Statistical Journal of the IAOS

Statistical services should offer a broad spectrum of products in order to meet the requirements of different user groups. This can only be achieved if demands for statistical information are well known and, on the other hand, if statistical prodcuts are well accepted as useful information tools. Results of a survey in the Federal Republic of Germany show that the general acceptance of official statistics is closely connected to the users' assessment of the services provided by the statistical offices. The study showed that there is great general acceptance of official statistics, particularly by those respondents who have experienced statistics to be relevant to their own informational needs. While the results of the study have served to develop a more target-group oriented approach in presenting official statistics to the public and to market the broad spectrum of statistical products accordingly, there remains a need for continuous press and public-relations work to enhance communication between the statistical service and its clients.

In the article, a typology of the use of official statistics is proposed. Following an introduction to the subject, three dimensions of the use of official statistics are distinguished analytically: user categories, objectives of use, and methods of use. Each of the three dimensions is subdivided into four categories. The dimension user categories is subdivided in government, science, business, and the public; the dimension objectives of use consists of knowledge, preparation for choice and action, evaluation of choice and action, and routinization of choice and action; and, finally, the dimension methods of use is subdivided into consultation, monitoring and comparison, formal analysis of aggregate data, and formal analysis of individual-level data. Some of the more frequently occurring combinations are illustrated with the aid of examples from the Netherlands, the United States, Canada, and the European Communities. After some introductory statements on priority setting (an evaluation of the significance and the cost of statistical projects), the relationship between type of use and priority setting is clarified. It is stated that, since a classification cannot automatically lead to a particular outcome in terms of priorities, the primary goals for a classification of types of use are relevancy and structure of information. The final paragraph discusses the likeliness of each of the 64 possible combinations; it is concluded that in the Dutch situation, for several reasons, almost half of these 64 combinations are more or less unlikely to occur.

The intensity of the changes taking place in the transition period and their effects on the statistical services differ from country to country. Hungarian statistics have not been hindered by central planning, as the role of this planning had decreased in recent years and statistics went far beyond the scope of planning.

In identifying users, government has been given first priority. However, with changing market conditions the statistical services are being faced with new types of requests coming from Parliament, political parties, the business sector, etc. The main challenge is the correct definition of the new requests, an optimal selection from them, and the anticipation of future needs.

This article identifies four types of data dissemination: statistical publications, special publications for the media; individually-tailored dissemination; publication through electronic media.

New conditions for data collection are created by the increasing role of the private sector, the growing number of economic units, and the decreasing readiness of data suppliers. One of the first tasks is the development of an appropriate business register. The basic information would come from the tax authorities.

This article provides a sample of fields of intensive activity. National Accounts and price statistics are discussed in some detail, together with the assessment of the impact of reforms on the population.

Statistical agencies serve many users. These users are heterogeneous in terms of their information needs, the ways in which they use statistical data, their analytic abilities to manipulate data, and their computing capacities. The impact of their uses of statistical data will also differ greatly depending on their roles in society. If the statistical agency is to be responsive to the needs of a wide range of users, it must be aware of these needs and must tailor its products and services accordingly. This paper reviews and assesses approaches that have been used in Canada to ensure that the products of Statistics Canada meet, and continue to meet, the needs of its various clienteles.

The paper begins with a brief review of mechanisms for keeping abreast of users' information needs. It then defines a set of categories of users and uses of the outputs of a statistical agency. For each category, the paper summarizes the uses to which statistical data may be put, the major characteristics of the data needs, additional mechanisms that may be used to keep abreast of these users' needs, and ways in which the products and services of the statistical agency may be tailored to satisfy the needs of this particular set of users. The paper concludes with a review of the problem of setting priorities among the conflicting demands of different user groups.

The Austrian Central Statistical Office (ACSO) is facing new requirements provoked by the emergence of the Single Market and by developments in Central and Eastern Europe. This will have consequences for ACSO's structure, output planning, applied statistical techniques and methods, and electronic data processing.

For example, if Austria becomes a member of the European Economic Area, statistical standards and rules of the Community must be applied. Therefore, ACSO's target for the 1990s is to achieve fully EC-compatible national statistics. Some fields are already highly adapted to international standards. However, the non-agricultural census – to give an example – lacks some methodological requirements. The enterprise needs to be introduced as the reporting statistical unit.

Along with those methodological changes, EDP-oriented facilities should be introduced. Up to now there exists an “all governing” EDP department, but it is planned to initiate a more user-oriented system of workstations for statisticians.

The Central Bureau of Statistics, Norway, CBS, is highly dependent upon computers for both production of statistics and economic research and analysis. Motivated by shortcomings of mainframes and PCs in this work, CBS has introduced and tested out Unix based workstations in economic research and analysis.

For computational intensive tasks such as parameter estimation and simulation workstations proved quite superior to PCs: CPU time is shorter and available memory is larger. A comparison with mainframes gives a less unambiguous conclusion. Due to quick input/output rate on the mainframe, the CPU time required on workstation sometimes exceeds that of a mainframe. However, because of the time sharing system of a mainframe, elapsed time on a mainframe may still be longer than on a workstation.

Communication problems of introducing workstations into the existing computer park has been solved satisfyingly. Data transmission is performed smoothly. When using PCs as terminals to workstations one faces minor shortcomings, which future network innovation hopefully will remove.

CBS intends to enlargen the workstation stock in the near future. Most macroeconomic models use the fourth generation language Troll, which so far has been available only on mainframe. When a workstation version is released in 1992, CBS plans to convert most macroeconomic models to workstation.

“Provision of statistical services in Romania at territorial level” deals with the statistical problems of the regional offices of the National Commission for Statistics. It is divided into three sections. The first briefly presents the past situation for statistical services at territorial level. At the same time it explains the characteristic elements, the deficiencies of the applied statistical system, and some considerations of the impact of the planning and restrictions imposed by the central level on statistics. The second section deals with the strategy for improving and adapting statistics to the requirements of the transition period. The third section presents specific aspects of regional statistical services during the transition period.

Ensuring that statistical offices are fundamentally service organizations requires a concerted and organized marketing effort. At Statistics Canada, new marketing initiatives have been the driving force behind major changes to the Agency's corporate culture, product line and business management approach. The marketing approach used determines market needs, develops suitable products, and then disseminates and promotes those products.

This paper first outlines how Statistics Canada determines client needs through the use of advisory committees, market feedback and program evaluations. It then explains the Agency's major product line redevelopment initiatives and its policy of balancing the costs of production (excluding data collection and processing) with earned revenues. The paper discusses how the Agency maintains its service to the public by making information available as a public good as well as its shift to more user-funded surveys and studies. The paper concludes with an analysis of how the new marketing policy has helped to maintain the integrity of Canada's statistical system in an era of budget reductions.

The word “marketing” in the context of public activities has been in recent years usually associated with the “privatization” of the public sector or at least with its development towards a set of market-oriented and businesslike activities. A traditional civil servant – or a statistician – has so far followed this development with rather strong reservations. No wonder, because the world view of a statistician actively marketing his or her services is so different from the world view of a traditional statistician as a civil servant or researcher.

In any case, we may find that the new kind of marketing-oriented statistician is a proper response to increasing requirements for more user orientation – a trend much discussed by statisticians for well over ten years. The traditional, somewhat inward-looking professional and the new outward-oriented salesman have to be combined somehow within the statistical service, because both are needed. Finnish experiences during the 1980s indicate that market orientation is useful for connecting the statistical service with the rest of the society, while the actual production of statistics still requires the traditional professional approach.

In April 1989 the then Prime Minister, Mrs Margaret Thatcher, announced measures designed to achieve a significant improvement in the quality and relevance of government economic statistics. This led to an enlarged Central Statistical Office which became a separate government department. It took over responsibility for most of the United Kingdom's economic statistics, including the over-seas trade figures and the retail prices index. It quickly emerged that dissemination of statistics would be a priority. In particular there was pressure on the CSO to get its press relations absolutely correct from the start.

The paper examines how the operation was carried out and has developed. It also looks at the publications policy inherited from the old CSO and other departments and possibilities for future improvements.

It can be seen that the creation of the new CSO led to a radical re-examination of how Britain disseminates its official statistics. This involved introducing new and innovative practices and is in keeping with the declared intention of the Central Statistical Office becoming one of the Government's new-style executive agencies. These are being formed under what is called the “Next Steps” policy aimed at providing a sharper focus and emphasis on the way government delivers its services to the public and other customers.

The economic censuses are mail surveys of business establishments in the United States, conducted once every five years. They cover seven sectors of the economy: retail trade, wholesale trade, services, transportation, manufacturing, mining and construction. These censuses present two marketing challenges: promoting the timely completion of census forms by companies during data collection, and delivering the results of the censuses to researchers and decision-makers.

Census forms were mailed to 3.9 million businesses in December 1987 with a due date of the following February 15. While response is required by law, only about half of all businesses responded by the due date, and a series of follow-up letters were necessary to obtain responses from most of the remainder.

To promote timely response, the Census Bureau included explanatory brochures with each questionnaire mailed, asked hundreds of business publications to write articles about the censuses, and conducted a multi-media public service advertising campaign under the auspices of the non-profit Advertising Council. We conducted an attitude and response behavior study during data collection, and the results have yielded a number of ideas on how we should change our approach to respondents in the 1992 Economic Censuses.

Most of the 1987 census results have now been published in a series of over 500 printed reports, 20 computer tapes, and seven compact laser discs (CD-ROMs). We have promoted their dissemination and sale not as a source of income – since the law puts all census publications in the public domain and allows us to recover only the cost of reproduction – but as a way to increase the value the nation receives from its investment in this data collection. Marketing efforts have included product redesign, simplified product ordering, publicity in user-oriented periodicals, publication of guides and brochures, and holding a nationwide series of user conferences. In the long run, our most important activities may well be our efforts to effectively package our published data on CD-ROM for use on microcomputers, and our work to establish a network of intermediaries – librarians and state data centers – that can provide census information to researchers, government, the general public, and especially the businesses from which we obtained the census information in the first place.

An outline is given on new tendencies in Bulgarian State Statistics during the nation's transition to a market economy. The statistical information service is distinguished by the way it publicizes statistical data and by its thorough and current submission to public, Government, and social authorities.

The marketing of statistical information and services will play an important role in the evolution of State statistics.

All services are in principle free of charge, but attempts are being made to provide certain services against pay.

Marketing of statistical information and services is carried out in several ways:

a) assessment of needs for statistical information and services

b) segmenting of the marketplace;

c) organization of promotion and price policies

Modern collecting, compiling, processing and publishing of statistical data demand automated work stations (AWSS) which should be part of the METAINFORMATION SYSTEM of the CSO as well as of the Marketing Information System. The work stations will provide complete information on marketing of statistical information and services. This system should consist of four subsystems – for home information, outside information, marketing investigations, and analysis of complex information.

An overview of practices is given in ECE member countries about costing and pricing of statistical products.

Although the research of good cost information is common to all countries, there are still great differences in the accuracy of costs calculation. The same is valid for their utilization.

As for the pricing of statistical products (publications, “tailor made statistics”, databanks), the general policy is to charge only at marginal cost, the involved costs for data collecting and data processing being considered as already paid by the tax payers.

The setting of fees for the utilization of databanks remains the most difficult one.