Collection, curation, and quality: The editor's responsibility

Balancing quality and inclusion The role of journals in the recording and dissemination of science and scholarly information has been a topic of discussion throughout the life of Learned Publishing, and it does not look like it is going away anytime soon. The responsibility of journals – and editors in particular – has increasingly come into the spotlight, especially with the rise of apparently fake news and article retractions (RetractionWatch, 2016), as well as the well-documented rise of journals which do not conform to accepted western standards. Adding to all of this are extensive discussions over the so-called reproducibility crisis in science, raising concerns about the influence of articles reporting non-replicable (and potentially dangerous) findings (see Fanelli, 2018). Several articles recently published in this journal raise issues over trust and integrity in the scholarly environment and within journals. Issues covered include the publishing decisions made by researchers which show how the pressure to publish can overcome the desire to publish well (high quality articles in high quality journals; Kurt, 2018; Shehata & Elgllab, 2018). This is exacerbated by too few (high quality) journals to accommodate the output of research – rising by 22% in some regions according to the most recent data emerging from both Scopus and Clarivate (see Fig. 1). So perhaps it is worth asking again about the level of responsibility of editors. Should editors be accepting and publishing more widely, or more narrowly? Should we be increasing the filtering of research, or widening our collections? In one of the plenary sessions at the recent SSP conference in Chicago, Safiya Noble made the point that if you are not published, then you are not part of the scholarly record. She argued that this has led to the exclusion of some sections of the global population (based on race, location, sex, etc.), which would advocate for more inclusive publications. However, this flies in the face of the more selective (and high impact) strategies of both journals and indexes which promote the promise of “the best of the best” to researchers and remain the most desired target for authors. Each of these three important questions would require a longer article than this editorial to discuss adequately. The global increase in article output, with increasingly diverse submissions has also brought about its own problems. Whilst the addition of articles from other regions is to be welcomed, language is a frequent concern. At the recent EASE conference (European Association of Science Editors), several sessions were dedicated to the problems (and opportunities) of publishing and editing in different languages. Whilst English may be accepted as the international language of science, many regions need local languages to be represented. This means that editors and publishers have to invest in multi-lingual editing and publishing (requiring specialized people and more money) for parallel language publications. For example the German Medical Journal, Deutsches Ärzteblatt, which publishes both a German-language and an English-Language edition, has to support translations where required (https://www.aerzteblatt.de/int/), which need their own level of quality control. The EASE meeting had a lively exchange about the linkage of different language versions, and whether the translated version could – or should – be considered identical to the original version, raising issues of version control and inflated citation counts. Scrutiny by readers appears to be increasingly keen, and so to avoid post-publication problems editorial oversight requirements increase each time something goes wrong. Fear of long-term repercussions of any error haunts both editors and publishers, and a recent interview with Deborah Kahn from Taylor & Francis about the “controversial past” of Dove Press (acquired by Taylor & Francis several years ago) demonstrates that memories are long. (The interview was spurred by a recent problem article, but framed by the so-called historical issues with Dove Press: Poynder, 2018.) It seems now that each submission cannot be taken on trust but must be checked for plagiarism, ethical approval, compliance with submission guidelines, data deposit, and a raft of additional author requirements. Suggested reviewers need to be checked to ensure they actually exist, and are not friends or colleagues. Authorship lists need scrutiny to ensure that they do not include guest authors, and that all authors deserve to be listed as authors according to the journal's authorship criteria – and they may also need to provide contributor statements. And, of course, all authors need to assert that they agree to the submission, and as average authorship numbers increase, this adds to the administrative burden. A recent interview with Richard Gray, Editor of the Journal of Psychiatric and Mental Health Nursing (Chilvers, 2018), raised the question of whether editors/journals were responsible for checking all references to ensure that none of them referred to retracted or corrected findings. A comment made at the EASE conference from one of the larger journals said that they actively discourage the citing of reviews, and ask authors to cite the original research, so the original researchers received citation – something else for editors to watch out for. Another large journal has editors specifically to reduce the “spin” that authors add to make their research look more impressive and important. There are some technical answers to these additional requirements, even the possibility of applying artificial intelligence (AI) to peer review (see Mrowinski, Fronczak, Fronczak, Ausloos, & Nedic, 2017); however, editorial oversight remains key, and (in the foreseeable future) AI is unlikely to entirely replace the human factor in peer review (in my own opinion). The EASE conference also had a session on how editors can contribute to sex and gender equity in research and publishing through journal guidelines and policies (particularly supporting the SAGER [Sex and Gender Equity in Research] reporting guideline: Heidari, Babor, De Castro, Tort, & Curno, 2015). This adds to the list of reporting guidelines that journals now ascribe to, and one speaker made a controversial presentation about why they would not be adopting the SAGER guidelines as “yet another” guideline that added to both the authors’ and editors’ workload. None of these checking requirements are wrong or should be criticized as being a waste of editorial time. However, the time-consuming nature of what is being required calls into question the ability of editors to ensure that everything is checked and OK'd. This is exacerbated by the fact that most editors fit their editorial and reviewing roles around regular, full time, employment, and many lack editorial office support. The build-up in workload may therefore lead to decisions in favour of “easier” articles, rather than unbiased consideration of each submission. The presentation at EASE which challenged the idea of adopting another set of guidelines along with the interview with Richard Gray have raised an important question that journals and editors need to ask: at what point does the checking stop, and how is it influencing the ideal of providing a curated yet broad collection of articles? The tension between increasing the range and inclusivity of what we accept to publish and the need to filter and curate valuable information are both heavily influenced by the increasing quality and compliance checks required. With the proviso that every journal is different, at what point does the combined requirement for editorial checking and the need to manage increasing articles become unrealistic? Where, and how, do we draw the line? The ability of journal editors to find the balance between collection, curation, and quality may be impossible.