دانلود رایگان مقاله پدیده های ماندگار در انتشار علمی: تمرین شبیه سازی

abstract

We model the academic production process understood as the creation, submission, evaluation and publication of papers: scientists produce manuscripts to the best of their abilities and try to publish them in academic journals, which rely on referees to judge the submissions. The resulting model is able to reproduce several properties of the journal-landscape but also illustrates that even under unrealistically optimistic assumptions the process of scientific publishing will give rise to several universal emergent phenomena for purely mathematical reasons: the efficiency of scientific publishing is delicate and very unstable.

6. Conclusion

In this paper we have demonstrated that in a perfect world, where estimates of scientific quality are always accurate and strategic behavior is absent, peer-review indeed is a viable tool suitable for objectively clustering academic research in outlets of different quality. However, we have also shown that even a minimal deviation from the idealized conditions drastically affects the outcomes of the academic production process. Already tiny misjudgements from authors/referees as well as minimal strategic considerations by authors lead to a clustering of journals and a high variability of quality among mediocre journals. Moreover, we observe a highly idiosyncratic development of the probability of rejection with respect to quality, where overall rejection rate increases with the quality of manuscripts for a majority of the population of scientific papers – this is tightly coupled to a mismatch between the quality of papers being produced and the distribution of the quality ofjournals. These results suggest a new dimension to the traditional criticism of practises in academic publishing: even in the absence of human fallibility, ‘hot’ topics, the pressure of grants and tenure, underlying phenomena emerge for purely mathematical reasons and are potentially harmful to the scientific process at large.