collected directly or obtained from administrative record sources; informing respondents of the purposes of data collection and the expected uses of the information; informing those asked to report whether their participation is voluntary or mandatory, and using appropriate informed consent procedures for voluntary data collections; minimizing the contribution of time and effort asked of respondents, which may be achieved in seeking the advice of respondents in the planning and design of data collection; and respecting the willingness of respondents to contribute to society by representing accurately the information they provide, and by making it available for important uses.

6.6 Openness

An effective statistical agency operates in a completely open manner. Openness requires providing full documentation of the data with acknowledgment of any uncertainty, a description of the methods used and assumptions made, and measures or other descriptions of error, for both sampling and nonsampling error. When a mistake occurs, the agency should acknowledge it promptly and publicly. For a correction to an important indicator, the agency should strive to inform as many current and potential users as possible.

6.7 Quality And Professional Standards

An effective statistical agency is committed to the highest standards of quality work and professional practice and ethics, in statistics and other relevant professions. For this reason, a strong and continuous relationship with appropriate professional organizations in statistics and relevant subject-matter areas is important.

6.8 Dissemination Of Data

An effective statistical agency seeks the widest possible dissemination of its data through a variety of means, including public-use data files and the agency’s website, and has policies to guide what data to archive and how.

6.9 Advancing Effective Use Of Statistics

Statistical agencies should also advance the uses of their data through making them easily accessible to users in convenient forms, involving data users in planning data collection and dissemination, and providing full documentation and other forms of technical support to users.

6.10 Analysis And Interpretation

Agencies can make their data more relevant to public policy concerns by analyzing and interpreting the data. Care must be exercised, however, that analyses not are perceived as policy advocacy.

6.11 Research

An active research program is integral to the activities of an effective statistical agency. Research may be conducted on the substantive issues for which the data were compiled, especially concepts and definitions; methods of measurement, collection, analysis, and dissemination; and topics to improve the utility of the agency’s data in policy analysis and decision making.

6.12 Professional Staff

A statistical agency must recruit and retain a professional staff of high caliber in statistics and fields relevant to its mission. Opportunities should be provided for staff to maintain and advance their capabilities, such as through appropriate activities of professional organizations.

6.13 Coordination And Cooperation

In the US context, an effective statistical agency continually seeks opportunities to cooperate with other statistical agencies to enhance the value of its own and other agencies’ information. Statistical agencies cooperate, not only with each other, but also with state and local statistical agencies in the provision of data for subnational areas. They cooperate, too, with foreign and international statistical agencies to exchange information, on both data and methods, and to develop common classifications and procedures to promote international comparability of information. Integrating data collected or compiled by different statistical agencies invariably requires much coordination and cooperation among the agencies, each of which have different missions. But the rewards are data more relevant to policy concerns, and a stronger statistical system as a whole. For these reasons, statistical agencies must cooperate as partners with one another, not only on the development of data, but also on the entire panoply of statistical activities, including definitions, concepts, measurement methods, analytical tools, professional standards, dissemination modes, ways to advance the effective use of their data, and means to protect the confidentiality of responses.

7. Challenges Faced By Statistical Agencies

We live in a rapidly changing, increasingly interdependent world with a global economy. Many countries face profound demographic changes. Primarily fueled by technological advances, widespread changes are occurring in society, the environment, and international relationships. Governments struggle to keep up with the changes. They are concerned with the need for new property rights, laws governing new markets, new forms of businesses, and investments in new technology for education. They are concerned too about mitigating human error in transportation, communications, healthcare, and other complex, critical systems; avoiding harmful effects on the environment and sources of food; and guarding against the loss of individual privacy. And governments are also concerned with disparities in economic growth and the distribution of income and wealth, to which disparities in access to technology may contribute, since these disparities could increase tensions among social groups and nations.

An efficient, flexible, and modern information system is essential for a government to address these and other concerns. The greatest challenge facing any statistical agency is for it to keep up with the times. ‘Public policy analysis often cannot wait until new surveys are undertaken, but must rely on knowledge drawn from the existing statistical database’ (Cook 2000). Those in charge of government statistics must anticipate and prepare for future needs.

Technology, which has induced so many rapid changes, itself can give statistical agencies the means to respond to those changes. The agencies must not only incorporate new technology into collecting, processing, analyzing, and disseminating government statistics, but they must also prepare for the plethora of new data in digital form.

Advances in information technology are also making it more and more difficult to protect confidentiality. If a statistical agency is to respect its data providers by putting their information to use in the best way to benefit society, stronger ways to protect confidentiality must be sought. Statistical agencies should lead in confidentiality research, including research on data encryption and altering or reconstituting data, and in fostering legal and administrative means to protect confidentiality.

8. Conclusion

Statistics are essential for a government to function effectively and to advance businesses and the wellbeing of its citizens. The nature of data as a public good provides an economic argument for a government to collect statistics, but there are many other reasons for it to do so. One is to assure accurate, timely, and credible information that its citizens may use to hold their government accountable on any issue. A statistical agency must be able to provide information relevant to current and future policy issues, and to establish credibility for itself and its data. Finally, the agency must be respected and trusted by respondents who provide it with data and with all data subjects whose information it obtains. That respect and trust can be fostered by protecting the confidentiality of individually identifiable information and by putting the information it obtains to use in the best ways to benefit society, including research.

More information on the principles and practices of statistical agencies can be found in Martin et al. (2000), and on the organization of US statistical agencies in Norwood (1995). Statistics for over 200 countries and areas of the world are compiled by the United Nations Statistics Division (2000), which provides links for the national statistics of countries online at www.un.org/Depts/unsd through a link to ‘Global Statistics.’ Links to the websites of statistical agencies in the USA are available at www.fedstats.gov.

Bibliography:

1. Cook L 2000 Statement by the first National Statistician on the launch of National Statistics on 7 June, www.statistics.gov./uk/nsbase/about,,ns/About.asp
2. Duncan G T, Jabine T B, De Wolf V A (eds.) 1993 Private Lives and Public Policies: Confidentiality and Accessibility of Government Statistics. National Academy Press, Washington, DC
3. Edmonston B, Schultze C (eds.) 1995 Modernizing the U.S. Census. National Academy Press, Washington, DC
4. Federal Interagency Forum on Aging-related Statistics 2000 Older Americans 2000: Key Indicators of Well-being, 2000. US Government Printing Office, Washington, DC, http://www. agingstats.gov
5. Federal Interagency Forum on Child and Family Statistics 2000 America’s Children: Key National Indicators of Well-being, 2000. US Government Printing Office, Washington, DC
6. Mackie C, Bradburn N (eds.) 2000 Improving Access to and Confidentiality of Research Data: Report of a Workshop. National Academy Press, Washington, DC
7. Martin M E, Straf M L, Citro C F 2001 Principles and Practices for a Federal Statistical Agency: Second Edition. National Academy Press, Washington, DC
8. Norwood J L 1995 Organizing to Count: Change in the Federal Statistical System. The Urban Institute Press, Washington, DC
9. United Nations Economic and Social Council 1994 Fundamental principles of official statistics of the Statistical Commission of the United Nations. Report of the Statistical Commission on its Special Session, New York, 11–15 April 1994. Official Records of the Economic and Social Council, suppl. 9
10. United Nations Statistics Division 2000 Statistical Yearbook Forty-third Issue. United Nations, New York
11. US Office of Management and Budget 1997 Notice of order providing for the confidentiality of statistical information and extending the coverage of energy statistical programs under the federal statistical confidentiality order. Federal Register 62: 35044–50
12. US Office of Management and Budget 2000 Statistical Programs of the United States Government, Fiscal Year 2001. US Government Printing Office, Washington, DC