Many scientists today have ready access to services like SciFinder and Reaxys to help them find and evaluate scientific papers and chemical information relevant to their research. It wasn’t always so easy, but beginning in the 1950s, Eugene Garfield helped make it so.
In those days, when it often could take more than a year to get a scientific paper published, Garfield started creating information tools such as the Science Citation Index that made it possible to learn about and analyze scientific ideas much more rapidly than had been possible before.
Garfield, who died last year, founded the Institute for Scientific Information (ISI), now part of Clarivate Analytics, to produce those tools. At the American Chemical Society national meeting in New Orleans last month, speakers discussed his life and innovative scientific contributions during a session sponsored by the Division of Chemical Information.
Garfield was born in 1925 in New York City and earned a bachelor’s degree in chemistry in 1948 at Columbia University. He began his career as a bench chemist in the Columbia University lab of physical organic chemistry pioneer Louis Hammett.
But a presentation he saw at a 1951 New York City ACS meeting, on the use of IBM punch cards to encode chemical information, set him on a new path. Later that year, he signed on as a research assistant at Johns Hopkins University’s Welch Medical Library to help investigate the potential of computer-based information services.
In his two years there, Garfield developed three concepts that would serve as the foundation for his life’s work: listing publications’ contents so they could be easily searched with punch cards; developing searchable indexes of chemical compounds; and using the literature cited by a publication to track the web of ideas connecting present, past, and ultimately future scientific studies.
It took him only about a decade to turn all three concepts into realities. He created Current Contents, which compiled the tables of contents from hundreds of scientific journals; Index Chemicus, which reported on newly published compounds; and the Science Citation Index, which cataloged literature references.
The total number of articles now cataloged in the current “expanded” online version of the Science Citation Index.
The number of scientific journals assigned impact factors in Journal Citation Reports in 2016. JCR started with 2,630 journals in 1975.
After leaving the Welch project and getting a degree in library science at Columbia in 1954, he started Eugene Garfield Associates, later renamed DocuMation and then ISI. In 1958, the company started publishing Current Contents.
In 1955 while continuing to run the company, Garfield became a Ph.D. student in structural linguistics at the University of Pennsylvania, where he developed procedures to translate the names of chemical compounds into molecular formulas a computer could read and index. In 1960, he put this technique into practice by initiating and publishing Index Chemicus.
“Index Chemicus began as a current awareness service on new compounds, with molecular formulas, structural diagrams, and codes representing compound fragments,” said information scientist and consultant Wendy Anne Warr at the New Orleans symposium. “Wiswesser Line Notation, text sequences that denote specific chemical structures to enhance computer searching, were added later.” Nowadays, graphics-based structure-search systems in services like CAS’s SciFinder have largely replaced text code searches, although fragment codes and line notations have not disappeared, Warr said. CAS is a division of ACS, which publishes C&EN.
And in 1963, Garfield developed the Science Citation Index to make it easier to identify highly cited scientific papers. At the time, “the literature of science was often organized by traditional subject indexing,” James Testa, vice president emeritus of editorial development at Clarivate Analytics said during the meeting. “An expert indexer would examine each paper and assign it to one or more subject categories. Each of these judgments takes time and bears a cost. This method is also limited by the expertise and skill of the indexer.”
The Science Citation Index “effectively replaced and amplified the indexer’s expertise with reference lists created by the authors themselves,” Testa said. The index “made visible each article’s association with prior articles” and, in time, made clear how that article influenced future lines of research. It not only allowed researchers ready access to the universe of scholarly research, he added, but also enabled them “to extract and make visible the order inherent in it.”
The Science Citation Index “is an effective information retrieval tool,” said chemistry librarian Svetla Baykoucheva of the University of Maryland, College Park. By making it easier to find out what others are doing, it allows scientists to avoid duplication of research effort, boosts interdisciplinary research and connections, and sparks new ideas through unexpected associations, she said.
A key offshoot of Science Citation Index was Journal Citation Reports, which in 1975 began reporting journal impact factors, a measure of the relative importance of different journals. Garfield and a colleague developed impact factors to select additional journals to include in the Science Citation Index. Impact factors later gained widespread independent influence, but they have also generated controversy.
An impact factor is the number of citations in Science Citation Index journals in one year pointing to articles published in a specific journal in the preceding two years, all divided by the total number of citable articles published in that specific journal in the preceding two years. “Citable articles” has a specific definition that does not include editorials and letters, for example.
“Who cares about impact factors?” Baykoucheva asked rhetorically during the session. Despite the controversy they’ve stirred up regarding their validity and use, she said, they are important to authors who want to publish in high-ranking journals, editors who want to track their journals’ rise and fall in the ratings, librarians who must decide what journals to keep or cut, and administrators evaluating grant proposals, among others.
Problems with impact factors include editors using tricks to inflate their journal’s rating, researchers citing themselves too much, and scientists colluding to cite each other frequently.
But perhaps the most controversial issue for impact factors is their use as proxies to evaluate researchers for hiring, promotion, and awards or to make science policy and research funding decisions, Baykoucheva said. For example, rewarding a researcher for publishing in Nature, a journal with a high impact factor, might be inappropriate because the paper might never be cited.
An impact factor measures the influence of a journal, not that of an individual paper or researcher, Baykoucheva continued. Garfield “was very vocal against using it for science policy or to evaluate individual scientists,” she said. Alternative metrics, or altmetrics, such as the number of times papers are downloaded from journal websites, are currently growing in importance as measures of the significance of journals and individual papers.
In 1986, Garfield capped a career of already impressive accomplishments by creating and publishing The Scientist, a newspaper in which “issues that were really important for doing research could be heard and addressed,” said former ISI executive vice president for database publishing Bonnie Lawlor.
Lawlor noted that Garfield “was not a bureaucratic manager” at ISI. “The place was crazy. People parked their motorcycles at their desk. There was no dress code. One person used to love to wear baby-doll pajamas to work, one of the senior directors had a sapphire on his forehead and a teddy bear on his belt, and those two people, Baby Doll and Sparkles, streaked at one of the company events. But beneath that surface, it was a very energizing, intellectually challenging environment. Garfield encouraged every one of us to do the best with the talents God gave us and to contribute what we could for the betterment of the company.”
On the other hand, Testa said, “100% quality was what he demanded, and when you’re dealing with billions of bits and bytes, try it”—meaning that it’s not easy.
When Garfield died in February 2017, “an era did end, no question about it,” Lawlor said. “But his legacy lives on.” For example, she noted that Garfield’s citation concept influenced the development of PageRank, an algorithm Google now uses to produce search engine results. “There are hundreds of people who have jobs because of him, and there are thousands of researchers around the world who use his tools every day.”