Is there such a thing as 'Least Publishable Unit'? An empirical investigation

INTRODUCTION"Academic publishing has already reached a point where too much material of too little substance is being published, and this trend is continuing" (Trimble, Grody, McKelvey, & Gad-el-Hak, 2010, p. 276). The foregoing claim is one that begs for both clarification and examination. We can take it almost as a hypothesis to be tested according to certain strictures. Before engaging in the investigation, we should note that the claim is not isolated. Goldsborough (2009) observed, "as the level of information input increases, our capacity to process and retain that information decreases" (p. 13). With these two statements we have claims about the body of literature and about people's abilities to absorb and make sense of it all. Additionally, there are cognitive and psychological-as well as time-limits to the development of a full understanding of the scholarly and research literature that exists. The problem is exacerbated by the annual growths of literature in every discipline. To refer to Trimble et al. (2010) again, "In short, we find the present to be inefficient, irrational, unfair, outrageously expensive, and environmentally irresponsible-in two words, unscholarly and unethical" (p. 280). Academic libraries need to become aware of the phenomenon so that services for researchers can be as completely informed as possible. Scholarly communication in general is affected by the practice. While we shall not address expense, the other topics will feature in the following analysis.LEAST PUBLISHABLE UNIT (LPU)There is a phenomenon that is said to typify publication in the sciences called "Least Publishable Unit" (LPU). William Broad (1981) said, "LPU is a euphemism in some circles for the fragmentation of data. A researcher publishes four short papers rather than one long one" (p. 1137). Broad and his colleague Wade (1982) included the phenomenon in a book they wrote. If they are correct, the literature of a scientific field or sub-field will be inflated. In fact, if, say, 10,000 scientists were to behave as Broad and Wade indicated, a literature of perhaps 20,000 to 30,000 papers published in a year might swell to 80,000 to 120,000 articles. Readers of the literature will then have to sift through a potentially massive number of articles in order to locate pertinent works. The amount of time required to make effective use of what is published likewise increases. The import of the effects of LPU expands beyond the individual. There are many drawbacks to rewarding large numbers of publications. Buddemeier (1981) offered, "Perhaps the most readily implemented suggestion would be to abandon the total bibliography as an evaluation tool and require applicant [for grants] to list some specific number of their most significant publications and provide a brief narrative outlining the importance of the work and the nature of the applicant's contribution" (p. 494). At this point in time, the applicants' curriculum vitae are limited in length, so the emphasis on quality should hold. "The practices of fragmented and repetitive publication are wasteful and inefficient" (Angell, 1986, p. 261). Huth (1986) added that, "The abuse of divided publication is the breaking down of findings in a single string of papers. . . . The research could probably be reported in a single paper. But why should the investigators confine themselves to one paper when they can slice up data and interpretations into two, three, four, five, or more papers" (p. 257).The criticism of LPU continues across disciplines and time. Yankauer (1990) wrote, "The 'least publishable unit' (LUP)-sometimes called salami science-is perhaps the most vexing and most difficult problem to control. It arises because of pressures to publish and the institutional promotion and tenure procedures that rely on publication quantity rather than quality" (p. 400). Gleeson and Biddle (2000) pointed to another problem related to LPU: "The main worry is that duplicate, fragmented or piecemeal publications is misleading if they appear to represent independent observations. …