This research has been funded by the Publishing Research Consortium.

Introduction

1Early career researcher is the name given to unestablished (untenured) researchers. Various definitions are in operation throughout the world although, it is generally understood that they are in their twenties and thirties and either have received their doctorate and are currently in a research position or have been in research positions but are currently undertaking a doctorate. Obtaining an understanding of the attitudes and practices of early career researchers (ECRs) is extremely important because they are the fuel that powers the world’s research projects. They are not only the largest community of researchers in most countries (thought to number around a million in China alone) and universities, but also, crucially, they are the new (tidal) wave of researchers, who are born digital and bring with them the millennials belief in openness, sharing and transparency (Pew Research centre, 2010). If they—and they are tomorrow’s great scientists and Noble prize winners- are set to change the current system of scholarly communications and adapt it to their own ways of thinking then this will be transformative and of immense consequence to the stakeholders—universities, publishers, librarians and, indeed, society. But strangely and, possibly, worryingly, given the aforementioned, we actually know very little about their scholarly attitudes and behaviour. They have largely been neglected by information scientists who have tended to concentrate on their tenured colleagues or lumped them in in analyses of the whole population of scholars.

2To set matters straight, an international research project was established in 2016, which focussed purely on this research group and their potentially changing scholarly attitudes and practices. A major focus of the study is to determine whether ECRs are taking the myriad opportunities proffered by new digital innovations, developing within the context Open Science, Open Access and social media to disseminate their research. The project, the Harbingers project, is three-years long and this paper provides to the French academic audience the highlights of the first-year benchmarking exercise (more details of which can be found in Nicholas et al., 2017a, 2017b 2017c, 2017d and Rodriguez-Bravo, 2017) and then investigates the strategic changes one year on.

Methodology

3A three-year longitudinal, interview approach to the research was adopted because it was felt that this offered the best opportunity to obtain deep insights into change and to calibrate it. This is partly because asking researchers whether things are changing/will change is clearly not as effective as ‘following’ them and then observing what has actually changed and then measure it (Cohen et al., 2013). Semi-structured interviews were used because questions about technological and social change challenges the ubiquitous questionnaire because you are typically asking about things people do not know the names of, do not fully understand and are difficult to articulate. Take questions, for instance, about altmetrics or open science; they are not easily answered.

4A sample 116 of ECRs was derived from 7 countries—China, France, Malaysia, Poland, Spain, UK and US. ECRs were approached via publisher lists and research networks. The number of ECRs was dictated by the funding available and the country selection made on the basis of funder preferences and the availability of interviewers on the ground. Again, because of funder priorities around two-thirds were scientists and the rest social scientists. Most were in their thirties. See Table 1 for more details of the sample.

Table 1. Characteristics of the ECR sample (N = 116)

Country	No. of ECRs	Social Sci. (%)	Science (%)	Male (%)	Age 20s (%)	Age 30s (%)	Post-doc (%)
China	13	31	69	54	46	54	92
France	14	21	79	64	64	36	100
Malaysia	12	42	58	50	0	100	100
Poland	10	20	80	60	40	60	50
Spain	18	22	78	56	39	61	72
UK	21	38	62	62	24	76	67
US	28	21	79	61	29	71	64

Literature Review

6Nobody else appears to have concentrated fully on this crucial initial stage in a researcher’s career, the early years when researchers are attempting to establish themselves in their field, from a scholarly communications point of view. The studies of young researchers tend to focus on their challenging circumstances (Fransman, 2014); most of the studies there are have investigated ECRs, as part of a broader study, to see how different/similar their scholarly communication practices and attitudes are to those of their seniors (Mulligan et al., 2013); Nicholas and Rowlands, 2011) and, finally, others have focussed on only PhD students (Carpenter et al., 2012). The imperfect knowledge we do have tells us that ECRs are conservative in their attitudes and behaviours (for example, Harley et al., 2010; Watkinson et al., 2016); they are rated almost exclusively on the volume of papers published in high-ranking journals and the number of citations obtained (for example, Mulligan et al., 2013; Nicholas et al., 2016); non-traditional research products (e. g., social media interactions) obtain little recognition and use (Wolff-Eisenberg and Schonfeld, 2015); and they rely on their mentors also militates against straying from well-trodden reputational paths (Brechelmacher et al., 2015). Despite evidence to the effect that open research is associated with increases in citations, media attention, potential collaborators, and job and funding opportunities (McKiernan et al., 2016), young academics appear to be no keener to employ innovative methods than their seniors; rather, the contrary sometimes prevails (Harley et al., 2010).

Results

Year 1

7The first year of the study was essentially a benchmarking exercise, which sought to calibrate attitudes and behaviours over a very wide range of scholarly communication tools, platforms and issues in order to identify and measure change in ensuing years. The topics covered are listed in Table 2 and were derived from focus groups with publishers and ECRs and informed by a literature review. Full details of the findings can be found in the project report (CIBER, 2016, 2017) and in the aforementioned papers (Nicholas, Boukacem-Zeghmouri, et al., 2017; Nicholas, Rodríguez-Bravo, et al., 2017; Nicholas, Watkinson, et al., 2017; Rodríguez-Bravo et al., 2017). In this paper we concentrate on what we regard to be the broad findings that shed light on how ECRs are changing the scholarly communication market.

ECRs are manifestly paper-driven

8Publishing papers—and all the task associated with it—dominates ECRs working lives. They are even more fixated with publishing papers than their seniors because of the precarious employment position they find themselves in. Papers are the ultimate reputational currency and provide a passport to the promised land - tenured work. So, despite increases in the number/range of research outlets, courtesy of Science 2.0, ECRs dance to the same reputational tune. However, not any old papers will do, they are focused very much on publishing papers in highly ranked journals and often they are prescribed, with ECRs having to choose from institutional or government lists of acceptable journals - acceptable generally meaning listed by Web of Science & Scopus although not so much in the US/UK. This biggest and most specific list is that which governs every publishing move of Polish ECRs and is currently under review.

Peer review

9ECRs are seasoned researchers with around half of our sample having experience of peer reviewing, which probably explains their general, but not complete, support for peer review. They tend to prefer double-blind review because of the anonymity it affords, something that is felt necessary as they would not want to be viewed as junior researchers as that might disadvantage them. For similar reasons they are not so keen on open peer review either, which they consider too risky and dangerous for them because it places them directly in the spotlight. Furthermore, this form of review would attract what they regard as unwelcome, additional comment from a larger and unknown army of critics. ECRs are also of the opinion that open review makes it more difficult to reject papers because it would mute reviewer criticism. With so much at stake (their jobs and future) ECRs are inevitably very wary when it comes to peer review.

10ECRs biggest criticism is, in fact, levelled against reviewers who they complain are frequently uninformed an inappropriate. They have two main suggestions to improve the situation: 1) editors should exercise greater control and intervene more often; 2) largely emanating from Spanish ECRs, there should be a reviewer rating/matching system which has author input. They argue that they are rated and indeed everything is rated (notions of TripAdvisor) so why not reviewers? Despite having issues with reviewers ECRs have no complaints over publisher control of the peer review system. They believe there are few alternatives as they do not trust scientific societies (and feel somewhat alienated from them) and, as we have already heard, certainly do not trust the ‘crowd’.

Social media and online communities

11There are patches of social media/online community use among ECRs and they are larger than we have witnessed before (Watkinson et al., 2016). What is most revealing is the sheer range of activities that platforms are used for—finding information, communicating information, sharing information, building a digital profile/presence, obtaining PDFs and engaging in outreach activities. ResearchGate is attracting the greatest interest and its social and gaming characteristics are liked by a good number of ECRs. It is also proving a good source for finding collaborators. LinkedIn, surprisingly perhaps, and Twitter (mainly a means for alerting people of new articles, conferences etc.) are also tools of choice for scholarly engagement.

Open access (OA)

12Characteristically, ECRs like the idea and principles behind OA as it fits their views on openness and sharing, but it is rarely a priority when it comes to choosing a journal in which to publish their papers. They and their more senior co-authors are very pragmatic and opt for journals that will benefit their reputation most and these are high impact factors journals and they tend not to be typically gold OA. Given that ECRs are the last in the queue when it comes to obtaining resources in terms of accessing funding there is the inevitable disquiet about (high) author charges making the playing field uneven between those that can pay (tenured researchers and those that cannot (ECRs). Regarding green open access—archiving in repositories—ECRs are completely non-plussed, they typically see that as a job for others (librarians were mentioned) and also do not see any reputational or career progression advantage depositing. Interestingly, they are not really aware of institutional or funding mandates requiring them to deposit.

Open science (OS)

13We have already talked about two components of OS, open peer review and open access, and we have seen a gulf between attitudes and practices, so what about the broad concept itself and the other major component of OS, open data, for instance? As the topic falls into the openness and transparency category you might expect it to have much traction with ECRs. But, again, while they are mostly warm towards the concept (although not so in France where researchers are either not well informed or almost antagonistic towards the concept), this turns out not to be the case and this might be partly put down to the fact that it is still early days in the gestation of the idea and practice. Despite increasing promotion of OS by the EC and research funders, most ECRs display little understanding of what it means.

14There is, therefore, a degree of confusion in their answers as they do not know they are practising OS or not. Thus, some think blogs, social media etc. are part of OS agenda and others simply equate OS just with OA. There are country variations, though, and as mentioned earlier OS is barely known in France and French ECRs and have a mistrust the agenda. “A new means of imposing control and evaluation.” However, there is a much higher awareness/appreciation in the US. Given all this, there is a need to be wary of questionnaire studies of the topic, where explanation and guidance cannot be provided, as is the case with deep interviews.

15What then of open data a topic being much promoted? Well, ECRs were not so interested in open data (and software) because many would like to exploit their data to the full themselves (for their publications and reputation) and not give it away where it could be misused or used more profitably (the later a case of the competitive instinct kicking in). Again, French ECRs are the main naysayers believing that it is not always relevant/useful to make their data available openly.

Collaboration and sharing

16Collaboration and sharing are much mentioned by ECRs as central to the way they want to run their scholarly lives and it is not simply a case of warm words as ECRs genuinely believe this. But they do feel their wings clipped a bit by traditional reputational requirements, which offer no rewards in respect to collaborative and sharing activities employing such tools as Facebook, Twitter and ResearchGate. Nevertheless, they still use these tools, although not as much as they would like, especially for finding researchers to collaborate with or talk to.

Metrics

17Metrics, if they are citation based, are, of course, something ECRs are very interested in because they, especially impact factors, are the means by which they benchmark their success and determine their reputational standing. Altmetics, on the other hand, is an example of a topic which is widely trialled by conferences and university research administration departments, but about which ECRs exhibit little awareness or interest. Again, this could well be because they have not caught up with these metrics yet, but it is also a case that they are not widely used and accepted by researchers or university system for reputation.

Transformations

18The traditional scholarly communication and reward system has been sitting on a potential time-bomb of digital innovation and social media for around a decade now and signs that it might explode anytime soon could be the runaway success of ResearchGate, a pure player (born and wholly digital with no industry antecedents) built entirely along social media and gaming lines which has a rapidly expanding following of 16 million researchers. So, ECRs were asked whether they could see the field transformed down the line (five years was the timeline given) and, if so, what would be the transformations. Now, interestingly, ECRs could see the opportunities to change and the benefits of so-doing. However, they felt unable to take them, because there were no opportunities to do so in the insecure/busy environment in which they worked and where they were shackled to a very traditional and unforgiving reputational system. However, we seemed to have moved on from situation which we found earlier (Watkinson et al., 2016) where they had no ideas about change and those who disliked it just railed against it. Now we are being presented with ideas for change and that is mainly moving away from the preoccupation with publishing papers and more towards outreach activities and the need for greater transparency

Year 2 – the main changes

19One year on and we found our ECR panel much more experienced, informed and calculating. It was as though they had experienced a fast-track year which had aged and seasoned them. They had obtained greater exposure to scholarly initiatives, which had certainly been ramped up, most notably by the EU and had experienced increased institutional pressures as universities attempt to raise their reputational game. They also had had time to observe and note what can be useful to them in their career. This all meant that their behaviour was likely to change. Thus, if they saw something that they once criticized could fast track them along their careers then they were going to adopt this practice. There was also a sense that they had learnt something from our previous interview and were just more aware.

20We used 4 descriptors to portray the scale of change—significant (e.g. a step change for the majority of ECRs), small (e. g. incremental change effecting a minority of ECRs), stirrings (e. g. ECRs talk about the desire for changing but no changes occur in practice) and none. No surprise, then, given the above that we saw a change of some kind occurring in respect to all the 23 scholarly categories (Table 2). Taking significant and small changes together, the categories that saw the most change (that is change for over 33% of ECRs), in ranked order were:

• Jobs (60%). Meaning project/role/status change and career aims/ambitions. This category also saw the most significant growth taking place (44%). Clearly this high ranking can be explained by the precarious, transient and turbulent nature of ECRs’ work. Take the case of Chinese ECRs, for instance, where two-thirds saw job changes. Two left their jobs as university researchers and started working for business; 2 have obtained tenure; two have left the university for research positions in the USA and UK as visiting scholars; 3 have changed research groups/centres; 2 have changed the direction of their research; and the other two have become mothers.

• Online scholarly communities (54%). There was widespread growth in active usage, especially of ResearchGate (previously ECRs were more voyeurs); more awareness of digital visibility opportunities they provide; and a stronger belief they are the future. The 54% score was boosted by a very high small change score (32%) suggesting, perhaps, there is greater growth in the pipeline.

• Careers (40%). Covering the career aims, ambitions and motivations held by ECRs. This result can again be put down to the turbulent environment ECRs inhabit.

• Authorship (36%). There appeared to be a general tightening up of the rules surrounding authorship, especially regarding the positioning of authors and the role of corresponding authors (e.g. they no longer obtain reputational credit) and, generally, ECRs were just more reflective on authorship. This was less the case in the UK and US and is, probably, explained by the greater maturity of these two countries, meaning that such things were long-settled.

• Open access publishing (35%). OA is becoming more acceptable and this seems not so much to do with mandates, but much more down to pragmatic thinking, with ECRs being swayed by the claims regarding the benefits of greater outreach, increased citations and (possibly, imagined) speed of publication that OA publications bring with them.

• Peer review (34%). ECRs: a) were more selective in accepting reviews and more proactive in obtaining them; b) showed more disquiet about the process, especially so in China and Malaysia, because of a flurry of recent retractions in these parts. The problems that arose were put at the door of the enormous pressures on researchers to publish.

• Social media (34%). Social media has obtained a low-level acceptance from many ECRs. There is more use of twitter, Facebook and LinkedIn and mainly for dissemination and outreach purposes, which are thought to be very important by ECRs.

21If we look at the categories where there has been very little change over the past year, then, metrics (traditional citation counts) clearly stands out where there was only a 11% change in behaviour/attitudes. Even if we take stirrings of change into account (5%) change has not occurred for 84% of the ECR sample, suggesting that metrics have still got a grip on scholarly communications. Little change is also happening regarding data (17%) and libraries (18%). Smartphones are not making much scholarly headway either, except in China.

22

Diversity

23There were big country differences, with China (18%) and France (17%) showing the most significant changes (Table 3). Much of the changes in France concerns jobs and careers rather than scholarly communication attitudes and habits. In China change concerned scholarly communications as well, with peer review and open access subject to changes. If we count significant and small change together then Malaysian ECRs are the harbingers of change (42%). In the case of Spain, it’s very much a case of stirrings of change (34%). In UK, Poland, and USA there is a preponderance of no change. There were differences between scientists and social scientists, with the former more: likely to work in groups; productive in terms of papers published; concerned with publishing in high impact factor journals; collaborative and knowledgeable about scholarly communications. However, there were no real differences in the scale of change. In regard to status, those doing a PhD were slightly less likely to change.

Table 3. country diversity and change (in order of ascending significance)

Country	None	Stirring	Small	Significant	N/applicable
Poland	75%	10%	14%	2%	0%
UK	77%	2%	17%	3%	0%
US	73%	1%	24%	3%	0%
Spain	40%	34%	20%	6%	1%
Malaysia	51%	7%	34%	8%	0%
France	48%	13%	22%	17%	0%
China	54%	3%	21%	18%	4%

Conclusions

24When a big, strategic international market notches up even a small degree of change it makes a real difference across the board and globe. And our research shows this, and more in fact, because in a number of areas of scholarly communication we have documented widespread and significant changes in the space of just one year. There are definite signs of ECRs rattling, if not yet breaking out of, the reputational chains that have constrained and regulated their behaviour and practices for decades. The most significant changes are occurring in respect to attitudes and behaviours in regard to authorship, open access, peer review and online communities. There are also are many small changes occurring, for instance, in the case of scholarly reputation, which could be prophetic—maybe, the sign of bigger, deeper and structural change down the line. The stirrings of change we have detected could also auger change even further down the line and, in this respect, it is worth marking our card in regard to open data and research impact. In the end, the scale of change that occurs will depend, of course, on whether ECRs take their millennial beliefs in sharing, openness and transparency into leadership positions. A good number are moving to permanent posts, so we should know a little more next year.

25Of course, ECRs are not the only harbingers or drivers of change, there are the innovative digital technologies, which are disrupting forces and none more so than ResearchGate, with its 16 million members. RG, a pure player, is far more than an online social community, it is, in fact, a one stop scholarly shop meeting with engagement and networking at its very heart. ECRs have taken to RG in increasing numbers, especially for the digital visibility and sharing opportunities it offers. Virtually all our ECRs are members and many active members in that they create profiles, upload and share. Give ECRs a tool that is built more on their millennium principles then we really have the fuel to transform scholarly communications.

26The other great attraction of RG much appreciated in fact by the whole scholarly community, but even more so by the lowest members of that community (ECRs) is that it is free and open. The paywalls of the publishers and aggregators’ sites in comparison are beginning to look like the walls of old mediaeval castle, no longer effective in the age of drones and helicopters.

27Finally, not only have we witnessed change on a wide front, but we have also detected a momentum that looks unstoppable. There are big warning signs here for the stakeholders—publishers, librarians, universities and researchers, most notably, that ECR beliefs in respect to openness, sharing and collaboration are not going away and this will challenge what they do (and their bottom line)