Decade-Long Legal Battle Focused On Journal Cost, Impact

Editor's Note: For the past two decades, librarians have used citation data to rank journals. When library budgets were seemingly unlimited, journals published both by nonprofit professional societies and by for-profit publishers were purchased, often regardless of their intrinsic value or cost-effectiveness. In the past decade, under severe budget constraints, university librarians-with the advice of faculty-have had to "de-accession" hundreds of journals that, by one test or another, could no longer be justified. As an extension of the quantitative methods of evaluation originally reported in my "Significant Journals of Science" (Nature, 264:60915, 1976), it was inevitable that peer judgments of journal quality would be combined with bibliometric indicators, such as the number of articles published (productivity) and average frequency of citation (impact). In 1988, Henry H. Barschall-a retired professor of physics at the University of Wisconsin, Madison, and a member of the American Institute of Physics' (AIP's) Publishing Policy Committee-published his now well-known study of physics journals (Physics Today, 41[7]:56-9, 1988). He combined subscription prices and journal impact data to produce an indicator of cost-effectiveness. To my knowledge, there has not been a comparable study in other branches of science. However, it is important to note that many scientists and librarians do not support the use of impact data. Today it is possible and more relevant to prepare a customized analysis of an institution's use of journals by using citation indexes to learn what journals its faculty members actually have cited in their current publications. These analyses are often supplemented with usage studies indicating what students and faculty read but do not necessarily cite. Shortly after Barschall published his study, Gordon & Breach Science Publishers S.A. filed suits in European courts claiming unfair competition. When this failed, Gordon & Breach filed a legal action in the United States Second District Court alleging unfair competition. The legal dispute between Gordon & Breach and AIP, the defendant society, dragged on for a decade. Last August, Judge Leonard B. Sands rendered the following conclusion: "Barschall's methodology has been demonstrated to establish reliably precisely the proposition for which defendants cited it-that defendants' physics journals, as measured by cost per character and by cost per character divided by impact factor, are substantially more cost-effective than those published by plaintiffs. Plaintiffs have proved only the unremarkable proposition that a librarian would be ill-advised to rely on Barschall's study to the exclusion of all other considerations in making purchasing decisions. This consideration in no way makes Barschall's study or defendants' descriptions thereof false." The Web site http://barschall.stanford.edu provides a good precis and history of the case. Since most scientists are not familiar with the case, we accepted an offer from Albert Henderson, an editor who worked for Gordon & Breach, to condense his view, followed by a rebuttal from Marc Brodsky, AIP's executive director. -Eugene Garfield .-Eugene Garfield By Albert Henderson Advertising executives know that science makes a sales pitch more persuasive. So when Henry H. Barschall devised a scientific-looking "cost-effectiveness analysis" to promote American Institute of Physics (AIP) journals, its managers did virtual cartwheels. Librarians gave him awards. The Association of Research Libraries (ARL), whose members dominate American sales of research journals, expanded on his claims in its Report of the ARL Serials Prices Project (Washington, D.C., Association of Research Libraries, 1989). His "findings" were boosted by an editorial in Science (P.H. Abelson, 244:1125-1989). His proponents continue to defend the work, which ranks physics publishers according to average unit prices and impact factors, and describe it as "scholarly" (APS News, January 1996). Writs against unfair competition were filed in Europe and the United States by Swiss-based Gordon & Breach Science Publishers, which named Barschall, the American Physical Society (APS), and AIP as defendants. In the United States, a federal judge recently decided that they had not violated prohibitions against false advertising largely because AIP had destroyed its promotional material before it could be mailed (Judge L.B. Sand, SDNY 93 Civ. 6656 1997). The litigation also disclosed departures from scientific custom that were ethical violations not considered by the courts. A tenet of research integrity is full disclosure of relevant influences, such as conflicts of interest and sources. Barschall's judgment was clouded by a conflict of interest shared by his publishers. Barschall presented himself as a physics professor interested in solving the library crisis. His solution asked readers to favor the "low-cost-per-character" journals of AIP and its members, conclusions relayed to librarians two months before the articles were published. He adopted the idiom of librarians, using the word "cost" to mean the nonmember price of a journal. He failed to mention his commercial objectives as a member of the AIP Board of Governors, the directorate responsible for its $42 million publishing business. He also neglected to acknowledge the contribution of the societies' managers. For example, in a letter dated Sept. 24, 1987, AIP director of publishing R.H. Marks suggested to Barschall, "It would be a good idea to include information about the size, total number of pages published as well as the annual subscription price. Most of our competition price their product by the volume with no guarantee on the number of volumes per year." These influences were forcibly brought to light under authority of the court. Ken Ford, executive director of AIP, and Harry Lustig, treasurer of APS, compounded the ethical fault in CBE Views (16[1] 8-11; 16[4]:81-2, 1993). They echoed affidavits sworn before European courts: "Neither the APS nor the AIP has ever 'supported' the preparation of the studies: We have published certain unsolicited manuscripts concerning the serials crisis." The court heard testimony that AIP wanted to increase market share and profitability. They aimed to distract librarians, who controlled 70 percent of AIP's income, away from their own planned price increases. Bias affected Barschall's work. Logic would dictate that he match impact factors for 1984-85 sources with easy-to-calculate unit prices for the same period. Instead, he used elusive 1987 prices and kiloword data. In a letter dated Sept. 8, 1987, he complained to the APS treasurer, "I don't know yet how many pages each journal will publish so that I had to extrapolate. A more serious obstacle [to his assault on competitors] is the fact that the library is following my advice and canceled the journals that charge the most per character." He must have found examples here because his data appeared to contradict his stated sampling parameters ("all the physics journals to which there was easy access"). His report also excluded examples that would interfere with promotional aims, such as AIP's translation journals. Dividing his report into "least publishable units" dodged comparisons. Data tables were buried among ephemeral meeting abstracts in the APS Bulletin (H.H. Barschall, J.R. Arrington, 33:1437-47, 1988). His summary appeared in Physics Today offering the key "Table 2," which ranked publishers according to average unit price. It put AIP and its members in five of the top six positions (H.H. Barschall, 41[7]:56-9, 1988). It misrepresented data in the Bulletin. The novelty that set Barschall's formula apart from other surveys enhanced unit prices with journal impact factors, many of which were not available. He claimed journal impact factors enabled him to analyze not only "cost" but also "cost-effectiveness." To sustain his theory and present cost-to-impact ratio averages for publishers, he invented them. He also made other errors, perhaps intentionally. Take, for instance, Barschall's handling of the Italian physical society, which issues four journals: G. Fis. and Nuovo Cimento A, B, and C. Only three of them have impact factors. In his Physics Today summary, he claimed the average ratio of cost to impact to be 12.4 for the four journals, ignoring the absence of data for G. Fis. The correct summary would admit that an average for the four is not available. He misrepresented nine other competitors with price-to-impact averages that were similarly impossible. Barschall also misrepresented prices. He estimated the cost-per-1,000-characters of the four Italian journals at 2.5 cents , 9 cents, 7.4 cents, and 5.5 cents, respectively, in the Bulletin. In Physics Today, he claimed the average to be 7.9 cents, much higher than the correct figure of 6.1 cents. This error moved this unfortunate competitor from the 14th to the 21st place in a field of 24. His averaging of five other publishers' unit prices was also incorrect. For three of those, the error resulted from segregating translation journals, "because they have a quite different cost basis." An impartial application of this principle should exclude all examples of cost differences affecting pricing (such as currency fluctuations, foreign postage, tax exemptions, payments to authors and editors, and author subsidies). Eliminating his selective reporting moves AIP's average unit price from 1.9 cents to 5.4 cents and its rank from the fifth to the 13th, next to where he should have placed the Italian physical society. The validity of Barschall's claims about solving the serials crisis was also not considered by the courts. His theory had no foundation in economics or information science. Donald W. King, former president of the American Society for Information Science and consultant to AIP's long-range planners, testified that niche journals serve a limited audience. Their fixed costs are borne by a smaller circulation, reasonably requiring higher unit prices. According to King, a librarian relying only on a Barschall-type study would make "some horrendous mistakes." The court's opinion also noted that AIP and APS "acknowledge that Barschall's analysis does not demonstrate abstract product superiority or quality." Judge Sand wrote, "a librarian would be ill-advised to rely on Barschall's study to the exclusion of all other considerations in making purchasing decisions," while favoring AIP in his decision. Barschall's provocatively presented "finding" that high and low unit prices differed by a factor of 80 was neither a new nor an abnormal condition. It had been preceded by similar observations well before the library crisis in Report of the Task Group on the Economics of Primary Publication (Washington, D.C., National Academy Press, 1970). Instead of solving the crisis, AIP and others aggravated and exploited it. They leveraged their unique advantages into a spurious, anticompetitive standard. By unreasonably exalting society journals, they Balkanized research publishing into a period of "ours vs. theirs": U.S. associations vs. foreign and commercial publishers. The roots of the crisis are more likely to be found in arcane financial practices, such as research overhead formulas and administrative priorities, than in competitors' prices. Digest of Education Statistics indicates that constant dollar research spending doubled between 1972 and 1993, generating many articles, while library spending rose barely 35 percent (Tables 335, 338, 1995). Present ethical guidelines are unequivocal. "Any professional relationship or action that may result in a conflict of interest must be fully disclosed. When objectivity and effectiveness cannot be maintained the activity should be avoided or discontinued" (Physics Today, 45[1]:62, 1992). AIP and APS managers still defend Barschall's work in spite of the obvious. They continue to appeal the French injunction against "publishing in their journals articles which, in scientific guise, have as their goal the denigration of competing journals" (Tribunal de Grand Instance de Paris. 3d chamber, 1st section. Nov. 27, 1991). I challenge AIP and APS to explain why they have not sought a solution to the library crisis that would benefit everyone. The crisis hobbles researchers across the entire spectrum of disciplines. It seems to me that the advantages enjoyed by these nonprofit associations, such as tax exemptions and other subsidies, are meant to serve society. Their objectives, "the advancement and diffusion of the knowledge of physics," suggest that they should help other physics publishers and refrain from cut-throat commercialism. They can influence policy and accreditation. As "official" voices representing the physics community, they also have a duty to be fair and authentic and to provide models of professional conduct. Albert Henderson has worked for Gordon & Breach Science Publishers and the Foundation for International Scientific Cooperation. He also edits Publishing Research Quarterly. By Marc H. Brodsky Albert Henderson in his accompanying article accuses the late Henry H. Barschall of the American Physical Society (APS) and the American Institute of Physics (AIP) of improper motives, including making "errors, perhaps intentionally," when APS and AIP published information on the unit pricing and impact factors of about 200 physics journals. The information was developed by Professor Barschall and included data on 14 physics journals published by the Gordon & Breach publishing group (G&B). Unfortunately, rather than engage in open discourse about the issues, G&B has sued in four countries, causing enormous expense to the societies in defending the right to discuss an important issue to their members. Indeed, a United States District Court has noted the evidence of G&B's "aggressive corporate practice of challenging any adverse commentary upon its journals, primarily through threatened [and actual] litigation." It is ironic that an effort by the societies to examine the high cost to libraries of physics journals should lead to such unnecessary expense. As executive director of AIP, a member of APS, and one who had personal contact with the late Professor Barschall, I resent Henderson's innuendoes and accusations. I am proud of the services rendered to the members of our societies by Barschall's publications. APS, AIP, and Barschall's estate are still actively defending the openness, care, truthfulness, and relevance of Barschall's writings and the importance of having the freedom to discuss and analyze the data behind science policy issues, such as Barschall addressed. Henderson, who serves as G&B's consultant, seems inclined to pick at details, allege errors in the data or analysis, and assist G&B in its legal attacks. In fact, I welcome open discussion (but not litigation). In the wake of the recent trial before Judge Leonard B. Sand in U.S. District Court, the observers of this discourse now have a wealth of information about Henderson's claims and can form their own judgments about them. I refer them to the Web sites recently established by librarians at Stanford (http://barschall.stanford.edu) and Yale (http://www.library.yale.edu/barschall/), which have gathered the Barschall articles, the set of U.S. District Court opinions, and the trial transcript to which Henderson selectively refers. The reader will find refutations of Henderson's incredible claim that the court ruled the way it did because AIP "destroyed" material. Rather, the court ruled for AIP and APS because Barschall's articles and our use of them were neither false nor misleading, citing evidence "which conclusively established the reliability of Barschall's methodology." The quality of APS and AIP journals was never an issue of contention in the suit. Witnesses for both sides testified that APS and AIP publish the leading journals in their fields. The only disputed issue was whether statements about prices and cost effectiveness were justified by reference to Barschall's methodology and articles. The court unequivocally found for AIP and APS. Extensive trial testimony from both G&B and society witnesses supports Barschall's conclusions on relative costs and cost-effectiveness, independent of his specific analysis and selection of journals. Indeed, the court noted that "[G&B] at one point concede[s] that 'there is no dispute' that 'whatever variant of cost per citation or impact is used, the societies do very well and the commercial publisher do not.'" The court also noted what to me seems obvious, namely that "Plaintiffs' [G&B's] evidence that there are reasons why their journals cost more in no way makes false the statement that they do cost more." The editor of Physics Today, a publication of AIP, published two articles by Barschall (Physics Today, 39:34-6, 1986; Physics Today, 41:56-9, 1988) because she thought they addressed an important issue of concern to the physics community, namely the high cost of library physics journals. Because her editorial style precluded Barschall from publishing extensive backup data, APS published them in a companion article in the July 1988 issue of its Bulletin of the American Physical Society (33:1437, 1988). Therefore, extensive data and the articles themselves are in the published record (and on the aforementioned Web pages) for anyone to examine and critique. In closing, I take up Henderson's backhanded challenge to AIP "to explain why they [referring also to APS] have not sought a solution to the library crisis that would benefit everyone." It appears that Henderson intends by this that the "solution" should include benefits for his client. But AIP has taken the routes available to it to help hard-pressed libraries. Through publishing the Barschall articles we have tried to inform our members about the issue. Moreover, despite inflation in production costs, growing journal size, declining subscriptions, and extra costs for electronic publishing, we are struggling to contain our subscription price increases. In fact, we are trying to match our price increases more closely to library budget increases. This is not an easy task, but it is our goal and we think have had some success in recent years. Perhaps, as one witness from AIP indicated in a deposition in the U.S. District Court case, not raising prices as high as the market will bear will leave some money left over so the libraries can buy more expensive journals from others. In this way, even G&B benefits from our efforts. Marc H. Brodsky is executive director of the American Institute of Physics in College Park, Md.

Decade-Long Legal Battle Focused On Journal Cost, Impact

Interested in reading more?

Become a Member of