Perspectives on Open Science and Scholarly Publishing: a Survey Study Focusing on Early Career Researchers in Europe

This question explored the general awareness regarding OS, in order to determine the most and the least aware groups across different grouping variables. It was the “key” question, unlocking the other questions from the “OS section” for those who answered positively. The majority of respondents (76%) had heard about OS. However, we found differences across groups, mainly relating to the region of Europe, GDP, and GERD as % of GDP (Figure 3).

Figure 3. Reported awareness of OS.

Respondents were asked if they had heard of OS (answered by indicating either “Yes” or “No”). Cumulative bar plots display the percentages of reported awareness for (A) all respondents, or (B) based on European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Statistical differences were evaluated with a Chi-squared test - results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question was available only for those who indicated that they had heard about OS (76% of the survey participants). It explored the level of OS knowledge among respondents, who were offered a 10-point scale for self-assessment, with the lowest mark (1) meaning “have just heard about it” and the highest (10) meaning “advanced expert.” The aim was to determine the most and the least knowledgeable groups across different grouping variables.

We found that the highest level of self-assessed knowledge of OS was in researchers in western European countries (Figure 4A), by a considerable margin. Equally, the highest knowledge level, based on GDP per region (Figure 7B), was found in the 40,000 – 55,000 range, with much of this group containing nations in western Europe. This trend shows that researchers from this region are most knowledgeable and positive in regards to OS; this trend is visible in other analyses below also). However, we found no significant differences between GERD as % of GDP categories. Among research disciplines (Figure 7D), the researchers in Information Science and Engineering and Interdisciplinary science reported the greatest knowledge of OS. We also found knowledge of OS appears to increase with increasing career seniority (Figure 7E).

Figure 4. Self-assessment rating of knowledge on OS based on (A) European region, (B) GDP, (C) GERD as % of GDP, (D) field of study, and (E) career stage.

Figure 5. Reported agreement with the statement ‘Open Science is generally a good thing”.

Figure 6. Most important features of Open Science identified by (A) all respondents or within (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage.

Abbreviations: Sc.: Science, Eng.: Engineering.

Figure 7. Most important advantage of Open Science.

The graphs illustrate the mean and standard deviation of the responses from each group, with the y-axes representing the numerical rating scale and x-axes displaying the different categories assessed. General statistical significance was measured by Kuskall-Wallis test and is indicated above each panel. Stars indicate significance for multiple comparisons with Dunn’s test, with corrected p-values. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001.

The next question sought feedback regarding participants’ general attitude towards OS, to determine the respective differences across various grouping variables and was, again, available only for respondents indicating that they had heard about OS (76%). Out of those, 90% (84% of the whole sample) agreed that OS “is generally a good thing”. We found no significant differences between any of the variables’ groups; at least 80% of respondents in any European region, GDP and GERD as a % of GDP ranges, the field of study, and career stage, kept the same belief. The biggest difference, although not significant, was found for field of study: representatives of Interdisciplinary and Life Sciences showed the most positive attitude, while representatives of Physics and Chemistry the least.

The percentages were calculated based on the classification of respondents. Cumulative bar plots show the results for (A) all respondents, (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Statistical differences were evaluated with a Chi-squared test - results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

We next explored which features and activities of OS were most important to respondents aware of OS, to determine which specific elements across different grouping variables are more appealing to researchers. We noticed relevant differences in the responses across the groups, the most significant in the region, GDP, and GERD as % of GDP; while attitudes towards the importance of reproducible research, and open and FAIR data varied the most (Figure 6, Table 1).

In general, respondents consider Open Access the most important feature of OS, followed by reproducible research and open and FAIR Data (Figure 6A). OPR in considered much less important. The rest of the features (Open Licenses for research outputs, Open Metrics and Impact, and Citizen Science) are not considered important by most of responders. Notably, reproducible research and open and FAIR data are considered much less important by respondents in eastern Europe (Figure 6B) and the countries within the 2,500-10,000 GDP rnage (Figure 6C) and less than 0.75% regarding GERD as % of GDP (Figure 6D). This was also the case for respondents in economics sciences (Figure 6E).

Respondents could choose more than one feature. Radar charts display the results in terms of the percentage of total respondents of each category. Statistical differences were evaluated with a Chi-squared test and results are shown in Table 1.

We next explored what advantages respondents who were aware of OS, felt were offered by practicing OS, in order to determine the main incentives. Looking across the entire cohort studied, researchers in Europe indicated that the most important advantage of OS was “greater availability and accessibility of research outputs” (51.8% of responders). At the same time, we found differences across groups, the most significant in the European region, GDP, and GERD as % of GDP (Figure 7).

“Greater availability and accessibility of research outputs” was less crucial for respondents in western Europe (Figure 7B) and the countries within the 40,000 - 55,000 GDP range (Figure 7C). “Greater reproducibility and transparency of research outputs” was much less important on average, but crucial for respondents in western Europe (Figure 7B), in the countries within the 40,000-55,000 GDP range GDP (Figure 7C), and in those spending the most on research (Figure 7D). Other options were considered much less important. Notably, the perceived importance of the “possibility of more transparent and rigorous peer-review process” tended to decrease with increased values of GDP (Figure 7C) and GERD as % of GDP (Figure 7D) and to grow with the career seniority (Figure 7F).

Respondents were asked to select only one main advantage from a standard list of answers or select the option “Other”. This question was not compulsory. Percentages were calculated based on the total number of respondents per category. Cumulative bars plots show results for (A) all respondents, or classified based on (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Statistical differences were evaluated with a Chi-squared test; these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question explored the main concerns respondents (only those aware of OS) had around OS, in order to identify the hurdles behind OS uptake. Researchers in Europe indicated that the “emergence of low-quality and false science” was their main OS-related concern (32.9% of respondents). Importantly, we found differences across groups; the most significant in the European region, GDP, and field of study (Figure 8).

Figure 8. Main concerns around OS.

“Emergence of low-quality and false science” was particularly concerning for responders in northern and southern Europe (Figure 8B), the countries within the 55,000 or more and 25,000-40,000 GDP ranges (Figure 8C), and for respondents in chemistry, physics, and interdisciplinary (Figure 8E). It was less worrying for respondents in western Europe (Figure 8B), the countries within the 40,000 - 55,000 GDP range (Figure 8C), and for responders in social sciences and humanities and life sciences (Figure 8E). Notably, the popularity of this concern decreased with career seniority (Figure 8F). “Missing sufficient training, tools and infrastructures” was the primary concern for responders in western Europe (Figure 8B) and the countries within the 40,000-55,000 range (Figure 8C) and was more prevalent among respondents in information science and engineering, social sciences and humanities, and life sciences (Figure 8E). “Potential misuse of scientific research outputs” and “the public may misunderstand research outputs” were more concerning for respondents in eastern Europe (Figure 8B) and less wealthy countries in terms of GDP (the ranges below 25,000 - Figure 8C) and those spending the least on research (Figure 8D). This was also relevant for responders in economic sciences (Figure 8E). “More amount of work required from researchers” was less concerning for responders in northern and southern Europe (Figure 8B), the wealthiest countries in terms of GDP, (Figure 8C), and those in the 1.25-1.75% range (Figure 8D). This was also the case for responders in chemistry and economics sciences (Figure 8E). Importantly, the popularity of this concern increased with career seniority (Figure 8F).

Respondents were asked to select only one main concern from a standard list of options or to select the option “Other”. This question was not compulsory. Percentages were calculated based on the total number of respondents per category. Cumulative bars plots display results for (A) all respondents, or classified based on (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Statistical differences were evaluated with a Chi-squared test; these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question explored the general awareness regarding OPR to determine the most and the least aware groups across different dimensions. It was the “key” question unlocking the other questions from the “OPR section” for those who answered “Yes.” Most respondents had not heard of OPR (46% answered positively). We found differences across groups; the most significant in the GDP and European region (Figure 9).

Figure 9. Reported awareness of Open Peer Review (OPR).

Unlike the rest, most responders in western Europe (Figure 9B) and countries within the 40,000-55,000 GDP range (Figure 9C) were aware of OPR. Respondents in southern and eastern Europe (Figure 9B) and countries within the 25,000-40,000 GDP range (Figure 9C) and in those spending below 1.25% of GDP on research (Figure 9D) were the least aware, which was also the case for the respondents in physics and economic sciences (Figure 9E). Awareness of OPR tended to grow with career seniority (Figure 9F).

Respondents were asked if they have heard about OPR. Cumulative bars plots display percentages of awareness for (A) all respondents or based on (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Statistical differences were evaluated with a Chi-squared test, these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question was available only for those who indicated that they had heard about OPR. It further explored respondents' experience passing through the OPR process to determine the level of engagement in this open practice from the author’s side. Only 22.6% of those who responded to this question had had this experience, and we found some differences among groups, the most significant in career stage and European region (Figure 10).

Figure 10. Reported previous experience on passing through the OPR process.

Experience passing through the OPR process significantly increased with career seniority (Figure 10F). The most experienced responders were in eastern Europe and the least in northern and central Europe (Figure 10B) and countries within the 10,000-25,000 GDP range (Figure 10C). This was also the case for researchers in physics and information science and engineering (Figure 10E).

Respondents were asked if any of their papers had been through OPR. Percentages were calculated based on the total number of respondents per category. Cumulative bar plots show the results for (A) all respondents, or classified based on (B) region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. This question was not compulsory. Statistical differences were evaluated with a Chi-squared test; these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question was again available only for those who indicated that they had heard about OPR. It further studied respondents' experience in OPR from the reviewer’s side. Even fewer respondents (17.6%) had had the experience of reviewing in OPR, and we found some differences among groups, the most significant in career stage and region again (Figure 11).

As in the previous question, the experience of reviewing research outputs openly significantly increased with career seniority (Figure 11F). However, unlike the previous question, the most experienced responders were in western Europe (Figure 11B) and in countries within the 40,000-55,000 GDP range (Figure 11C). The least experienced responders were in northern Europe (Figure 11B). Regarding the field of study, respondents in chemistry, physics, and social sciences and humanities had had the least experience (Figure 11E).

Figure 11. Reported previous experience as a reviewer on an open peer review process.

Respondents were asked if they have been a reviewer in open peer review. Percentages were calculated based on the total number of respondents per category. Cumulative bar plots show the results for (A) all respondents, or classified based on (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. This question was not compulsory. Statistical differences were evaluated with a Chi-squared test, these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

The next question sought feedback regarding participants’ attitudes towards OPR compared to the conventional one to determine the differences across various dimensions. This question was available only for respondents indicating that they had heard about OPR. We found no significant differences between groups in any of the studied variables. Overall, 41.8% of those who responded to this question thought that OPR is better, while 45.3% answered “Not sure,” which matches the lack of OPR-related awareness in general.

We next explored what advantages respondents, who are aware of OPR, felt were offered by practicing OPR to determine the main incentives. Looking across the entire cohort studied, researchers in Europe indicated that the most important advantages of OPR were “encourages reviewers to be more tactful and constructive,” and “improves communication and understanding between authors, reviewers, editors and the broader community in general” (both options gained 35.0% of votes of those who answered this question). We found differences across groups, the most significant in the field of study, European region, and GDP (Figure 12).

Figure 12. Reported most important advantages of Open Peer Review.

“Encourages reviewers to be more tactful and constructive” was selected less by respondents in eastern Europe (Figure 12B), the least wealthy countries in terms of GDP (Figure 12C), and those spending less than 1.25% of GDP on research (Figure 12D). This was also the case for responders in chemistry, interdisciplinary, and physics, while for respondents in environmental and geosciences and life sciences, this advantage was significant (Figure 12E). “Improves communication and understanding between authors, reviewers, editors and the broader community in general” was more important for researchers in northern Europe (Figure 12B) and those working in physics and interdisciplinary fields (Figure 12E). This advantage was less critical for responders in eastern Europe (Figure 12B), the least wealthy countries in terms of GDP (Figure 12C) and in those spending the least on research (Figure 12D). This was also the case for respondents in environmental and geosciences and economic sciences (Figure 12E). “Leads to more objective reviews” was much less critical for respondents in northern and southern Europe (Figure 12B) and those working in physics (Figure 12E). This advantage was more important for respondents in chemistry and economic sciences (Figure 12E). “Helps to detect reviewers' conflicts of interests” was most important for respondents in eastern and southern Europe (Figure 12B), the least wealthy countries in terms of GDP (Figure 12C) and in those spending less than 1.25% of GDP on research (Figure 12D). This was also the case for respondents in chemistry and social sciences and humanities (Figure 12E), as well as for senior researchers (Figure 12F). This advantage was less critical for respondents in central and western Europe (Figure 12B), in countries within the 40,000-55,000 GDP range (Figure 12C), and researchers in interdisciplinary and economic sciences (Figure 12E).

Respondents were asked to select only one main advantage from a standard list of answers or select the option “Other”. This question was not compulsory. Percentages were calculated based on the total number of respondents per category. Cumulative bar plots display the results for (A) all respondents, or classified based on (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Statistical differences were evaluated with a Chi-squared test, these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

The next question explored the main concerns respondents (those aware of OPR) faced when practicing OPR to identify the obstacles in the OPR implementation. Researchers in Europe indicated that it “may disadvantage early-career researchers and be an advantage for established ‘big name’ researchers”. This was their main OPR-related concern (39% of responders), followed by “open reports may be less critical” (24%), “increased likelihood of reviewers declining to review” (21.8%), and “more amount of work required from reviewers” (10.8%). We found no significant differences between groups.

This question explored the level of perceived reward for peer reviewing among respondents, who were offered a 10-point scale for self-assessment, where the lowest mark (1) meant “not rewarded at all” and the highest (10) - “extremely rewarded.” The aim was to determine the most and the least rewarded groups (self-assessed) across different grouping variables. On average, the self-assessed level of feeling rewarded was 4.4.

The results are displayed as follows: (A) European region, (B) GDP, (C) GERD as % of GDP, (D) field of study, and (E) career stage. The graphs illustrate the mean and standard deviation of the responses from each group, where the y-axes represent the numerical rating scale and x-axes display the different categories assessed. General statistical significance was tested using the Kuskall-Wallis test and is indicated above each panel. Stars indicate significance for multiple comparisons using the Dunn’s test, with corrected p-values. * P<0.05, ** P<0.01, *** P<0.001, **** P<0.0001. The statistical test did not include the overall results as its main purpose was to detect differences between the different subcategories.

The number of answers to this question were substantially lower than the others (7.7% of responses), matching the generally low level of experience in OPR among European researchers. Hence, we abstained from performing analysis of these responses.

We studied which factors respondents felt were decisive when choosing where to publish research outputs to identify desirable features of publishing venues. Most researchers in Europe still consider the journal impact factor the most important variable, followed by a high-quality peer review process. Moreover, we found significant differences across groups, especially in the European region and GDP (Figure 7, Table 2).

The results are displayed as follows: (A) results including all respondents, (B) region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. Respondents could choose more than one channel of preference. Radar plots display the results in terms of the percentage of total respondents of each category. Statistical differences were evaluated with a Chi-squared test and results are shown in Table 2.

This question explored the general awareness regarding open publishing venues to identify the most and the least aware group across different grouping variables. It was the “key” question unlocking the other questions from the “open publishing venues section” for those who answered positively. The majority of respondents were not aware (only 39.6% answered positively). However, differences across groups were found, the most significant in the GDP, region and GERD as % of GDP (Figure 15).

Figure 13. Self-assessment regarding feeling rewarded writing peer reviews

Figure 14. Reported most important factors for deciding where to publish research outputs.

Abbreviations: Sc.: Science, Eng.: Engineering.

Figure 15. Reported awareness of open publishing venues.

Respondents in western Europe (Figure 15B), in countries within the 40,000-50,000 GDP range (Figure 15C), and within the 1,25-1.75% range (Figure 15D) were the most aware of existing open publishing venues. Researchers in eastern Europe (Figure 15B), in the least wealthy countries in terms of GDP (Figure 15C) and in those spending below 1.25% of GDP on research (Figure 15D), were the least aware. Junior researchers were the least knowledgeable compared to their more senior colleagues (Figure 15F).

Percentages were calculated based on classification of respondents. Cumulative bar plots display the results for (A) all respondents, or classified based on (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. This question was not compulsory. Statistical differences were evaluated with a Chi-squared test; these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question was available only for those who reported being aware of open publishing venues. It further explored respondents' experience in OS practices to identify the level of engagement with existing open publishing venues. From the offered list of such venues, respondents were mostly experienced with PLOS One, but significant differences among groups, especially in the GDP, European region, and field of study were again found (Figure 16, Table 3).

Figure 16. Reported experience with Open published venues.

Abbreviations: Sc.: Science, Eng.: Engineering.

Although some of the open publishing venues listed in the survey were more popular among respondents (Figure 16A), it is evident that generally, some groups of survey participants are much more experienced than others. Respondents in western Europe (Figure 16B) and countries within the 40,000-55,000 GDP range (Figure 16C) are the most experienced. In contrast, respondents in eastern Europe (Figure 16B), the least wealthy countries in terms of GDP (Figure 16C), and in those spending the least on research (Figure 16D) have much less experience. Regarding the field of study, experience with specific publishing venues might reflect their thematic focus. However, respondents in life sciences, interdisciplinary, and information science and engineering have more experience, while respondents in physics and economic sciences have less (Figure 16E).

Respondents were given a set of open journals from which they could select multiple venues. They were given the option to add “Other” venues, but only individual venues that accounted for more than 5% of the responders are shown. Radar charts display responses by (A) all respondents, or classified by (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. In total, 393 (34%) respondents answered this question. The results are shown in terms of percentage of total respondents of each category. Statistical differences were evaluated with a Chi-squared test and results are shown in Table 3.

This question was again available only for those who reported being aware of open publishing venues. We explored which features of open publishing venues were reported as being the most important to identify the main motivations underlying their use. “Ability to address a wider audience” and “potential scientific impact and citations” were reported as the most important motivations, and we found important differences between categories, especially in the GDP, GERD as % of GDP, and European region (Figure 17, Table 4).

Figure 17. Reported motivations to publish research outputs in an open publishing venue.

Abbreviations: Sc.: Science, Eng.: Engineering.

Respondents were given a set of motivations and could select multiple options. Results refer to analyses including (A) all respondents or classified by (B) European region, (C) GDP, (D) GERD as % of GDP, (E) fields of study, and (F) career stage. In total, 429 (37%) respondents answered this question. Radar charts display the results in terms of the percentage of total responders of each category. Statistical differences were evaluated with a Chi-squared test and results are shown in Table 4.

As the survey was initially developed to inform the ORE team, we explored the awareness level of this project across Europe. At the time of the survey, only 19.9% of respondents had heard about the ORE platform. Despite the general low awareness level, we spotted only minor differences across categories, namely in European region and Career stage (Figure 18).

Figure 18. Reported awareness of the coming ORE establishment.

Respondents were asked if they have heard about the EC’s plans to establish ORE. Cumulative bar plots display percentage of reported awareness for (A) all respondents, (B) European region, (C) GDP, (D) GERD as % of GDP, (E) field of study, and (F) career stage. This question was not compulsory. Statistical differences were evaluated with a Chi-squared test; these results are shown below each graph. Degrees of freedom and sample size per categorization are shown between parentheses and p-values are shown between brackets.

This question explored respondents' experience in science communication through social media to identify the level of engagement with the most popular online social media platforms. From the offered list of platforms, responders reported being mostly experienced with ResearchGate. Significant differences were found among groups, particularly in GDP, European region, and field of study (Figure 19, Table 5).

Figure 19. Reported usage of different social media channels for dissemination of research outputs.

Abbreviations: Sc.: Science, Eng.: Engineering.

Unlike the others, respondents in eastern Europe (Figure 19B), in the least wealthy countries in terms of GDP (Figure 19C), and in those spending the least on research (Figure 19D) rarely use Twitter to disseminate their research outputs but use Facebook and Academia more often. Respondents in western and northern Europe (Figure 19B), in the wealthier countries in terms of GDP (Figure 19C), and in those spending more on research (Figure 19D) tend to choose Twitter more often. Using LinkedIn to disseminate research outputs is more prevalent in northern Europe (Figure 19B). Respondents in social sciences and humanities, and economic sciences use Academia and Facebook more, while Twitter was more prevalent among life sciences and interdisciplinary respondents (Figure 19E).

The results are displayed as follows: (A) results including all respondents, (B) European region, (C) GDP, (D) GERD as % GDP, (E) field of study, and (F) career stage. Respondents could choose more than one channel of preference. Respondents were given the option to add “Other” social media used for dissemination, however these responses are not shown as each account for less than 5% of responses. In total, 1017 (88%) respondents answered this question. Radar charts display the results in terms of percentage of total respondents of each category. Statistical differences were evaluated with a Chi-squared test and results are shown in Table 5.

This question explored the level of satisfaction in regards to research outputs dissemination through social media channels, in order to determine the most and the least successful groups in this regard. Respondents were offered a 10-point scale for self-assessment, where the lowest mark (1) meant “not satisfied at all” and the highest (10) was “extremely satisfied.” In general, respondents were moderately satisfied with disseminating their research through social media channels, and the reported average level was 5.3. We found no significant differences between groups.