This study was funded and sponsored by three organizations: the National Commission on Libraries and Information Science (NCLIS), the National Science Foundation (NSF), and CONTU. The need for it became apparent when the Working Group of the Conference on Resolution of Copyright Issues found itself unable to agree on the actual volume of library photocopying. The conference, which had been organized in 1974 by the Register of Copyrights and the chairman of NCLIS, agreed in 1975 to participate in drawing up a request for proposals. NCLIS was joined by NSF, and in 1976 the contract was let to Market Facts, Inc. (later King Research, Inc.). Soon thereafter, CONTU added funds to enable a detailed analysis of the transactions of the Minnesota Interlibrary Telecommunications Exchange (MINITEX).

{Page 132}

The Survey

The contractor secured from the National Center for Educational Statistics and other sources a master list or sample frame of more than 21,000 libraries in the United States, a list believed to include most of the libraries of any consequence, except for public and private elementary and secondary school libraries, which are considerably more numerous. The libraries were divided into four types: academic, public, federal, and special (the latter frequently but not always serving for-profit organizations). A sample of 360 libraries was drawn so as to fully represent each type. While most of the sample was chosen randomly, a number of the largest libraries were deliberately chosen because of the scale of their photocopying activities. Responding libraries reported only on photocopying done on machines operated or supervised by staff members; unsupervised (including coin-operated) machines were excluded. The libraries in the sample frame had more than thirty-five thousand photocopying machines, of which twenty thousand were used exclusively by the staff.

The numbers found throughout the King study generally constitute estimates based on projections against the nationwide sample frame rather than actual data or observations from which the estimates are extrapolated. The estimates are subject to varying amounts of uncertainty, depending on the number of observations or the length of time in which they were made. Estimates concerning photocopying in one kind of library are therefore often subject to greater uncertainty than estimates concerning all libraries. One of the reasons for adding the MINITEX part of the study was to provide a basis for checking the results of the national library survey, and the results agree quite well.

Findings

The following tables present some of the basic results for the 1976 calendar year. Some totals may not add exactly due to rounding of the numbers. Materials of indeterminate copyright status are not included in the stated totals of numbers of copies made from copyrighted works. The unit of count is a complete document, whether one page or twenty pages.

TABLE H-I

HALF OF COPYING COMES FROM COPYRIGHTED WORKS

Type of Library

No. of Libraries

(x 1,000)

Average No. of Copies

(x 1,000)

Total No. of Copies

(col. 1 x col. 2)

(x 1,000,000)

Percent Copy-righted

No. of Copyrighted Copies

(col. 3 x col. 4)

(x 1,000,000)

Public

Special Academic

Federal

8.3

8.5

3.0

1.4

7.7

3.1

5.5

4.9

U.S. TOTALS

21.3

5.4

114

The above table shows that for all kinds of materials, about one-half of the copies made came from copyright-protected works. The 114 million copies amounted to about 1 billion pages; copyright status could not be determined in 17 million copies. The next table shows that, for all kinds of libraries throughout the United States, serial publications accounted for most of the copying of copyrighted works.

TABLE H-2

MOST COPYING OF COPYRIGHTED WORKS COMES FROM SERIALS1

Type of Materials

No. of Copies

(x 1,000,000)

Percent Copyrighted

No. of Copyrighted Copies

(col. 1 x col. 2)

(x 1,000,000)

Serials

Books

Others

Due to the unclear interpretation of some of the definitions in section 108 of the new law, {Page 133} and without prejudice to their resolution, King Research, Inc., felt it was necessary to break down copying into three kinds of services: (1) copying for ordinary local users, induding employees of organizations served by the library; (2) copying for users at another branch within an overall library system; and (3) copying for interlibrary loan. As the next table shows, for all kinds of libraries and materials, copying for local uses was the dominant activity.

TABLE H-3

MOST COPYING Is FOR LOCAL USERS1

Kind of Service

No. of Copies

(x 1,000,000)

Percent Copyrighted

No. of Copyrighted Copies

(col. 1 x col. 2)

(x 1,000,000)

Local

Intrasystem

Interlibrary loan

An area of particular interest was that of copying from copyrighted serials for interlibrary loan. The King study revealed that the CONTU guidelines, in combination with provisions of the copyright law itself, greatly reduced the number of such copies needing authorization. There were 3.8 million such copies made in 1976, a number reduced to 2.4 million if one excludes copies made from serials over five years old. The exemptions for replacement of damaged or missing items and for classroom use further reduce the number to 2 million. After applying the CONTU guidelines, which permit up to five copies per serial title for each requesting library in a given year, there are 500,000 remaining copies needing authorization. The status of material over five years old remains unclear, however, making this estimate a lower limit.

The distribution of copying, by size of library, was quite uneven. Large libraries dominated; in particular, 20 percent of all libraries accounted for almost 80 percent of copies made for local users and almost 75 percent of those made for interlibrary loan. Indeed, since the number of supervised machines in the libraries surveyed was smaller than the number of libraries themselves, some must have had no supervised machines at all.

The distribution of copying was also uneven with respect to source materials, especially serials. Although copying from journals seemed to bear little or no relation to circulation levels, 20 percent of them accounted for almost 70 percent of copies made for local use and over 85 percent made for interlibrary loan. If all the exemptions for interlibrary loans were applied, 90 percent of serial titles would have fifty or fewer copies made needing authorization from them throughout the country. Very few, if any, would have one hundred or more such copies.

In addition to counting photocopies and estimating totals, the King study also questioned libraries about their preferences regarding the design of a mechanism to collect and distribute royalties for photocopies needing authorization under the copyright law. Describing the choice to be made between a system of complete reporting of copying activity at one end and a system of minimum reporting at the other, the report noted that greater accuracy in collection and distribution of payments would require a more complex and costly system. Librarians seemed to prefer a simpler system which, although less exact in its payments, would be easier to administer.

1 Cf. King study. pp. 45, 47, 49. The table on p. 47 has two typographical errors in the “All Libraries” row in the sixth and tenth columns.

REPORT TITLE: Library Photocopying in the United States, with Implications for the De­velopment of a Royalty Payment Mechanism

Background

The Survey

Findings

REPORT TITLE: Library Photocopying in the United States, with Implications for the Development of a Royalty Payment Mechanism