University of Tennessee School of Information Sciences Center for Information Studies Paper presented to the ICSU Press Workshop, Keble College, Oxford, UK, 31 March to 2 April 1998
This paper summarizes costs of publishing scientific scholarly journals. Activities are described for five publishing components: article processing (e.g., manuscript processing, editing, composition, etc.), non-article processing (i.e., similar activities related to covers, tables-of-content, letters, book reviews, etc.), reproduction (e.g., printing, collating, binding, etc.), distribution (e.g., wrapping, labeling, sorting, mailing, subscription maintenance, etc.), and support (e.g., marketing, administration, finance, etc.). A model is derived for each of these components consisting of cost parameters (e.g., number of issues, pages, subscriptions, etc.) and cost elements (e.g., cost per page of editing, set-up cost per issue, postage cost per issue copy mailed, etc.). Total costs of a "typical" journal are presented where cost parameters are estimated from a sample of scientific scholarly journals and cost elements derived from estimates reported in the literature. Costs of electronic publications are also derived.
In preparation of two invited articles on electronic publishing [1, 2] it was apparent that a great deal was said about electronic journals with little data and known facts to support opinions concerning the future of paper and electronic journals. Yet journal system participants such as authors, publishers, libraries, readers and funders (of readers and libraries) can all be profoundly affected by the transition to electronic publications; perhaps negatively if some lessons from the past are not learned. A Special Libraries Association Steven I. Goldspiel Research Grant was given to pull together results of over 50 research studies performed by King Research and University of Tennessee, School of Information Sciences [3,4,5,6] and research reported by others. The objective is to establish trends and provide data regarding authorship, publishing, distribution and use of scientific scholarly journals. These data were analyzed and are being reported to help the system participants mentioned above deal with changes in a positive and rewarding manner.
These studies provide trends concerning scientific scholarly journals dating back to 1960 including publishing (cost and number of journals, articles, pages, citations, etc.), authorship (number of authors/authorships, cost, revisions, time, etc.), readership (number of articles read, distribution of reading, cost, means of identification, sources, photocopying, usefulness and value of information, etc.), pricing (by type of publisher, sensitivities of personal and library subscriptions), library services (number of readings by type of service, service attributes, importance and satisfaction ratings, service costs, etc.), and interlibrary borrowing/document delivery (number of articles obtained and provided, cost, attributes of service, etc.). The studies involved surveys of 12,668 scientists (and over 8,000 other professionals, not discussed here) tracking a sample of scientific scholarly journals from 1960 to 1995 and in-depth cost studies.
This paper addresses publishing costs. The objectives of the paper are to:
Detailed costs are presented below to achieve these objectives.
In order to assess publishing costs it is helpful to describe all the activities that are necessary to publish scholarly journals. Journal publishing activities are sub-divided into five types of activities including article processing (i.e., all article-related activities necessary to produce the first copy of a journal issue); non-article processing involving letters, editorials, book reviews, and so on; journal reproduction (i.e., printing and binding); distribution (i.e., subscription maintenance, wrapping and mailing); and publishing support activities (i.e., marketing and sales, administration, finance, non-allocated resources and other indirect cost activities). The cost of these groups of activities is discussed below, along with an approximate cost model of relevant activities including publishing cost parameters that affect cost (e.g., number of pages published, number of subscribers, etc.) and specific cost elements related to these parameters. Cost parameters are taken from estimates of 1995 scientific scholarly journal averages  and cost elements are derived from one or more sources of data found in the literature or educated guess. Estimates of total cost of each of the five sets of activities are derived which lead to an estimate of the total cost of a "typical" scientific scholarly journal. From these costs we provide average costs per subscription . Thus, we have some picture of how much publishers must charge in order to recover publishing costs at various levels of circulation.
2. Journal Publishing Activities
The activities presented here are those which publishers and their representatives perform as part of traditional print journal publishing. We have chosen to divide these activities into five functions: article processing, non-article processing, journal reproduction, distribution and publishing support. The total cost of article processing activities is largely a function of the number and size of articles handled, and, thus, the unit of cost of processing articles is important to know. The cost of non-article processing is isolated because it is often ignored in costing scholarly journals, yet it can be a significant cost that must be considered in assessing journal prices. The costs of journal reproduction and distribution are largely a function of number of subscribers, frequency of publication, and the size of the journals. The costs of these functions are important because they are the ones most affected by switching to electronic publishing. Total article and non-article processing costs are fixed costs identified with a particular journal. These are fixed costs in the sense that they are not affected (or affected very little) by number of subscriptions and are most often called first copy or prerun costs. Another kind of fixed cost involves indirect publishing support costs of marketing, administration (i.e., accounting, personnel, utilities, etc.), managing rights and copyright protection, and finance. However, these costs do vary some over large changes in size of journals and number of subscribers. A partial list of scholarly journal publishing activities are given below by the five categories mentioned above. These activities are extracted from several sources [7, 8, 9, 10, 11, 12].
Article processing involves a number of activities that begin with receipt of an article manuscript and ends with "first copy" composition, typesetting, and/or engraving. These activities include receipt processing, initial disposition decision-making, identifying reviewers/referees, review processing, subject editing, special graphic and other preparation, formatting, copy editing, processing author approval, redaction, and preparation of master images. Most of these activities are designed to enhance favorable attributes of article content and its presentation. Non-article processing involves essentially the same activities as article processing although the authors are often internal staff and peer review is not necessary. Subject editing may require less effort and special graphics may be minimal. Master images may be provided by the advertisers. One special kind of non-article activity involves preparation of issue covers and another requires preparation of tables-of-contents for issues. Reproduction activities include printing, collating, and binding. Publishers often print more copies of issues than required by circulation in order to have back issues available for sale. They also will print individual copies of articles for authors or for sales of reprints. Distribution includes wrapping, labeling, sorting (by ZIP code), and actual mailing. One important activity involves subscription maintenance including keeping an updated list of subscribers, their addresses, and their payment status.
Support activities (sometimes called conducting business, general and administration, indirect costs) include:
These support activities and costs represent an appreciable portion of total publishing costs. They sometimes increase as a proportion of total direct costs (i.e., of the other four components).
3. Sources of Publishing Cost Data
We have gathered together several sources of traditional print scholarly publishing cost data. We went back to three studies sponsored by the National Science Foundation (NSF) in the 1970s because they appear to be the most recent comprehensive studies performed on scholarly publishing costs. Two of the studies [13, 14] involved surveys of scholarly publishers performed in 1975 and 1973 respectively. These surveys obtained information on journal (and book) publishing costs, circulation, price, and financial aspects. While providing very useful data and insights, the two studies suffered from low response rates. Presentations of results also were limiting because amount of costs attributed to publishing activities such as editing, composition, printing, mailing, etc. were presented as proportions (%) of total direct costs. The problem with this is that total direct costs involve large fixed ("first copy") costs and small variable (printing and distribution) costs per subscription. The proportions attributable to fixed or variable costs depend a great deal on number of subscriptions. That is, the proportion of all fixed cost activities (e.g., editing, composition, overhead, etc.) of a low subscription journal could sum to 90 percent, but for a large subscription journal it could be as low as 10 percent. Thus, presentation of proportions does not help one interpret relative costs unless more information is provided concerning total costs, circulation amount, size of journals and so on. Fortunately, Machlup & Leeson  provide some of these data to accompany the proportions.
The third NSF-sponsored study of journal publishing costs involved a series of studies (summarized by King, et. al. ). The approach taken in these studies was to visit three publishers and their presses to establish how cost resources (i.e., staff, equipment, etc.) might be allocated to specific publishing activities (15 in all). This approach led to detailed cost models and estimates of publishing costs among fields of science and over time based on trends in costs of resources (i.e., labor, paper, etc.) and cost parameters (i.e., number of articles/pages, issues, subscriptions, etc.). Extension and modification of these models are discussed below involving the five categories of journal publishing costs.
While we are unaware of more recent comprehensive publishing studies, some articles and book chapters have presented typical publishing costs of some activities. These include 1996 cost per page of paper and parallel paper and electronic , 1995 ranges of cost data for various American Chemical Society publishing activities  and 1983 total cost budgets for several American Institute of Physics journals  along with total pages and subscriptions for these journals. Marshall  provides an example of a commercial journal cost for the years 1988 and 1979 and Page et al.  provide similar data for the later years. Some scientific activity cost data are reported for editing [17, 18]; internal and external review ; composition ; and copy editing . Finally, we obtained some 1995 vendor cost quotes of printing, binding and mailing and postal rates from the U.S. Postal Service. Cost data presented in past years are all updated by inflation (CPI) to 1995 to provide some semblance for comparison. Some costs related to electronic publishing and storage are provided by Garson  and Bowen  respectively.
4. Cost of Article Processing
Article processing involves all publishing activities leading up to the "first copy" from which copies are ready for reproduction (at least in traditional print journals) and distribution. These "prerun" activities must be performed regardless of how article copies are distributed (i.e., electronic or paper). Since costs are largely a function of number of articles and their length, the principal parameters for article processing costs are number of articles published in a year, the average number of pages per article, number of special graphic pages produced, etc., and number of issues involved (because each issue requires some effort to "put the issues to bed").
Article processing costs are sometimes called fixed costs because the cost parameters and elements should not vary with journal output (i.e., number of subscriptions). However, there is some evidence that average editing and other related costs are less for small subscription journals than large ones. We also believe that average cost per page is less for non-article text than for articles because less effort is necessary for review and subject editing. Thus, the cost of processing non-article material is covered separately.
A rough cost model for article processing is as follows:
CA=C1I + C2M + (C3 + C4)APA + C5G
where CA is total annual article processing costs, C1 is the fixed direct cost per issue, C2 is the "setup" cost of receiving, processing and reviewing a manuscript, C3 includes costs associated with editing and proofing articles, C4 is the composition/typesetting costs, and C5 is the cost of processing special graphics and other non-text materials. Parameters of the model include I- number of issues, A- number of articles, PA- average number of pages per article, and G- number of special graphics and other materials.
With the observed model parameter and derived costs we arrive at "typical" total publishing cost of article processing as follows:
CA= $500 (8.3) + $230 (205) + ($50 + $70) (123)(11.7) + $60 (260)
or about $1950 per article published or about $165 per article page published. We emphasize that costs and model parameters vary dramatically from journal to journal and publisher to publisher.
5. Non-article Processing Costs
Scientific scholarly journals are estimated to have an average of 289 non-article pages published annually. The corresponding cost is not believed to be as high as article processing. A rough cost model for non-article pages is as follows:
CN = (C17 + C18)I + (C19 + C20) PN
where CN is the total annual cost of processing non-article pages, C17 is the "first" cost of handling the non-article materials for each issue, C18 is the cost of preparing covers for each issue, C19 is the cost of editing and proofing these pages, and C20 is the cost of composition/typesetting. There may also be special graphics costs, but these are ignored here. I is 8.3 issues per year and PN is289 pages published annually. In the absence of evidence we assume C17 plus C18 to be $50 and $200 or about $250 per issue, C19 to be ½ of article costs (i.e., $25 per page) and C20 to be the same as article pages (i.e., $70 per page). Thus, total non-article processing costs with these assumptions would be about:
CN = $250 (8.3) + ($25 + $70)(289) = $29,530
or $102 per non article page (compared with $165 per article page).
6. Reproduction Costs
Once a master image is produced a traditional paper journal can be reproduced. Reproduction involves printing (from a composed image) and binding. These two activities involve labor, equipment, paper, and other supporting resources such as space. The reproduction cost model is as follows:
CR = I[C6 + (C7 x PI) + (C8 + C9)S + (C10 + C11)PI x S]
where CR is the annual print and binding cost; C6 is the setup cost associated with one issue; C7 is the platemaking and/or collating cost per page; C8 is binding cost per copy; C9 is the cost per copy of special covers; C10 is the labor and equipment cost per impression (one copy of one page); C11 is the paper cost per impression (i.e., twice one piece); I is the number of annual issues; PI is the number of pages in an issue; and S is the number of copies made for an issue (number of subscriptions plus some overrun). Instead of multiplying by I one could more accurately sum the equation (without I) for each issue over a year. One could also simplify the model by ignoring the parameter number of issues (I) by adding a constant to the equation (I xC6) and using annual number of pages and impressions.
A typical 1995 scientific scholarly journal  has 8.3 issues, 5,800 subscribers (plus 200 copies for overrun), and 1,723 pages (or 208 pages per issues). Thus, one issue has an average of 1,248,000 impressions. We estimate C6 (one issue setup cost) to be about $980 per issue. C7 is roughly estimated to be about $6 per page (one issue). C8, the binding cost per copy, depends on the type of binding (e.g., saddle stitch, perfect binding, etc.), but are observed to be about $.015 per copy. Special covers appear to cost about $0.15 per copy (C9) but can be much higher. Total cost per impression (C10 + C11) comes to roughly $0.00095. These costs depend on size of journals (i.e., number of pages) and number of copies (i.e. number of subscriptions). They also depend on other factors such as quality of paper. Applying the parameters and elements above, a typical journal would have the following reproduction cost:
CR = 8.3[$980 + ($6 x 208) + ($0.15 + $0.15)6,000 + (0.0095) 208 x 6,000]
or $23 per subscription.
7. Distribution Costs
The distribution cost includes the same three parameters as reproduction costs (i.e., number of issues, number of subscriptions and number of pages per issue). The cost model is CD = C12I + (C13 + C14) SI + C15S + C16SIPI where CD is the total direct distribution cost, C12 is the fixed cost associated with mail processing, C13 is the mail processing cost per item or copy (including inserting, sorting, sealing, labeling, etc.), C14 is the postal cost per copy rate (which is dependent on proportion of advertising pages), C15 is the subscription maintenance cost per subscription, and C16 is the postage cost per page per copy mailed (also based on proportion of weight due to advertising). Again, the typical scholarly scientific journal parameters are 8.3 issues, 5,800 subscriptions, and 208 pages per issue. Typical costs appear to be about $50 per issue (i.e., mailing) for C12; C13 = $0.35 per copy mailed; C14 is $0.263 (assuming 5% advertising pages); C15 is about $7.00 per subscription; and C16 is about $0.001 per page per copy mailed (assuming 5% advertising pages). Thus, the total distribution cost of a current journal cost parameters and elements above is as follows:
CD = $50 x 8.3 + ($.035 + $0.263) 5,800 x 8.3 + $7 x 5,800 + $0.001 x 5,800 x 8.3 x 208 = $80,538
or roughly $14 per subscription. There are a number of factors that affect these costs such as whether journal copies are wrapped, how heavy the pages are (including special covers) and whether a mailer has a non-profit designation for preferred postal rates.
8. Publishing Support Costs
Publishing support involves non-direct cost elements such as general and administration (G & A) costs, space and utilities, taxes, and so on. It also includes marketing, cost of research, and reinvestment. We assume that support costs (excluding finance) are about 28 percent of article processing and non-article processing costs (or about 40% of direct costs) since these costs are labor intensive and at 21 percent of reproduction and distribution costs since these are often contracted out. Thus total costs of publishing support would be:
CS = 0.40 (CA + CN) + 0.21 (CR + CD)
where CS is the costs attributable to support activities, CA is total article processing costs, CN is total non-article processing costs, CR is the total reproduction costs and CD is total distribution costs. As an example,
CS = 0.40 ($239,592 + $29,530) + 0.21 ($131,837 + $80,538) = $152,248
Thus, publishing support costs represent an appreciable proportion of all direct costs.
9. Total and Unit Publishing Costs
Based on the approximate costs of the four functions or group of publishing activities the total costs of publishing a typical scientific scholarly journal is represented by the following equation:
CP = CA + CN + CR + CD + CS
= $239,592 + $29,530 + $131,837 + $80,538 + $152,248
If CA and CN are considered fixed costs and CR and CD variable costs with support costs allocated (at 40% and 21%), we find that the minimum price necessary to cover publishing costs for various circulation levels would be as shown in Figure 1.
The price necessary to recover costs at 500 subscribers is over $847 per subscription, but decreases to $120 at 5,000 subscriptions. At 50,000 subscriptions the necessary price drops to about $48 per subscription which approaches the incremental reproduction and distribution cost of about $40 per subscription. At 500,000 subscribers the price is less than$1 above the incremental cost.
The median number of subscriptions is about 1,900 and the price necessary to cover costs is $252. Thus, one would expect one-half of the journals to have a subscription price over $252. The average subscription price is $258 per journal (and much less averaged by subscriptions). It is emphasized that the costs above and model parameters are merely indicative of what a journal might cost. Below (Table 1) we vary some of the cost parameters and unit costs to test sensitivity of total costs to them. The cost parameters are primarily number of subscriptions, number of issues, number of article and non-article pages, and number of special graphics pages. The average scientific scholarly journal is estimated (1995) to have the following parameter values: 5,800 subscribers; 8.3 issues per year; 1,434 article pages and 289 non-article pages; and 260 special pages. We varied number of subscribers above. Below we vary number of articles (with pages per article remaining constant):
Table 1 Model Prediction of Scholarly Journal Total Costs and Cost per Subscriber by Number of Articles Published: 1995 Dollar Values
|No. of Articles||Total Cost||Cost Per Subscriber|
Thus, increasing article number (and other related parameters) by increments of 50 articles results in an increase in cost per subscription of about $36 per subscription, or about $0.72 per article per subscriber. Increasing the number of articles four-fold from 50 to 200 results in an increase in cost per subscriber of only 2.9 times ($57 to $164). This is one of the reasons that publishers tend to increase the size of journals rather than publishing more journals.
Over 20 years (1975-1995), the number of scientific scholarly journals published in the U.S. has increased at a rate far less than growth of number of scientists and engineers , but number of articles published per scientist or engineer in the U.S. had remained relatively constant because the average size and frequency of publishing has increased substantially. From 1975 to 1995, the number of scientists and engineers in the U.S. had increased from about 2.6 million to 5.7 million (an increase of about 119%). The number of scientific scholarly journals is estimated to have increased from 4,170 to 6,771 (an increase of 62%). Referring to cost model parameters the changes in size and frequency of publishing are as follows:
Table 2 Scholarly Journal Publishing Parameters: 1975 and 1995
|Cost Model Parameter||
|No. of Issues||
|No. of Articles/Title||
|No. of Manuscripts Submitted||
|No. of Article Pages||
|No. of Special Graphics||
|No. of All Pages||
|No. of Subscriptions||
Sources: King, et al. 1981, Tenopir & King 1997.
Applying 1995 cost estimates in the cost models to the 1975 parameters, we get a total cost of $334,180 per journal. Below we compare 1975 and 1995 size journals in terms of total cost and unit cost.
Table 3 Scholarly Journal Model Costs for Journals Published in 1975 and 1995: 1995 Dollar Values
|Total and Unit Costs||
|Cost per Issue||
|Cost per Article||
|Cost per Page||
|Cost per Subscriber||
Thus, it may be that increasing the size of journals as opposed to number of journals has kept costs down at least on a per cost basis. Unfortunately looked at from the buyers perspective, the prices have increased far more than one might expect considering increased size of the journals (i.e., information provided). Even with cost elements given for both years in 1995 dollars, the cost per subscription increases from $58 to $109 (even though average number of subscribers is nearly equal.)
10. Cost of Electronic Publishing
Electronic publishing of scientific scholarly journals is likely to follow three paths: electronic publishing only in which the journals replicate processes and articles similar to traditional paper journals; journals that provide subscriptions in both paper and electronic media (with electronic access also to single articles on demand); and journal publications that depart from traditional journal formats and processes and/or provide value-added features to traditional journals.
Totally electronic journals save the cost of reproduction and distribution (about $40 per subscription), some costs associated with paper issues (about $550 per issue or $4,150 in a typical journal), and non-article processing of issue covers and other information ($250 per issue or $2,075). For a typical journal the fixed direct costs would decrease by about $6,225 or about two percent of the total. These savings are partially offset with electronic storage, software and typically higher labor costs. Even ignoring the additional electronic-related costs, the savings to subscribers is not all that great. For example, the cost difference for journals having 500 subscribers is about $66 per subscription or at most about 7.8 percent of the total cost per subscription. With journals having 5,000 subscribers the cost difference is about $50 per subscription or at most about 41 percent of the total cost per subscription.
Parallel paper and electronic publications include those in which only full subscriptions are available in both media, those in which electronic access is only for separate copies of articles, and those in which electronic access is through subscriptions and separate copies of articles. In all instances, the cost of parallel publications is likely to be higher because of additional (and more expensive) resources required for the added electronic access. Electronic costs also include additional coding and tagging required for electronic access. However, the additional costs appear to be in the 3% to 8% range [e.g., 7, 22]. The excess costs should be easily recovered through access of separate copies over the long run (in lieu of interlibrary borrowing and document delivery). However, pricing the separate copies will be particularly important concerning how much revenue is generated.
The wide range of value-added processes that may be offered with totally electronic journals will cost more. To begin with, publishers will be able to provide a database of journals, single journals, individual articles, or parts of articles. Various levels of information can be made available on examination, including titles, abstracts, reviews of the article, accompanying data, appendices, and so on. Sets of articles can be sent automatically to users, based on profiles of readers' interests. Quality of older articles can be rated by citation counts of authors (before or after publication), ratings made by readers, or ratings made by a panel of referees. Multimedia (including sound, motion, extended graphics, etc.) and interactivity (between readers and data, readers and readers, readers and authors, etc.) are all features added to electronic-only journals. Since any changes will affect costs, information and service attributes, and use, pricing strategies must be established for each.
This study was funded in part by a Special Libraries Association Steven A. Goldspiel Research Grant. The authors would like to thank Danielle Green for her assistance.
 C. Tenopir. "Authors and readers: the keys to success or failure for electronic publishing." Library Trends, vol. 43 no. 4, pp 571-591, Spring 1995.
 D. W. King and J. M. Griffiths. "Economic issues concerning electronic publishing and distribution of scholarly articles." Library Trends, vol. 43 no. 4, pp. 713-40, Spring 1995.
 D. W. King, D. D. McDonald and N. K. Roderer. Scientific Journals in the United States: Their Production, Use and Economics. Hutchinson Ross Publishing (Academic Press), 1981. Available from D. W. King.
 C. Tenopir and D. W. King. "Trends in scientific scholarly journal publishing in the United States." Journal of Scholarly Publishing, vol. 28 no. 3, pp. 135-170, April 1997.
 C. Tenopir and D. W. King, The Transformation of Scholarly Journals: 20 Year Trends in the Economics, Use and Information Seeking Patterns of Scientific Print and Electronic Scholarly Journals. Washington, DC: Special Libraries Association, 1998 (in progress).
 J. M. Griffiths and D. W. King. Special Libraries: Increasing the Information Edge. Washington, DC: Special Libraries Association, 1993.
 A. Holmes. "Electronic publishing in science: reality check." Canadian Journal of Communication, vol. 22, no. 3/4 pp. 104-116, 1997.
 Robert Ubell Assoc. Cost Centers and Measures in the Networked Information Value Chain. Washington, D.C.: Center for Networked Information, 1996.
 R. H. Marks. "The economic challenges of publishing electronic journals." Serials Review: Economics of Electronic Publishing, Spring 1995, pp. 85-88.
 L. Scovill. Librarians and Publishers in the Scholarly Information Process: Transition in the Electronic Age. New York: Association of American Publishers, Inc., 1995.
 G. Page, R. Campbell and J. Meadows. Journal Publishing: Principles and Practice. London: Butterworths, 1996.
 D. W. King and N. K. Roderer. Systems Analysis of Scientific and Technical Communication in the United States: The Electronic Alternative to Communication Through Paper-Based Journals.
 F. Machlup and K. W. Leeson. Information Through the Printed Word: The Dissemination of Scholarly, Scientific and Intellectual Knowledge: Vol. II Journals. New York: Praeger, 1978.
 B. M. Fry and H. S. White. Publishers and Libraries: A Study of Scholarly and Research Journals. Lexington, Mass.: Lexington Books, 1976.
 R. Lerner. "The professional society in a changing world." Library Quarterly, vol. 54, no. 1, pp.36-47, 1984.
 G. Marshall. "The economics of journal publishing: a case study." In Serials Information from Publisher to User: Practice, Programs, and Progress. Ed. by L. A. Chatterton and M. E. Clack. New York: The Haworth Press, 1988.
 R. A. Day. "Economics of printing." In Economics of Scientific Publications. Washington, D.C.: Council of Biological Editors, 1973.
 J. H. Fisher. "The true costs of an electronic journal." Serials Review: Economics of Electronic Publishing. Ed. by E. F. Duranceau. Spring 1995, pp. 88-90.
 D. H. M. Bowen. "Costs in selecting manuscripts." Scholarly Publishing. Oct. 1979, pp. 43-46..
 J. Lago. "A decade of electronic editing." Scholarly Publishing. Jan 1993, pp. 101-112.
 M. Brogan. "Costs in copy editing." Scholarly Publishing. Oct. 1979, pp. 47-53.
 L. R. Garson. "Can e-journals save us? A publishers view." in The Economics of Information in the Networked Environment. Ed. by M. A. Butler and B. R. Kingma. Washington, D.C.: Association of Research Libraries, 1996, pp. 115-121.
 W. G. Bowen. "JSTOR and the
economics of scholarly communication." The Economics of
Information in the Networked Environment. Ed. by M. A. Butler
and B. R. Kingma. Washington, D.C.: Association of Research
|Last updated : 30 September 1998
Copyright © 1998 ICSU Press and individual authors. All rights reserved.