graphic distribution. Additional data elements include net sales, total employment product, field of applied research and development, a breakdown of costs (e.g. payroll, materials, etc.) and size of company by R&D expenditures for pollution abatement and energy purposes, R&D performed by American firms outside of the U.S. and R&D contracted out to others by industrial firms. Various ad hoc studies have provided additional in-depth information on the basic research needs of industry and on R&D in various industries such as chemicals and allied products, electrical equipment and communication, and aircraft and missiles.

Universities and Colleges

NSF has two ongoing annual surveys dealing with the university and college sector.

The first obtains Federal obligations data from the 14 agencies that provide major support data for academic science in the university and college sector. Data are collected on obligations to individual institutions for R&D; R&D plant, facilities and equipment for instruction in science and engineering; fellowships, traineeships, and training grants; general support for science; and other science activities. These data are described in an annual NSF study, Federal Support to Universities and Colleges, and Selected Nonprofit Institutions: A Report to the President and Congress.

R&D expenditure data are also collected by NSF directly from educational institutions and reported annually by NSF Expenditures for Scientific and Engineering Activities at Universities and Colleges. Data are provided by source of funds, geographic distribution, field of science, institutional control, type of R&D activity, type of institution, according to degrees granted in science and engineering. Information is also provided on capital expenditures for research, development and instruction in science and engineering; current direct expenditures for instruction; and estimates for departmental research. Nonprofit Sector

Data collected in an NSF survey conducted periodically on the R&D activities of independent nonprofit institutions include information by source of funds, field of science, type of institution, and geographic distribution. In addition, there is information on capital expenditures. Surveyed organizations include research institutes, nonprofit-administered federally funded research and development centers, voluntary hospitals, private foundations and other nonprofit organizations. Information is reported in an NSF publication, R&D Activities of Independent

Nonprofit Institutions. Nonprofit R&D data are also reported in an annual NSF publication, Federal Support to Universities, Colleges, and Selected Nonprofit Institutions, for those institutions which have received $300,000 or more in Federal R&D obligations.

National Overview of R&D Funding Activities

The publications which provide overviews of funding activities deal with all major economic sectors and derive their data entirely or mainly from NSF's recurring individual sector surveys. As such, their data components are restricted to those elements describing R&D activities common to all

sectors.

Data collected from the sector surveys are consolidated into national totals and appear in the annual NSF overview publication, National Patterns of R&D Resources. The report focuses on the flow of funds from source to performer and on the flow of funds among sectors. The data are arrayed in a series of matrices which provide a long-range view of the allocation of funds and the changing roles of the performers. R&D funding patterns are illustrated in terms of Federal and nonfederal funds, by selected objective (e.g., defense and space-related activities) with respect to gross national product. Trends in intersector transfers of funds and in the character of work performed are also included. An important feature is the time series (from 1953 on) provided for intersector transfer of funds and sources of funds by character of work.

Another overview of R&D funding is provided in a publication, Science Indicators, prepared by NSF staff for the National Science Board. The report incorporates current data provided by NSF surveys and other sources and presents a series of indicators designed to demonstrate the strengths and weaknesses of U.S. science, including international comparisons of R&D funding levels.

A third perspective is provided by another NSF study, 1985 R&D Funding Projections. Projections are provided for total R&D funds by source and performer.

Other Federal Agencies

In addition to NSF, there are other agencies which collect and analyze science and technology statistics.

The Smithsonian Science Information Exchange collects, processes for storage and retrieval, and disseminates information on basic and applied research being conducted by the Federal Government and the private scientific community. Studies are prepared on request of the number and types of

research activity underway in a specific area or discipline. Data elements provided include source of funds, R&D performer, location of performance, most current year project was underway, and research area or discipline (e.g., dental research, cancer research). Data are not aggregated except in response to a specific request. Reporting is voluntary, although some Federal agencies and private organizations have established mandatory internal reporting requirements. Frequency of reporting varies with the organization (monthly, quarterly, annually, etc.). The coverage of federally supported research is estimated to be 80 to 90 percent of total projects; coverage of major nonprofit foundations is considered to be good particularly in the medical sciences; the coverage of industrial and university sponsored research is very limited. There is also international coverage of cancer and energy research, collected from international organizations and government ministries.

The Department of Commerce's Bureau of Economic Analysis (BEA) conducts a periodic survey of U.S. investments in foreign countries; this survey collects information on total R&D expenditures abroad. The last survey collected 1966 data and the next one is expected to be for 1977 data. BEA will publish the results in the Survey of Current Business.

The National Institute of Health's Division of Resources Analysis prepares a series of specialized and recurring reports focusing on various dimensions of financial support for medical and health-related research and development. One of these provides an overview of national health R&D support, and is based on Federal data collected biennially from agencies involved in biomedical and health-related research activities, reports published by government and private organizations, and estimates made by NIH.

The Federal data elements include source of support, performance sector by grants versus contracts and intramural, field of science, and individual medical and academic institutions receiving Federal health R&D support. Pharmaceutical R&D data are obtained annually from the Pharmaceutical Manufacturers Association. Other industry data are obtained from the NSF industry survey and NIH estimates. Voluntary health agency and foundation data are elicited from respective annual reports and periodic surveys. State government expenditures are derived from periodic NSF State agency surveys and annual university expenditure data adjusted by NIH to extract the health R&D portion. A report on health R&D expenditures was published in 1977 by NIH, "Dollars for Health Research and Development, 1968-1975."

Summary figures on health R&D have appeared in NIH publications, such as the annual NIH Data Book and NIH Almanac, and in the Social Security Bulletin, which uses NIH data for its national health expenditures series.

NIH also publishes periodically a report on the Department of Health, Education, and Welfare's obligations to institutions of higher education, health professional schools and other nonprofit organizations, based on data collected by NIH for the National Science Foundation (Survey of Federal Support to Universities and Colleges, and Selected Nonprofit Institutions). The NSF report contains data by source of support, beneficiary institutions, type of obligations, and regional distribution. Information is also provided separately by type of institutions, medical schools, universities and colleges, nonprofit hospitals, and other nonprofit organizations.

The Securities and Exchange Commission (SEC) annually collects R&D funding data from publicly owned corporations as part of its general financial reporting requirements. Funding information is provided on total company R&D expenditures (both domestic and foreign) and the sources of funding, whether company-owned or Federal funds. Only those companies whose R&D expenditures exceed 1% of total sales and revenues are required to report this information. The data are maintained by SEC in files by company and are not aggregated. SEC does not publish or analyze the data.

The Federal Power Commission publishes a number of studies containing data on the operations of publicly and privately owned utilities. Included in these reports are statistics on the financing of R&D activity. Two reports, Statistics of Privately Owned Electric Utilities in the United States and Statistics of Publicly Owned Electric Utilities in the United States. are published annually. They include summary data on total R&D expenditures for all utilities and total R&D expenditures as a percent of total expenditures by source of funding (intramural and extramural). Data are provided by various classes such as hydroelectric, nuclear power, fossil fuels, transmission, distribution, environmental effect and systems planning. Data cover a three-year period including the current year and the preceding two years. Similar information is also developed for gas utilities.

Gaps and Limitations in Statistics on Funding of Science and Technology

As noted earlier, the only area of science and technology activity financing for which a major, ongoing data collection and analysis effort is

maintained is research and development and even here, a complete picture is not currently available. Missing from national R&D and R&D plant expenditure totals are data on social science research in industry, departmental research in universities and colleges, R&D plant expenditures in the industry and academic sectors. No more than periodic benchmarks are available on R&D funding in State and local government agencies. Taken together, these elements are felt to represent only a small portion of the Nation's total R&D expenditures.

There are also examples of data on R&D financing whose utility is restricted somewhat because of their limitations. Such is the case with data on industrial R&D expenditures. Requirements for confidentiality of individual company data significantly reduce comparative analysis with other types of industrial data. Another limitation of the data on industrial R&D financing results from the current practice of obtaining data at the company level. Data for large conglomerates whose industrial activities encompass several Standard Industrial Classification categories are included under the primary product line of the company. Efforts are currently underway by NSF to determine the feasibility of obtaining data at the division or establishment level of the company. Although individual company R&D data for public companies do exist in SEC files, definitional and classification limitations prevent meaningful comparisons with other forms of industrial economic indicators.

Also missing are nationally aggregated financial data on research and development by objective or function such as environment or transportation. Such data do exist for Federal obligations but not for the R&D activities of the remaining sectors. Furthermore, even the Federal data are derived by NSF from budget information of major organizational units of Federal departments and agencies, and, as such, represent only a rough estimate. However, these data could probably only be improved by means of project-by-project reporting by the agencies; this, however, could be a very costly investment of financial and staff resources.

Similarly, little financial data are separately available for R&D activities of Federal in-house laboratories. To date, other priorities within NSF and the participating agencies have prevented the compilation and reporting of data at this level. An effort is planned, however, to obtain these data in the near future.

Data collection for science and technical information (S&TI) expenditures on a national level is currently limited to science information activities

of the Federal Government. Even here, data are known to reflect only a portion of the total Federal funds that go into the storage, retrieval and dissemination of scientific and technological information. The amount representing science and technical costs associated with R&D contracts and grants is excluded from available S&T data. At present, there is no feasible means of obtaining this information.

In both of these areas, the National Science Foundation has attempted to overcome these limitations and gaps through continuous efforts to develop concepts, definitions, data collection methodologies, and respondent cooperation.

Financial data covering other activities that fall within the scope of science and technology activities are not available from any public or private source, even on a nationally aggregated basis. Data on total science education expenditures by educational level are an example. At the present time, only data representing Federal obligations to universities and colleges for science education activities are available. These data are published annually in an NSF report entitled Federal Support to Universities, Colleges, and Selected Nonprofit Institutions. Other examples of major types of science and technology activities for which little financial data are currently available include:

-total expenditures used in prospecting for mineral, geological, hydrological, and other natural resources;

Unfortunately, these measures are often greatly affected by other inputs as well. Consequently, it is extremely difficult to relate changes in these measures solely to developments in science and technology. Of the few measures that have been identified as more or less exclusively the result of scientific and engineering activity, several represent a comprehensive set of available data. These are in the areas of patents and citations. Studies of these indicators have appeared in the National Science Board's report, Science Indicators.

The Patent Office works to make the enormous file of patent grants useful for a variety of purposes and to present the data in the file in various forms through a number of reports. These reports spotlight the areas of technology experiencing a high level of overall activity or of foreign activity. Technologies of interest are followed over time, and geographic breakdowns are made of patents from the various States and foreign countries. The latest report adds descriptions of some individual patents and presents patent activity according to date of application, in addition to date of grant. There is also a description of the five U.S. patents and one foreign patent cited most often in U.S. patents granted in 1975.

All these published materials are taken from the Patent Office's computer file of patents, which was begun in 1970. At present, the file includes: subclasses of the U.S. Patent Classification System, and the Standard Industrial Classification, in 52 Product Fields and Product Field combinations. For U.S. patents issued since 1963, data are available on: category of ownership at time of issue (U.S. Government, foreign government, U.S. corporation, foreign corporation, U.S. individual, foreign individual, and unassigned); country or State of residence of the inventor; date the application for patent was filed in the U.S. Patent and Trademark Office; specific (i.e., named) ownership of all patents which, at time of issue, were organizationally owned (e.g., by a corporation, foundation, or government agency).

Currently, data on patents granted by foreign governments are being added to the files for the first time. The Patent Office also plans to add information on 600 large companies that own patents. This will include data on the R&D expenditures and sales of those companies going back about 20 years.

Transactions in royalties and licensing fees can be used as indicators of the transfer of technical knowhow. The U.S. Department of Commerce keeps records of U.S. transactions in royalties and fees and publishes series data in Survey of Current Business under receipts (direct investment-related and unaffiliated) and payments (also separated into direct-investment related and unaffiliated). Additionally, a new report has been published by the Bureau of Economic Analysis entitled Revised Data Series on U.S. Direct Investment Abroad 1966-1974

Publication counts serve as a quantitative output measure of scientific research, especially basic research. Correspondingly, one measure of the overall quality of research is provided by citation counts, which indicate the frequency with which specific papers are cited in other publications. Extensive data compilations and analyses involving publications, citations and such related variables as field of science, sector of origin, and country have been developed in time series covering the last decade and have been published in the National Science Board's Science Indicators report.

The state of the art in science and technology output measurement is in its early infancy. Even those measures cited aboved are in need of further work. Attempts to develop additional indicators of output have been made for R&D activity and technological innovation but, to date, none have proved satisfactory. However, these developmental activities are continuing and new approaches should be explored.

Introduction

The objective of the Federal Government's collection of statistics on the environment is to facilitate and evaluate programs which are designed to help ensure public health and safety. Included are data on working conditions, the quality of the ambient air and water, and the effects of noise, wastes, and chemicals and industrial processes on man as well as factors which contribute toward his survival, health, and general well-being. There are no measurable parameters which directly correspond with the objectives of air quality, water quality, safe work environment, and ecological balance any more than there is a medical parameter called health or a socioeconomic measure termed quality of life. Nevertheless, there is a need for data on environmental conditions which (1) facilitate public understanding, (2) are produced on a timely basis, (3) impart an understanding of their significance, (4) are amenable to tests of validity and reliability, and (5) satisfy the needs of private and public decisionmakers.

The facts and factors measured and analyzed by the environmental statistician are the product of complex industrial and social phenomena. There are many important interdependencies, for example, chemicals discharged into a stream may interact to form substances assimilated by fish which are hazardous to man, because of his dependence upon fish as food.

Although interaction and interdependence characterize the products and processes measured by the environmental/workplace statistician, the material presented in this chapter shows that environmental statistical activity is marked by arbitrary fragmentation of responsibility and insufficient communications among interdependent activities. Consequently, this chapter concludes that a comprehensive review of the need for and means of producing Federal environmental statistics is required.

A comprehensive assessment of the need of decisionmakers for environmental/workplace data is needed, because perceptions of what should be

measured in describing the environment have changed. For example, it is now recognized that measurement of pollutant discharge is insufficient, since the interactions of some pollutants in the ambient air or water often result in products which bear only a remote resemblance to, and can pose a greater threat to health and environment than, the original substances. Moreover, the environment transports and transforms pollutants in such a manner that their end states and effects on man may require complex investigative procedures and instrumentation for adequate detection and control. For example, the effects of toxic substances and particulates disguise themselves as the degenerative diseases of old age.

In addition, monitoring and assessment systems have been developed on the basis of perceptions of data needs which are now undergoing review. Water quality monitoring systems, for example, which measure biochemical oxygen demand (BOD) and fecal coliform for use as proxies for other pollutants, were developed to meet regulatory requirements. These systems may not meet the needs of research directed toward the development of new standards. The same assessment is relevant to occupational safety and health monitoring, air quality monitoring, and monitoring of pesticides production and utilization.

Responsible Agencies

The Federal Government collects, analyzes, and publishes environmental statistics on air quality and emission data; water quality and water supply; environmental contamination from radiation; noise pollution; toxic substances; the production, distribution and use of pesticides; the effect on the environment of energy sources now in use; the potential impact of new forms of energy upon the environment; and occupational injuries, illnesses, and exposure to workplace hazards.

Although the Environmental Protection Agency (EPA), through the use of federally mandated State programs, is the primary data collection agency for statistics on the environment, these data are