Implementation of Principles of Official Statistics 2004 PDF

United Nations E/CN.3/2004/21 Economic and Social Council Distr.: General 10 December 2...

United Nations E/CN.3/2004/21 Economic and Social Council Distr.: General 10 December 2003 Original: English Statistical Commission Thirty-fifth session 2-5 March 2004 Item 6 (a) of the provisional agenda* Activities not classified by field: implementation of the fundamental principles of official statistics Implementation of the Fundamental Principles of Official Statistics Report of the Secretary-General Summary The present report was prepared at the request of the Statistical Commission at its thirty-fourth session. a It presents the main results of the survey on the implementation of the Fundamental Principles of Official Statistics conducted by the Division between May and November 2003. Concluding remarks and points for discussion by the Commission are contained in paragraphs 119 and 120. a Official Records of the Economic and Social Council, 2003, Supplement No. 4 (E/2003/24), chap. I.A. * E/CN.3/2004/1. 03-65162 (E) 200104 * 036 516 2* E/CN.3/2004/21 Contents Paragraphs Page I. Introduction.......................................................... 1–3 3 II. Survey results........................................................ 4–118 4 A. Principle 1: Relevance, impartiality and equal access.................... 4–21 4 B. Principle 2: Professional standards and ethics.......................... 22–30 8 C. Principle 3: Accountability and transparency........................... 31–37 10 D. Principle 4: Prevention of misuse.................................... 38–45 12 E. Principle 5: Sources of official statistics............................... 46–53 14 F. Principle 6: Confidentiality......................................... 54–68 17 G. Principle 7: Legislation............................................. 69–85 20 H. Principle 8: National coordination.................................... 86–93 24 I. Principle 9: Use of international standards............................. 94–102 26 J. Principle 10: International cooperation................................ 103–109 28 K. Awareness of the Fundamental Principles.............................. 110–118 29 III. Concluding remarks................................................... 119 30 IV. Points for discussion................................................... 120 31 2  E/CN.3/2004/21 I. Introduction 1. The Fundamental Principles of Official Statistics were developed by the Conference of European Statisticians in the early 1990s, a time at which official statistics in various countries, particularly in Central Europe and the former Soviet Union, went through a period of existential crisis. Political and economic systems were transformed and a number of new nation States emerged. As with many other functions of government, official statistics in those countries had to be re-invented. Public trust in official statistics had to be rebuilt and Governments had to learn to understand the place of official statistics in a changed context. To support these processes, it was deemed useful to develop an international document that would set out the role of official statistics, as well as provide some general guidelines for the functioning of statistical systems. The Conference of European Statisticians adopted the Fundamental Principles of Official Statistics in 1992 and the United Nations Statistical Commission endorsed them in 1994 (after a few minor amendments to the preamble). 2. At its thirty-fourth session, the Statistical Commission requested the Statistics Division to conduct a global review of the implementation of the Fundamental Principles and to submit a report on the issue to its thirty-fifth session in 2004,1 which coincides with the tenth anniversary of the adoption of the Fundamental Principles by the Commission. To comply with this request, the Division developed a questionnaire allowing countries to report their experiences with the Fundamental Principles in a uniform way. Drafts of a questionnaire were sent to international experts and to the regional commissions for comments. 2 The final version of the questionnaire, which was translated into Arabic, French, Russian and Spanish, was sent to 194 national statistical offices3 and posted on the Division’s web site on official statistics.4 By 30 November 2003,4 the Division had received responses from 112 countries. Table 1 below gives an overview of the responses. Table 1 Overview over recipients and respondents Recipients Respondents Response rate for this Percentage Percentage geographic of all of all area Total recipients Total respondents (percentage) Developing countries 147 76 73 65 50 Including least developed countries 49 25 15 13 31 Developed countries 47 24 39 35 83 Total 194 100 112 100 58 Africa 53 27 23 21 43 Americas 36 19 14 13 39 Asia 48 25 36 32 75 Europe 42 22 34 30 81 Oceania 15 8 5 4 33 Total 194 100 112 100 58 3 E/CN.3/2004/21 3. The questionnaire contained 54 questions, structured according to the 10 Fundamental Principles, and a few introductory questions. For principles 1 to 9 the questionnaire started with a general question on the level of implementation of the particular principle in the country. The replies to these general questions are presented in table 2 below. The detailed results for each principle are outlined below. 5 Table 2 Implementation of principles In your country, overall, this principle is: Somewhat Non- Fully implemented Largely implemented implemented Not implemented Response Response Total Percentage Total Percentage Total Percentage Total Percentage Total Total Principle 1 51 44 50 45 10 9 1 1 112 0 Principle 2 66 59 41 37 4 4 1 1 112 0 Principle 3 48 43 56 50 7 6 1 1 112 0 Principle 4 41 37 41 37 21 19 8 7 111 1 Principle 5 54 49 47 42 9 8 1 1 111 1 Principle 6 90 80 21 19 0 0 1 1 112 0 Principle 7 86 77 19 17 4 4 3 3 112 0 Principle 8 34 31 48 44 21 19 7 6 110 2 Principle 9 50 45 55 50 5 5 1 1 111 1 Principle 10 - - - - - - - - - - II. Survey results A. Principle 1: Relevance, impartiality and equal access “Official statistics provide an indispensable element in the information system of a democratic society, serving the government, the economy and the public with data about the economic, demographic, social and environmental situation. To this end, official statistics that meet the test of practical utility are to be compiled and made available on an impartial basis by official statistical agencies to honour citizens’ entitlement to public information.” 1. Introduction 4. Official statistics are one of the cornerstones of good government and public confidence in good government. Official statistics, which are produced by government agencies, can inform debate and decision-making both by governments and by the wider community. Objective, reliable and accessible official statistics give people and organizations, nationally and internationally, confidence in the integrity of government and public decision-making on the economic, social and environmental situation within a country. To meet the test of practical utility, statistics must be relevant and of a quality suitable and in a form that facilitates easy and correct use. The key to achieving this is maintaining an understanding of users’ 4  E/CN.3/2004/21 needs. Statistical agencies use various instruments to interact with users, including advisory bodies and user satisfaction surveys. In addition, good planning is essential in order to respond to the changing needs of users. Compilation and release of data should be free from political interference, so as to ensure impartiality of the national statistical office. In many countries this independence is enshrined in statistics legislation (see also principle 7). Statisticians need to act professionally by the sound application of statistical methods (see principle 2), by openness about concepts, sources and methods used (see principle 3), and by avoiding partisan commentary. Furthermore, making information available on an impartial basis requires dissemination activities, which provide information in the form useful for the users, and release policies, which provide equal opportunity of access. Sound statistical principles need to be followed when presenting statistics so that they are easy to understand and impartially reported. 2. Survey responses Has a “user council” or other advisory body been established? 5. Two thirds of the countries replied in the affirmative, although the councils and advisory bodies come in many different shapes and sizes. The number of members of such councils and other bodies ranges from about 10 to more than 100. Some meet frequently, others only once a year. Apart from sharing one common characteristic, which is “representing the user’s interests”, they have a variety of tasks and responsibilities. These may be grouped under three broad headings: Strategic advice on statistical policy and priorities; Technical advice, in general or on specific statistical programmes and topics; Coordination of statistical activities. 6. Many councils combine several roles. Some have set up subcommittees for technical advice while in other cases technical committees function independently from the “general” council. 7. Examples to illustrating the variety of solutions are the following: Central Statistical Commission (51 members), which assures the link between users and producers of statistics; it meets annually and has 17 advisory subcommittees. In addition, there is a statistical council (15 members), responsible for monitoring the quality and objectivity of statistics, which also gives recommendations concerning the coordination of statistical activities; National Statistics Council, which advises on issues of statistical policy and priorities. Professional advisory committees advise on major programme areas; Advisory Committee for planning and statistical coordination, consisting of members from all ministries and provinces, as well as the private sector, which meets once a month. 8. Several countries explained why they had no statistical council, or had not yet established one. In some cases the reason given was that there was either no statistics law (or that such a law was under consideration) or that the existing statistical legislation did not provide for a council. Several countries mentioned that although there was no formal council there were committees to ensure a user- producer dialogue. Several countries mentioned that they had plans to establish a council. One country mentioned that many ministries and organizations were in the 5 E/CN.3/2004/21 middle of restructuring processes and that it was therefore difficult to set up a council. Is regular feedback of user satisfaction with statistical products and services actively sought? 9. In addition to user councils, feedback can be sought through various other mechanisms. More than two thirds of the countries reported that they use such mechanisms. Some examples are: Periodical surveys of users on the quality of the statistical yearbook and on which indicators should be added; All publications contain addresses, telephone numbers, web site information, etc. and the submission of comments is encouraged; Each programme is required to obtain and analyse user feedback on its outputs; Regular and ad hoc meetings are held with different stakeholders and user groups; Regular and ad hoc user satisfaction surveys are carried out; An annual “open house” event is held; Monthly “meet the user” sessions are held during “a day with clients”; In each publication a questionnaire is enclosed to enable the users to give their feedback on the product. 10. The main reason that countries do not actively seek feedback is for want of resources. Does the national statistical office have a (multi-) annual work programme? 11. As mentioned above, good planning is an essential part of the implementation of principle 1; to that end, the most commonly used planning instruments are annual and multi-annual year programmes. 12. Accordingly, more than 90 per cent of the countries have such programmes, and many have both annual and (rolling) multi-annual programmes. Is a systematic dissemination policy being pursued? 13. Approximately 90 per cent of the countries reported that they had such a dissemination policy. Many offices had set up specialized dissemination units and/or units dealing with the media. Many countries also mentioned the need to diversify the dissemination channels and the increasing importance of the Internet as a dissemination medium, in particular: Rapid growth of electronic dissemination is being pursued; All publications are on the Internet, access to which must be paid for; All figures are in online databases, for which access is free. 14. The lack of resources is mentioned as a main hindrance to developing a systematic dissemination policy. 6  E/CN.3/2004/21 Do you produce catalogues of publications, documents and other services? 15. Almost 90 per cent of the countries reported that they had such catalogues. Many countries had both a hardcopy and an Internet version. Lack of resources and lack of qualified staff were among the reasons mentioned for not having a catalogue. Is the national statistical office free from political interference when preparing the annual and (multi-) annual work plans? 16. Ninety-five per cent of the countries reported that political interference had not been encountered. From the details provided, however, it appears that this question may have been interpreted in different ways. In particular, countries mention the influence of certain “political” users regarding the formulation of work programmes. In addition, the limits imposed on statistical programmes by (lack of) budgetary allocations were mentioned a few times. Some countries also stressed the importance of legal and practical safeguards. Does the national statistical office need political approval to publish statistical information? 17. Eighty per cent of the countries reported they never needed political approval. Twenty per cent reported that they needed approval in specific cases. Examples of cases when political approval was needed are: population census results; the consumer price index; poverty numbers; gross domestic product numbers; and, in one case, “sensitive issues that may have adverse effects on the economy as a whole”. A few countries mentioned that, before publishing, they always submit statistics to the authorities for comments and approval. Does the national statistical office publish an advance release calendar announcing when the various sets of statistics will be published? 18. About one third of the countries did not have an advance release calendar. However, some respondents mentioned that they were considering the idea. Some others reasoned that there were so many uncertainties in the statistical production system that it was impossible to predict when statistics would become available. 19. However, almost two thirds of the countries reported that they published an advance release calendar. Compliance with the dissemination standards of the International Monetary Fund (IMF) was mentioned a few times. The examples given by the respondents revealed that calendars varied in: Coverage: some calendars only cover the main economic indicators, others cover all statistics; Periodicity and time span: some national statistical offices publish such a calendar once a year, others publish updated versions on a weekly basis; Reliability: some national statistical offices always adhere to the pre-release calendar; others have to adjust the set dates. One respondent mentioned that while regular publications were always published according to the schedule, irregular or large-scale publications were subject to rescheduling. Are statistics made available to all users at the same time? 20. About three quarters of the responding countries reported that this was, in principle, the case. The others stated that while this was true for some, it was not true for all statistics. Even in the countries where there was, in principle, equal 7 E/CN.3/2004/21 access for all at the same time, special rules may apply. This is clear from the replies to the following question. If any government departments are given access to statistics prior to their release, is this publicly identified? 21. In about half of the countries where the above practice exists, it is published, although not necessarily in detail. Prior access is given to the relevant ministries to allow them to prepare comments for the official release of the data. The time gap ranges from “one hour before” to “on the afternoon before their release” and documentation of this practice varies from “not publicly identified” to “fully documented in a protocol and release calendar” and “identified on the agency’s web site”. Some detailed examples of pre-release access for Government offices are given below: Given that ministers’ comments on information can have repercussions on financial markets, it has been deemed essential that a few key officials have advance access to a small number of economic statistics on the afternoon before their release the following morning. This is indicated on the agency’s web site; Consumer price index numbers are released to the minister a day prior to public release to allow him time for possible press interviews; Some regular statistical indicators are provided to senior policy officials one hour prior to their public release. This is not a secret, but there is no systematic list of instances when this takes place; Privileged pre-release access is in some cases given to ministers and their advisers; this is, however, fully documented in a protocol on release practices. B. Principle 2: Professional standards and ethics “To retain trust in official statistics, the statistical agencies need to decide according to strictly professional considerations, including scientific principles and professional ethics, on the methods and procedures for the collection, processing, storage and presentation of statistical data.” 1. Introduction 22. This principle further extends the impartiality element of principle 1. Sound statistical methodology, based on the use of frameworks and statistical standards, the correct application of statistical methods and objective presentation of statistics, must be used for the production and presentation of statistics. Such methodology should be chosen by the statistical agency, free from political interference and in accordance with professional ethics. To ensure proper application of methodology, staff need to receive statistical training and research, and innovation needs to be supported. 2. Survey responses How satisfied are you with the number, skills and experience of your staff? 23. Over 80 per cent of the chief statisticians were “satisfied” or “fully satisfied” with the number, skills and experience of their staff, whereas 17 per cent said that 8  E/CN.3/2004/21 they were “not satisfied”. The major problems mentioned by the national statistical offices, whether they were satisfied with their staff or not, were very similar and seemed to differ mainly in magnitude: Insufficient and/or decreasing number of staff: budget cuts have led to an insufficient number of staff in many national statistical offices. Though some offices reported that the number of staff had been increasing, the increase was insufficient to keep up with growing national and international demand for statistics; Difficulties in recruiting and retaining qualified staff: mainly, but not exclusively, developing countries experienced competition from other employers in the private as well as the public sector, where salaries and work conditions are better. This often led to difficulties in hiring, in particular in hiring young and highly qualified staff, as well as to a high turnover of such staff. Bureaucratic recruitment procedures sometimes presented an additional difficulty to the recruitment of staff in some countries. 24. The major shortcomings of skills and experience of staff as mentioned in the questionnaires are: Lack of analytical capabilities: the increased use of new methods of data collection, for example the use of administrative data, as well as the increased demand for the more analytical outputs of the national statistical offices, led to a higher demand for academically trained staff, which, for the reasons stated above, was difficult to recruit and retain; Lack of information technology (IT) capabilities: the increased use of IT in data collection and data dissemination, for example Internet enumeration and dissemination, lead to an increased demand for IT specialists, as well as to an increased demand for IT capabilities among substantive staff; Insufficient language capabilities: since most of the methodological publications as well as international standards and classifications are published in English only, language skills have become an important asset. The lack of sufficient language skills can be a major hindrance for applying international agreed methods and standards (see also principle 9); Insufficient international experience. Is the budget of the national statistical office for training of staff adequate? 25. Closely related to the previous question is the issue of training facilities and the training budget to enable staff to enhance their skills. About one third of the chief statisticians answered that the budget for training was adequate and about two thirds found it inadequate. Staff capabilities and training budget are positively correlated: chief statisticians that are satisfied with their staff also report that they have an adequate training budget. Is the National Statistical Office free from political interference in relation to methodology and survey design? 26. Ninety-eight per cent of the national statistical offices reported that they chose their methodology free from political interference. In most countries, this methodological independence is formally laid down in the statistical legislation that forms the basis of the work of the national statistical offices. In addition to, or 9 E/CN.3/2004/21 instead of, this legal status, a number of arrangements serve alone or in combination as practical safeguards against political interference. These are: Statistical/methodological councils take the formal decisions on the methodology; The use of internationally recommended standards and methods; The involvement of advisory groups, external experts and consultants; Full disclosure of the methodology applied; A tradition and reputation of independence and professionalism; A system of career civil service professionals. 27. In addition, a few countries also mentioned that owing to the very technical nature of statistics and the lack of understanding of the work involved, political interference had not yet been a problem in their work. Do you have guidelines on professional ethics for staff? 28. More than three quarters of the national statistical offices reported that they had written guidelines on professional ethics. However, the understanding of “guidelines on professional ethics” varied and there was a wide spectrum of existing codifications, as follows: The statistics law provides a general framework; Internal regulations and staff rules give more specific guidelines; Ethical codes for civil servants in general; Ethical codes specifically established for statistics provide guidance in ethical behaviour. 29. Existing guidelines, regardless of their form of codification, are usually supported by the following measures: Introduction and training programmes/seminars; Swearing in of new staff and receipt of the relevant laws and guidelines; Memoranda, brochures and the Intranet; Supervision. 30. Of those countries that did not have any ethical guidelines, some planned to develop guidelines in the near future, whereas a few did not deem such guidelines necessary. C. Principle 3: Accountability and transparency “To facilitate a correct interpretation of the data, the statistical agencies are to present information according to scientific standards on the sources, methods and procedures of the statistics.” 10  E/CN.3/2004/21 1. Introduction 31. From the design stage through the dissemination stage of a statistical collection or compilation, there are many ways in which errors can be introduced into the results. Some errors, in particular those resulting from the use of sampling, are random and their magnitude measurable. Other errors, mainly from non- sampling sources, can introduce bias into the results but are difficult to measure. This information on known sources of error as well as concepts, sources and methods used to compile statistics should be readily available to users so that they can judge the fitness of use of the data. 2. Survey responses Do you provide an indication of the quality of data published, for example adequacy of the source data, biases the data may have, response rates, non- response and its treatment, imputation? 32. Almost 90 per cent of the respondents reported that they provided an indication of the quality of data published. Overall, responses indicated that increasing attention was being given to this aspect. Many respondents emphasized that they provided information on the sources, methods and procedures used to compile the statistics, either in the statistical publications themselves or in special publications. In this context, a number of special points were made, as follows: The information is provided in a methodological note in the publication or web site or as part of a separate methodological publication or web site; This kind of information is especially provided when any new indicator is produced, or when large-scale changes in methodology occur for a survey; The practice might vary in comprehensiveness from one statistical field/ publication to another; Some information is published and additional information is available on request; National offices meet with analysts from ministries and the central bank to provide technical explanations for analysts directly; Special “quality” reports are issued from time to time; Standard quality indicators are developed in some countries. 33. On the other hand, about 10 per cent of the respondents indicated that they did not provide any indications of the quality of the data. Reasons given were: Lack of staff; Publishing this information was not thought to be necessary. Does the national statistical office routinely provide analytical/explanatory text with published statistics? 34. Ninety-five per cent of the respondents indicated that they routinely provided explanatory or analytical text with published statistics. The extent to which such text is included varied between countries and among publications within a country. Responses frequently emphasized that the objective was to ensure a good understanding of the data by users. In some cases, explanatory, but not analytical text was provided. Many respondents seem to have understood this question to relate 11 E/CN.3/2004/21 to the issue of providing technical, methodological descriptions and replied in the affirmative. This may mean that the publication of analytical explanatory text may not be as pervasive as suggested by the 95 per cent affirmative answer. 35. A number of special points were made as follows: The need to reach users using simple language; Text was included as part of the publications, or in some cases also in press releases; It was necessary for the national statistical office to remain impartial and “simply publish the main facts”; There was a requirement that the text be apolitical and objective; Graphs and charts are also published with the data; Peer review is used for new and sensitive publications; Analytical text will be included in an expanded number of publications. 36. Reasons given for not including analytical text were that: There is an insufficient number of staff; The staff in a national statistical office was not qualified for the task; The responsibility lies outside the national statistical office. 37. In some countries, the debate as to whether the statistical authority should or should not publish analytical text continues. D. Principle 4: Prevention of misuse “The statistical agencies are entitled to comment on erroneous interpretation and misuse of statistics.” 1. Introduction 38. While statistics can be acceptably used and interpreted in many different ways, it is important to maintain trust in, and the credibility of, official statistics. Hence, statistical agencies should draw attention to obvious public incorrect use or interpretation. In addition, other measures to minimize misuse include the publication of documentation explaining key statistics and education programmes for users to increase awareness and knowledge of official statistics. 2. Survey responses Is the national statistical office entitled to comment on erroneous interpretation and misuse of statistics? 39. According to this survey, more than 90 per cent of the national statistical offices are entitled to comment on erroneous interpretation and misuse of statistics. The extent to which national statistical office react to misuse/misinterpretation of data ranges from “systematically inspect” and “always correct” to “only rarely”. Indeed, many countries apply a modest policy, for example commenting only on 12  E/CN.3/2004/21 uses “when they are considered dangerous or particularly misleading” or in “cases of obvious misuse”. Practical reasons given for this are: The lack of resources; The fact that most instances of misuse/misinterpretation are discovered and commented on by others; A general hesitation regarding the danger of engaging in any political debate as such. 40. Most misinterpretations are reported to occur in the mass media and usually the chief statistician or the substantive unit of the national statistical offices responds through letters to the editor, which in many countries the concerned press is obliged by law to print, as well as press conferences and press releases. 41. National statistical offices also reported that they directly contact other government agencies (for example the central bank), international organizations and universities when they are the sources of misinterpretation of data. While most national statistical offices reported that they comment solely on misinterpretation and misuse of their own data, some offices reported that they provide comments on data published by private, other governmental or international organizations, though mainly as a matter of clarifying why different data producers publish different figures. One prominent example given is poverty data. Additional examples, where national statistical offices commented on misinterpretation or misuse of data are: inflation rate/consumer price index (introduction of the euro as the single European currency), national accounts/GDP, unemployment data, income, foreign trade, population estimates and life expectancy. Seven national statistical offices, all but one in developing countries, reported that they were not entitled to comment on misinterpretation of their data. Does the national statistical office carry out activities to educate users, including the media? 42. Activities to educate users are conducted to “improve the statistical literacy” of key users, such as the mass media and to prevent misinterpretation. According to the survey, about 80 per cent of the national statistical offices carried out activities to educate users. Some did so regularly, giving as many as 30 courses a year, others provided such activities more intensively before large censuses and surveys or when a major change of methodology had been adopted. On the other hand, some countries reported that they conducted such activities irregularly. Most countries, however, seemed to see user education as part of a broad public relations strategy, “to deepen the general public’s understanding of the importance of statistics”. From that perspective, many different groups have been identified and are targeted by national statistical offices, inter alia: Government, namely staff of ministries, “assistants to lawmakers”; Mass media: print media, radio and television; Businesses; Non-governmental organizations (NGOs); Trade unions; Academia (professors and students) and high school students; 13 E/CN.3/2004/21 Users of anonymized census microdata; The general public. 43. As examples, the following activities are reported by various national statistical offices: (Re-)designing publications to make them user-friendly; Publications and booklets focusing on user groups, for example “Statistics and the mass media”; Seminars, partly focusing on certain user groups (media, microdata users); User-friendly homepages and courses on how to find data on the web site (one country also reports to have a homepage especially for children); Press conferences and press releases, which have contact names and numbers for help with interpretation; Participation in annual conferences of user groups, book fairs and other suitable events; Open house events; Visitors service/special unit for user education; Awareness campaigns, such as a “National Statistics Day/Week/Month”; Events such as the “National competition of graphic statistics” or the “Statistics quiz”, which two national statistical offices describe as useful tools. 44. It is clear from the above list that educating users and seeking feedback of user satisfaction (covered under principle 1) are clearly linked and involve similar activities. 45. Those national statistical offices that do not pursue user education actively almost uniformly give the lack of financial and human resources as the main reason. Such problems are in some cases reported to be accompanied by a “weak statistical culture in the country”. Most of the national statistical offices that report they lack resources for user education purposes are in developing countries. E. Principle 5: Sources of official statistics “Data for statistical purposes may be drawn from all types of sources, be they statistical surveys or administrative records. Statistical agencies are to choose the source with regard to quality, timeliness, costs and the burden on respondents.” 1. Introduction 46. Statistical offices should be cost-effective, making the best choice of concepts, sources (including administrative records) and methods by balancing quality, timeliness, costs and the reporting load of respondents. Agencies should, therefore, have policies to minimize the reporting burden and should implement quality management programmes to achieve the quality and timeliness required by uses of their statistics. 14  E/CN.3/2004/21 2. Survey responses Does the national statistical office have access to administrative data? 47. To some extent, this appears to be the case in almost all countries, although there are many variations in the scope and conditions of such access. In several countries, the right and conditions of access to administrative data are described in the statistics law. Some countries note that ministries, not the national statistical office, compile statistics based on administrative data. The following examples illustrate some policies and issues related to the use of administrative data for statistical purposes: According to the law, data from all administrative registers must be submitted to the national statistical office; For economic statistics, access to administrative data is ensured. For some social statistics, the access to data of several ministries is not ensured, due to legal restrictions; The statistics law states: “Collection of data for statistical purposes shall primarily rely on data collected in other contexts”; There is a provision on the use of administrative data in the statistical law, but there is a problem actually obtaining data from the tax authorities; administrative data are not very well-developed; There is access, but not an automatic flow of data; There is no systematic access, but there is access on the basis of mutual agreement when administrative data meet a statistical need. Does the national statistical office systematically work on improving timeliness? 48. More than 90 per cent of the respondents reported that this is the case. Measures applied for improving timeliness are: Frequent meetings with data providers, especially other government departments, and setting deadlines for submitting data; Releasing preliminary data based on partial returns and not fully validated data; Improving overall procedures and the use of new technologies, including web- based data collection, which enhances timeliness. 49. Many examples of successes are mentioned, including the following: Timeliness of monthly retail statistics has been improved by 4 weeks over the last 10 years; Using computer assisted personal interviewing in some household surveys has reduced data-processing time; Release of data from national, large-scale surveys has been reduced from 12 to 14 months to less than 6 months after completion of the survey. 15 E/CN.3/2004/21 Does the national statistical office systematically work on reducing the reporting burden on respondents? 50. Eighty-six per cent of the countries reported that they worked systematically on reducing the reporting burden on respondents. One country mentioned that the Statistics Act stated that: “The data shall be collected in a manner that is economical and causes the respondents a minimum of inconvenience and costs”. Specific examples of policies and practices are: Simplification of questionnaires, including shorter, less detailed questionnaires; Integrated surveys; Rotating sampling for annual surveys; Use of administrative data. 51. Successful examples are given as follows: A range of initiatives, including use of administrative data, has resulted in a 40 per cent reduction in costs imposed on small business over the last seven years; The use of tax data is expected to reduce the number of businesses surveyed for the monthly business frame update survey by approximately 2,000 per month; a 40 per cent reduction. Does the national statistical office have a quality management programme for its statistical outputs? 52. Sixty-three per cent of the countries replied in the affirmative. However, from the specifics provided to illustrate what is being done, it is clear that the label “quality management programme” covers many different approaches designed to enhance statistical quality, from relatively simple procedures and ad hoc measures to more sophisticated, wide-ranging multi-year strategies. Some of the details provided were: All key national statistics outputs are subject to a rolling five-year programme of quality reviews involving external expertise; There is an editorial committee that checks and edits all publications prior to their release; A new central methodological unit has been set up to development statistical guidelines and recommended practices; ISO 9002 certification has been awarded for the price statistics programme; three more outputs are in the process of application for certification; Quality management is based on generic systems such as the European Foundation for Quality Management (EFQM) and Total Quality Management (TQM). 53. Many of the countries that reported that they did not have a quality management programme reported that they did, however, have various systems in place to ensure quality. Some countries mentioned that they considered developing a formal programme in the near future and a few mentioned the importance of staff training in this context. 16  E/CN.3/2004/21 F. Principle 6: Confidentiality “Individual data collected by statistical agencies for statistical compilation, whether they refer to natural or legal persons, are to be strictly confidential and used exclusively for statistical purposes.” 1. Introduction 54. Reliable official statistics depend on public cooperation and goodwill to provide accurate and timely information requested in surveys. Such cooperation and goodwill is maintained by protecting the confidentiality of information provided by respondents. Key aspects of confidentiality protection are maintaining information securely, avoiding disclosure of identifiable information and providing access to anonymized microdata for non-statistical purposes, such as statistical research. 55. Under this principle, it is important to make a clear distinction between: (a) Disclosure of identifiable individual data; (b) Providing access to microdata from official statistics: making available non-identifiable individual responses, that is, anonymized individual (non- aggregated) information. 2. Survey responses How well developed are practices to prevent disclosure of individual data? 56. According to this survey, confidentiality protection is part of the culture in statistical agencies in almost all countries and as such is seen not only as a legal obligation, but as a “valuable means to maintain high-quality statistical data”. Hence, many countries continuously seek to improve the protection of privacy and confidentiality through improving laws, working procedures and by making use of technological developments. 57. Despite this, 77 per cent of the countries responded that practices to prevent disclosure of individual data were highly developed (no individual data is ever disclosed), and 21 per cent responded that practices were developed (usually individual data is not disclosed, but there have been exceptions). Two per cent of the respondents assessed such practices as being undeveloped in their countries. 58. The practices in place to guarantee that individual data will not be disclosed vary. The main mentioned mechanisms were: During data collection and data processing: No entry of individual names of persons or enterprises in databases; Procedures in place to control transfer of identified data files even within the statistical office; For publication of aggregated data: Suppression of information if the number of respondents allows for easy disclosure; Standard software (such as Argus) applied for checking tabulations and microdata against disclosure, also other special software; 17 E/CN.3/2004/21 Review (by authorized staff) of all data prepared for dissemination for possible indirect disclosure; When releasing individual data: For identifiable individual data: ask explicit permission from the individual or company concerned; Examination of all applications for access to confidential data by a statistical disclosure committee at the national statistical office and in some countries to the data protection authority; Release of individual data only as anonymized microdata for research purposes; General security measures within the office: Restrictive access to the working place and to databases, with access to individual computers and terminals secured by both identification and password; Ethics of the profession and/or internal regulations, including an “oath of office”. 59. Some countries reported occasions where individual data have been disclosed or were used for non-statistical purposes. These incidents can be classified as: (a) Release of individual data mandated by law, regulations or other authorization; (b) Release of individual data in violation of the law and other regulations. 60. While most countries reported either that they never released identifiable individual data or that such releases were rare exceptions, circumstances and procedures for the authorized release were given as follows: Whenever the individual who provided data consents to the use for other purposes; Individual data are used as evidence in a court of law; Government entities ranging from federal institutions such as tax collection bodies to state representatives, courts and local authorities have the legal right of access to identifiable individual data; Data that permits identification of respondents may be transmitted without the consent of the respondents for the purposes of scientific research, pursuant to procedures established by the government; Individual data can be disclosed in emergency situations such as a public health crisis; Individual data can be disclosed, in some cases, on the agreement of the High Statistical Council. However, it is not possible to disclose individual data related to families or individuals. 61. Examples of the authorized release of identifiable individual data were given as follows: 18  E/CN.3/2004/21 Population census data was exploited for establishing a voters’ register. The lifting of confidentiality was dealt with by a presidential decree which authorized it; “The one known occasion was that when a serious public health problem had arisen all over Europe. This had to do with the import of beef meat and animals foodstuffs from a certain country”. 62. On the other hand, one country reported that, by law, more than 20 organs of the State had the right to demand and receive identifiable information. 63. Examples of the unauthorized release of identifiable individual data were given as follows: In the past there have been cases of data being exchanged for cash to supplement low salaries; There have been isolated cases where data was stolen. 64. Additional points made by respondents were: Rules and procedures are generally much stricter regarding securing the confidentiality of individuals and households, rather than enterprises; Many respondents also mentioned that violation of confidentiality provisions may lead to penalties, ranging from fines, termination of employment/contract to imprisonment. What is your practice in granting access to microdata from official statistics for statistical purposes (for example, to researchers)? 65. While this matter is under discussion in many countries, many others reported that they provided anonymized microdata files for non-statistical purposes, typically for research. The decision to do so is taken by different bodies. Examples provided by the respondents are: The provision of microdata is provided for in the law; A committee, national council or internal advisory board is to decide on data disclosure requests; In a few cases, the authority lies with the director general of statistics; In one case, such decisions are taken by the minister responsible for statistics. 66. Once granted, access to microdata is provided as: A public use file available to everybody; Microdata under contract, where data is made available solely to the applying user. 67. The main means of making the data available are identified as: Access via the Internet (usually for public use files); Sending the anonymized data file by e-mail or other electronic means; On remote access facilities and controlled remote data processing; 19 E/CN.3/2004/21 On an onsite strictly controlled access facility (usually given to specific researchers after strict scrutinizing, under written agreements and under heavy fines if disclosure occurs). 68. Other provisions and arrangements used when granting access to microdata are: For statistical offices with more than one type of access available, the choice of which kind of access is to be granted for a certain case is made by an advisory board, based on the importance of the research project; Use of a researcher’s licensing or registering procedure to grant access to confidential data; In order to be allowed access, researchers must have an approved project, obtain a security clearance and sign a sworn agreement to preserve the confidentiality of the data; Decisions are taken on a case-by-case basis; Researchers are required, or encouraged, to share the results with the national statistical office. G. Principle 7: Legislation “The laws, regulations and measures under which the statistical systems operate are to be made public.” 1. Introduction 69. Openness in the production of official statistics is important for maintaining trust and credibility in both statistical agencies and the data they produce. In many countries, the production of statistics is governed by statistics legislation that sets out the authority and powers of a statistical agency, including the position in the national administration as well as its obligations, such as publishing the results of collections and protecting the confidentiality of information collected from respondents (see also principle 6). 2. Survey response Do you have a general statistics law? 70. Many countries stressed the importance of a legal basis for official statistics and the constant need to modernize it. Many countries reported that their statistics law is currently under revision or expressed the need for such a revision. In addition, many respondents stressed the importance of making such legal documents available to the public. 71. More than 90 per cent of countries responding had a general statistics law providing the authority and rules under which the national statistical office operated. 72. General statistics laws regulate, inter alia, the following issues: The organization of the national statistical system as a whole; 20  E/CN.3/2004/21 The functions, rights and responsibilities of the statistical entities within a decentralized system, in particular of the national statistical office. 73. In greater detail, statistics laws define or regulate: The coverage of statistical activities; The independence of the national statistical office and the position of the chief statistician; The relationship between the national statistical office and respondents, in particular their obligation to provide information and at the same time confidentiality protection on the part of the national statistical office; as well as penalties for non-compliance in both cases; Access to administrative records; Release practices and principles of dissemination of data. 74. In some cases, the law also mentions/covers: “The right of citizens to obtain official statistics”; The provision of anonymized microdata for research purposes; The establishment and functions of an advisory council; International statistical cooperation; The obligation to minimize the reporting burden, for example through coordination mechanisms within the statistical system; Coordination within the country’s statistical system. 75. Some countries stated that their statistics law was very old and too general, and expressed the need for modernization. 76. Of the countries that do not have a general statistics law, the majority of the statistical systems involved are regulated by government decrees, orders and regulations or by a “range of separate legislative provisions”, as it is the case in two developed countries. Generally, are respondents obliged by law to respond to statistical enquiries from the national statistical office? 77. In many countries, the law regulating official statistics promulgates an obligation for physical and legal persons to provide the national statistical offices with “complete, accurate, reliable and timely statistical information” free of charge. In fact, a total of 70 national statistical offices (64 per cent) reported that their respondents must “always” respond to official surveys. Thirty-two national statistical offices (29 per cent) reported that “in many cases” respondents were obliged to respond to statistical inquiries and seven offices (6 per cent) reported that this was “sometimes” the case. One country also reported that this was “never” the case. 78. In most of the countries, the general statistics law provides for both mandatory and non-mandatory participation of respondents and this is either: Specified in the general statistical law or in other legal provisions ordering certain statistics issued by the legislature or the Government; 21 E/CN.3/2004/21 Left to the national statistical office to decide on the status of an inquiry on a case-by-case basis. 79. Distinctions between mandatory and non-mandatory participation in statistical inquiries are being made along various lines: Some countries have defined a “national statistics programme” and participation in statistical inquiries that are part of this core set is mandatory, while participation in any additional survey is voluntary; Other countries distinguish between enterprise and household surveys; participation in the first usually being mandatory and participation in the latter being voluntary. 80. Participation in the country’s population and housing census seems to be mandatory in almost every country, although some national statistical offices specified that participation in any inquiry that involved either the private life of a person (for example personal health and political opinions) or another organization or institution was voluntary. 81. Despite the existing obligation in many countries and the sanctions that are usually (but not always) provided for by the law, many national statistical offices “mostly obtain information through ‘requests’” and “enforce the relevant legal provisions in [their] Act … as a last resort”. However, some offices claimed that they pursue non-respondence or misinformation rigorously, including through police involvement. In a statistical survey, are respondents informed about the nature of the survey and their rights? 82. Ninety-eight per cent of the national statistical offices reported that they informed respondents about the nature of the survey and about their rights. This was done in various ways, following both tradition and regulations: Awareness campaigns often precede census and large surveys. These usually make use of all media, in particular radio, television and newspapers as well as the distribution of leaflets; In face-to-face and telephone interviews, interviewers introduce themselves and explain the nature of the survey as well as the respondent’s obligations and rights; In almost all countries, when questionnaires are being mailed to respondents they contain an explanatory text about the survey and the respondent’s obligations and rights on the front page, usually accompanied by a “letter to respondents”; In addition, advance letters announcing the visit or call by the interviewer or the sending of a questionnaire are often sent by mail, in particular when a respondent participates in a survey for the first time; Many countries also publish information on the rights and obligations of respondents on their web sites; Respondents are given the phone number of the person in charge of a particular survey and can call to ask questions and give comments related to it. 22  E/CN.3/2004/21 83. The information given through any of the above-mentioned means is reported to be generally related to the nature of the survey as well as to the rights and obligations of the respondents and can be specified as follows: Explaining the goals, purposes, methods and other main characteristics of the survey; Including a description of the manner/method in which the respondent got selected as a participant in the survey; Giving information on whether participation in the survey is compulsory or voluntary; Guaranteeing the confidentiality of the information given by the respondent; Explaining the consequences of non-response or providing false information; If applicable, introducing the organization/institution on behalf of which the survey is conducted. To whom/what office within the Government does the head of the national statistical office report? 84. The position of the national statistical office with the Government is an important part of the legal environment in which it operates. That position and the terms of reporting may vary a great deal. Most prominently, national statistical offices report to the following entities: (Deputy/Vice) Prime Minister or the President; Cabinet or Council of Ministers; Ministry in charge of planning and/or development; Ministry in charge of economic affairs, industry and/or trade; Ministry in charge of finance; Ministry of Statistics; Ministry of the Interior; Ministry in charge of information and communication; Parliament directly; Planning and/or development commission/council/authority; Statistics Council; Various ministries depending on the subject area. 85. The above arrangements vary not only in terms of the position of the national statistical office within the overall government structure, but also in terms of how strong the relationship with the “parent body” is. While some national statistical offices have a high degree of administrative independence others are actually part of a ministry. In addition, the person to whom the national statistical office reports to within a ministry or other supervisory body varies considerably to include the minister himself/herself, a permanent secretary in the ministry, a general director, the director of a department or advisers. 23 E/CN.3/2004/21 H. Principle 8: National coordination “Coordination among statistical agencies within countries is essential to achieve consistency and efficiency in the statistical system.” 1. Introduction 86. Official statistics are broad in scope and are often produced by many different government agencies in a country. Usually, there is a central or national statistical office that produces the greater share of official statistics, while sometimes there is more than one statistical agency handling different areas of statistics. In all cases, the majority of official statistics are produced by government departments as a by- product of their activities, and sometimes by separate statistical units within the departments. 87. No matter what the organizational arrangements are for producing national statistics, coordination of statistical activities should be undertaken to avoid duplication of work, to minimize the reporting burden of respondents and to facilitate the integration of data from different sources through the use of statistical standards and participation in international initiatives. 2. Survey responses Are there any other producers of official statistics other than the national statistical office in your country? 88. More than 90 per cent of the countries reported that in addition to the national statistical office there are other producers of official statistics in the country, although the national statistical offices are the dominant producers of official statistics in most countries. Many countries provided examples of the other producers of official statistics that produce and publish statistics in their own spheres of concern. The following examples were given by the respondents: Line ministries; Subnational offices, which sometimes produce statistics for their regions. 89. Countries reported various relationships between the national statistical office and the other producers of official statistics, including: Other entities that can collect statistics only subject to permission from the national statistical office; Each ministry produces data within its field of work and submits it to the national statistical office; Other federal institutions may also act as organs of the national statistical office; Other institutions are “marginal producers” compared to the national statistical office; the work of these institutions is well integrated with programmes of the national statistical office; Other government entities produce statistics based on questionnaires and methodology supplied by the national statistical office; 24  E/CN.3/2004/21 In one case, a central office carries out coordination of other statistical producers but does not produce statistics itself; In another case, there is no national statistical office, but the statistical system is decentralized among government offices. However, the statisticians working in those government offices constitute the government statistics service. Are there organizational arrangements to coordinate data collection for statistics at the national level? 90. Eighty-six per cent of the respondents indicated that organizational arrangements are in place to coordinate data collection and avoid duplication of statistics at the national level. Coordination is implemented in different ways: National statistical legislation regulates coordination; Coordination is included in the annual or multi-annual plan for data collection. 91. In many cases, the national statistical offices play a major role in coordinating data collection, as the examples below indicate: Joint data collection by the national statistical office and other agencies, including provincial agencies, avoids duplication; The national statistical office approves all questionnaires; The national statistical office has “clearing house” responsibilities for any planned data collection by other agencies; Memorandums of understanding or agreements are made between the national statistical office and other government agencies; Contact is maintained with other organizations at both high and operating levels to maintain coordination. Are there organizational arrangements for setting statistical standards (terminology, definitions, classifications, geographical classifications, methods, sampling frames, etc.) at the national level? 92. Almost 80 per cent of the countries indicated that they had organizational arrangements for setting statistical standards (terminology, definitions, classifications, geographical classifications, methods, sampling frames, etc.) at the national level. Arrangements were specified as follows: Statistical legislation establishes and specifies common standards, or specifies the body that has the responsibility; A central body, either the national statistical office or another body cooperating with the national statistical office, is assigned this responsibility by law; National statistical boards/councils/committees carry out the coordination role; Joint committees of the national statistical office and other agencies are established for surveys in specific subject matter fields; There is no formal organizational arrangement; however, the national statistical office promotes common classifications, uniform sampling frames and other standard elements; Standards are set through consultations with producers and users. 25 E/CN.3/2004/21 93. In addition, respondents mentioned that: International recommendations are taken into account; Special efforts are made to carry out new surveys. I. Principle 9: Use of international standards “The use by statistical agencies in each country of international concepts, classifications and methods promotes the consistency and efficiency of statistical systems at all official levels.” 1. Introduction 94. In order to facilitate international comparisons of statistics, as well as to achieve efficiency and quality in their production, international statistical standards (i.e. frameworks, concepts, and classifications) should be used as far as possible in the production of official statistics. However, compromises are usually required in the application of international standards to suit the conditions and requirements of users within each country. 2. Survey responses 95. Ninety-five per cent of the national statistical offices reported that this principle was fully or largely implemented in their countries. This is also reflected in the answers to the more detailed questions as shown in table 3 below. Table 3 Application of international standards6 Do you apply international standards? Yes, as they were Yes, adapted to Both are No, Statistical field recommended national circumstances applicable not at all National accounting/other economic statistics 42 62 5 - Censuses and demographic statistics 52 51 6 - Social statistics 38 69 4 - Do you apply international standards in national accounting/other economic statistics? 96. The application of international standards in national accounting and economic statistics in general includes general frameworks such as the System of National Accounts, as well as associated standards and classifications that also have other applications in the statistical system and elsewhere. 97. According to the present survey, of the large frameworks, the System of National Accounts (1993) is the most widely accepted, with the majority of countries reporting that they had adopted it. For European countries, the adoption of the European System of Accounts (1995) is mandated by European regulation, but since it is consistent with the System of National Accounts, adoption of the international standard is almost universal in Europe. As an exception, a number of 26  E/CN.3/2004/21 developing countries still work with the previous version of the System of National Accounts (1968), indicating their desire to follow international standards. Countries also reported that efforts to convert to the 1993 System are being made. 98. The underlying classification standards often have additional statistical and non-statistical applications, which add considerable national requirements to their formulation and implementation. Nonetheless, most countries use activity classifications, product classifications, classifications of expenditure, occupational classifications, health classifications and others that are either identical to or based on the most important international standard classifications. For some of these classifications, notably activity and product classifications, regional classifications have emerged that serve the more detailed needs of specific groups of countries. However, in almost all cases, these regional classifications are based on the international standard. For classifications that are not directly derived from international standard classifications, the ability to convert data to the international standard classifications is still a highly desired property, underlining their important central role. Do you apply international standards in censuses and demographic statistics? 99. Since its early years, the United Nations has issued a series of international recommendations on population and housing censuses to assist countries in planning and carrying out improved and cost-effective censuses, under the title Principles and Recommendations for Population and Housing Censuses, last published in 1998. 100. For the current census round, many countries reported that they followed the United Nations recommendations or related recommendations issued by the European Union or the Latin-American and Caribbean Demographic Centre. Furthermore, most countries used related international classifications, such as activity classifications, occupational classifications and health classifications. 101. About half of the countries reported that they departed from international standards as particular circumstances within the country required. In this context, several countries mentioned that the registers they used to compile demographic data, which were established for non-statistical purposes, do not necessarily use definitions that are in accordance with international standards, for example, the adaptation of the definition for “refugee” to national circumstances, as reported by one country. The importance of documenting and publishing such adaptations was also mentioned. Do you apply international standards in social statistics? 102. As in other areas, countries reported that they applied international standards where such standards existed, adapting them to national circumstances as necessary. Respondents also reported that they followed international recommendations, standards and classifications in the areas of education statistics (United Nations Educational, Scientific and Cultural Organization (UNESCO), labour statistics (International Labour Organization (ILO)), health statistics (World Health Organization (WHO)) and poverty statistics (World Bank). The use of registers was mentioned as one reason for adapting existing standards. In addition, some respondents mentioned the importance of regional standards (European Union, Latin America) and others pointed out the lack of international standards in areas such as cultural statistics. 27 E/CN.3/2004/21 J. Principle 10: International cooperation “Bilateral and multilateral cooperation in statistics contributes to the improvement of systems of official statistics in all countries.” 1. Introduction 103. The sharing of information and practices, as well as cooperation in the joint development of statistical standards, international statistical activities and so forth is an essential ingredient for the continuous improvement of the quality and range of official statistics in all countries and for the efficiency of their production. In general, such cooperation is facilitated by various international activities supported by international statistical and other agencies as well as by professional associations. 104. Technical cooperation, either organized bilaterally among agencies or organized through international organizations and activities, is important for the development of the range and quality of official statistics in the developing countries. Programmes and institutions are available for statisticians to receive statistical training relevant to the production of official statistics. 2. Survey responses Has the national statistical office been engaged in any international cooperation projects in the last five years? 105. Ninety-six per cent of respondents confirmed that their offices have been involved in international cooperation projects in the last five years; 55 per cent indicated that they had been involved as recipients only; 11 per cent indicated that they had been involved as donors only; and 29 per cent indicated that they had been involved both as donors and recipients. Table 4 Involvement in international technical cooperation project in the last five years As donor Yes No Total Yes 33 (29%) 62 (55%) 95 (85%) As recipient No 12 (11%) 5 (4%) 17 (15%) Total 45 (40%) 67 (60%) 112 (100%) 106. In the majority of countries that have been involved in international technical cooperation in the last five years, there is a special unit, or in some cases a high- level committee, within the national statistical office responsible for organizing cooperation activities. Some donor countries mentioned that funding for technical cooperation activities in statistics was normally provided by government departments other than the national statistical office. 107. Four per cent of respondents indicated that they had not been involved in international technical cooperation projects over the last five years. In two cases, this was expected to be a temporary situation. 28  E/CN.3/2004/21 In your country, has international cooperation in statistics contributed to improving your system of official statistics? 108. All but one country that had been involved in technical cooperation indicated that international cooperation had contributed to improving their statistical systems. 109. Additional comments made included the following: International cooperation should be an integral part of the work of national statistical offices; Bilateral consultations, expert group meetings and seminars provide benefits to both developed and developing countries; Not only do recipient countries benefit from international cooperation, but experts from the donor countries get new insights too; Many modern technical cooperation projects do not fit in the donor-recipient scheme; Better coordination of international efforts is needed. K. Awareness of the Fundamental Principles 110. Application of the Fundamental Principles necessitates knowledge of them, and the key persons who should be aware of their existence are the chief statisticians themselves. Have you previously been informed about the Fundamental Principles? 111. Almost 90 per cent of the chief statisticians reported that they were aware of the Fundamental Principles, most of them, about two thirds, having learned about them through the report of the Statistical Commission of 1994 in which the Fundamental Principles were adopted. Other sources of information were presentations at meetings and conferences (50 per cent), the web sites of the Statistics Division and the regional commissions (36 per cent) as well as various other sources (19 per cent). In the latter category were: the United Nations Handbook of Statistical Organization, the publications and meetings of the Statistical Observatory of Sub-Saharan Africa (AFRISTAT), other statistical publications, participation by national offices in the working group that drafted the Fundamental Principles and work with the General Data Dissemination System, in which the Fundamental Principles are integrated. Are the directors of statistics of other entities that produce official statistics in your country aware of the Fundamental Principles of Official Statistics? 112. As many statistical systems are decentralized, it is necessary to ensure that the directors of statistics of other entities that produce official statistics in a country are aware of the Fundamental Principles. 113. Just over half of the directors of national statistical offices said that the directors of statistics of other entities were aware of the Fundamental Principles. However, about one third said they did not know whether the directors of other entities were aware of the Fundamental Principles. On the positive side, some respondents indicated that they were taking steps to disseminate the Fundamental Principles in their countries. 29 E/CN.3/2004/21 114. Directors of other entities had been made aware of the Fundamental Principles in a number of ways: The Fundamental Principles themselves, or information about them, had been distributed to directors; References had been made to the Fundamental Principles in the statistics law or its equivalent or they had even been included in the law; the Fundamental Principles have been referred to in the annual federal statistics programmes; Discussion of the Fundamental Principles had been conducted in national conferences, workshops and seminars; Articles on the Fundamental Principles had been published in national statistical journals; The Fundamental Principles have been translated into the national language; During discussions with the other directors, the Fundamental Principles have been referred to. 115. About 10 per cent of the respondents said that the directors of other entities were not aware of the Fundamental Principles, which might be due to the fact that many institutions produce statistics “as a sideline of their main functions”, as suggested by one respondent. Is the person/office in the government to whom the national statistical office reports aware of the Fundamental Principles? 116. Two thirds of the respondents reported that their superiors knew about the Fundamental Principles, whereas one third reported that they did not. 117. The most important ways of keeping the superior office/official informed are: Most of the national statistical offices explicitly informed the appropriate individuals and offices about the Fundamental Principles; In many countries, the Fundamental Principles are either an integral part of the general statistics law or the law refers to the Fundamental Principles; The Fundamental Principles are being referred to in the law, in published and unpublished reports, in strategy and policy papers and in publications and meetings; Two national statistical offices reported that the officers they reported to were actually statisticians and were well aware of the Fundamental Principles. 118. Some chief statisticians mentioned that while their superiors were not aware of the existence of the Fundamental Principles as a document, they were, however, well aware of the principles under which the national statistical office operates, which fully reflect the Fundamental Principles. Very few chief statisticians reported that their superiors were not at all aware of such principles. III. Concluding remarks 119. On the basis of this self-assessment, it seems that the Fundamental Principles of Official Statistics are remarkably well implemented. Confidentiality (principle 6) and Legislation (principle 7) seem to be the best implemented principles, although 30  E/CN.3/2004/21 incidents of disclosure have been reported and quite a few countries mentioned that their law was outdated. On the other hand Prevention of Misuse (principle 4) and National Coordination (principle 8) seem to be the least implemented principles. Major problem areas that prevent a better implementation of the Fundamental Principles mentioned included: authorized and unauthorized disclosure of individual data for non-statistical purposes; political interference at the dissemination stage; the need to adapt international standards to national circumstances; and lack of resources. At a general level, there seems to be a strong correlation between statistical capacity and adherence to the Fundamental Principles of Official Statistics. IV. Points for discussion 120. The Statistical Commission may wish to: (a) Discuss whether the present survey represents an adequate picture of the level of implementation of the Fundamental Principles of Official Statistics and the main factors preventing their full implementation; (b) Provide guidance to the Statistics Division on whether/what focused action should be taken to further improve the implementation of the Fundamental Principles; (c) Endorse a global review of statistical capacity. Notes 1 See Official Records of the Economic and Social Council, 2003, Supplement No. 4 (E/2003/24), chap. I.A. 2 The Division gratefully acknowledges the contributions of Mariam Al-Awadhi, Jean-Louis Bodin, Heinrich Bruengger, John Cornish, Hermann Habermann, Andrew Flatt, Lucie Laliberté and Tamas Mellar. 3 These are: the 191 Member States, Niue, the Cook Islands and Palestine. Hence, in the present report “countries” refers to the national statistical offices of the Member States, Niue, the Cook Islands and the Palestinian Central Bureau of Statistics. 4 http://unstats.un.org/unsd/methods/statorg/review.htm. 5 In addition to the general question opening each principle, there was a concluding phrase asking for “any additional comments on the implementation of principle xx in general in your country”. Responses to these questions will, wherever possible, not be reported separate