Publishing and literature search in the online era
I. Misztal
Abstract
In the distant past, animal breeders had to follow just a few journals to stay up to date. A convenient way to browse them was to have paper subscriptions and occasionally visit a good library. Picking up the latest issues at a library and browsing relevant sections was a pleasure. However, not everyone had subscriptions or access to a good library. If a title of a paper was known, a reprint could be requested. Having many requests was a sign to the respective author that a topic was “hot.” Paper publishing was expensive because of printing and mailing costs. As the world moved into the digital era, publishing became less expensive. New sets of journals popped up to exploit the online opportunity and promised faster publishing with open access, for example Frontiers and PlosOne. Because research institutions adopted “analytics” to measure faculty productivity by the number of published papers, “special” online journals appeared to allow the publishing of any content at a price. Existing journals adapted to the digital era by making their papers available online and allowing them to opt-out of print copies at a reduced cost; some existing journals moved entirely online (e.g. GSE). With many more journals available than before, scientists need to decide where to publish. A common choice is a journal where a particular paper is most likely to be read. If paper acceptance at such a journal is not expected, the next choice could be a more applied journal, regional journal or (as a last resort) a journal that will publish anything. After all, any paper will be searchable. The choice of the journal also depends on whether the impact factor is deemed important. In some countries, promotion documents and grant applications require attaching an impact factor to each publication, and scientists are incentivized to submit to journals with a high-impact factor, even if those are not commonly read by the animal breeding community. Having a paper in a journal with a very high impact but less relevance requires different skills, as reviewers from those journals may be less qualified to see shortcomings in the methodology and instead may focus on dramatic claims and the quality of presentation. Publishing online only provided limithed help in lowering the time to publication. Finding reviewers among those overwhelmed with too many requests is difficult. Reviewing takes time (despite artificially short deadlines that are not treated seriously by serious reviewers), and both scientific and professional editing takes time. To find references, long visits to a library were replaced by fast Internet searches, which often return an astounding number of potential references. How can we find the relevant information in the heap of “noise”? To understand which papers can be useful for our purpose, we can assign them into several overlapping categories. The first category contains papers published by graduate students to document their work to secure a job in the future. The quality of such papers is strongly determined by the interest and energy of their main advisor. The second category contains papers viewed as progress reports of a project, often supported by a major grant. Those papers are necessary to show that some work is being done, even if not much is done. Their intellectual contribution depends on the stage of the project and maybe small if the goals of the project were unrealistic. Another category is review papers. These are highly sought after by newcomers as an introduction to a topic but add little to new science. Special sessions at meetings sometimes lead to invited papers that are submitted separately by each author, or as one joint (possibly eclectic) paper. A completely different category contains papers that present new methods that are expected to be highly cited and lead to new research areas. If those papers use simulations, they may be followed by many studies evaluating the new methods. If the results of any project are spectacular, there is a desire to publish in a high-impact journal. Categorizing a paper helps in understanding what kind of content to expect. With many papers found by online search, one needs to quickly decide whether each paper is worth reading in-depth. I use a few rules. Theoretical papers that deal with complicated formulas but end with a simple formula are likely to be of greater interest than those that end with a “giant” “cryptic” formula. Simulation papers may show great achievements (because 3 of 1000 QTLs were set to explain 90% of the additive variance) but are likely less interesting than results based on a large body of real data. Papers with many (20) tables, each with perhaps 50 numbers, may hide the fact that nothing important is present. When we see results in 4 decimal digits where even two are not statistically significant, perhaps a major author or major advisor was not paying attention, and perhaps neither we should. Do we publish too much, or should we publish just one summary paper at the end of a (possibly 5-year) project? Projects are not static, and progress report papers often cross-fertilize other projects. Therefore, publishing during the process of a project instead of only once after its conclusion can be beneficial to the scientific community. Also, online publication, especially with open access, is a boon to scientists from less wealthy regions.