Introduction

Research data management is a central part of the European Open Science Agenda in the field of research and innovation. For the European Commission (EC),

Open Science is the transformation, opening up and democratisation of science, research and innovation, with the objective of making science more efficient, transparent and interdisciplinary, of changing the interaction between science and society, and of enabling broader societal impact and innovation (Ramjoué 2015, p.169).

Open research data is one of the key components of the emerging new ecosystem of standards and services, and the EC priorities include raising awareness regarding data management, interoperability of infrastructure and datasets, and re-usability of data (ibid.). In France, the Conference of University Presidents put the issue of research data preservation and sharing at the top of their priorities during their annual conference in 2015, the Ministry of Higher Education and Research supports and promotes related actions, and all major public research organizations such as the CNRS (Centre national de la recherche scientifique) contribute to the development of data infrastructures and repositories.

The main issues and objectives are the same as in Germany or other countries, i.e. long-term preservation of scientific output and a global policy of open data in order to increase transparency and stimulate research, innovation and economic activity. There is a general consensus about the complexity and the diversity of the field. Not only are research data difficult to define but their handling and processing largely depends on institutional and disciplinary practices and behaviours; obviously, the term research data

must always be viewed in relation to a particular subject discipline (…) all requirements for the management and long-term availability of research data must be differentiated from each other in regard to both general and discipline-specific aspects and solutions (and) thus far, there is no general agreement on the definition of digital curation, not only in Germany, but on international levels as well (Neuroth et al. 2013, p.11);

one size does not fit all. Also, organizational or national policies and projects (top-down) require support and back-up from the scientific communities and local structures (bottom-up) to meet the scientists’ needs with success.

To support the step-by-step development of sound research data management practices, you must first understand researchers’ needs and perspectives (Ward et al. 2011, p.265). At Lille, we started in 2013 to work in the field of research data management, in particular on the social sciences and humanities campus (Lille 3) with 19,000 students, 580 PhD students, 850 academic and 600 technical staff, in order to gather empirical evidence for the development of data literacy programs and new library-based data services and to raise awareness among scholars and PhD students. This research is conducted together with the GERiiCO laboratory (information, communication and cultural studies), the graduate school and the academic library, with the support of the University’s research department. The following study is work in progress. We present the results of a campus-wide survey on data practices and needs in 2015 and compare them with other studies, including those from Humboldt-Universität zu Berlin (Simukovic et al. 2013, 2014). The complete results have already been published on our institutional repository HAL-Lille 3 (Prost & Schöpfel 2015). The further reading section contains other papers on our research work.

Methodology

The campus-wide survey was conducted in April and May 2015 (five weeks). The questionnaire contained 22 questions adapted from the survey from Humboldt-Universität zu Berlin mentioned above, comprising six sections: information about the respondent, data typology (sources and results), preservation and backup behaviour, sharing behaviour, opinion and motivation regarding data sharing and data repositories, data-related needs. The questionnaire (in French) was communicated to the whole research community on the social sciences and humanities campus (1,800 persons) in the form of an anonymous online version on a local LimeSurvey server. The data analysis and interpretation was done between June and August 2015, and the data were compared to other results from Berlin (Simukovic et al. 2013, 2014), Strasbourg (Rege 2015, Rebouillat 2015), Iowa (Averkamp et al. 2014), LIBER (Reilly et al. 2011), the European Commission (Kuipers & van der Hoeven 2009, see also Herb 2015) and Austria (Bauer et al. 2015).

Results

Response rate

The survey received 270 responses, equivalent to 15% of the whole sample of scientists, scholars, PhD students, administrative and technical staff (research management, technical support services). All scientific departments and research laboratories are represented. Larger and representative sub-samples are from psychology, history, education, information and communication sciences.

Also, all professional groups took part in the survey. The largest group of sub-samples are PhD students (n=73) and senior lecturers (n=69), followed by professors (n=40), scientists (n=16) and other staff (n=13). But the most representative group are professors (26%), followed by senior lecturers (17%) and PhD students (13%).

All respondents were asked to answer to the whole list of 22 questions but no question was obligatory. Multiple answers were allowed for most questions. As a result, no question received 100% answers, and the response rate per question varies between 12% and 82%.

Current research data management

One part of the questions was about research data behaviour – which kind of data are used and produced, how they are stored, preserved, safeguarded; and how they are shared with other researchers. The responses show a wide variety of practice and usage.

Data sources

The survey identified text documents as the most important data source, followed by observations, interviews and survey data (figure 1).

Figure 1: Research data sources (n=214)

Other, less important data sources are experiments, archival material, statistics, images, and audio and video recordings, while log files or simulations are missing, at least in our sample.

Data types

Figure 2 shows the research data produced by the respondents. Again, text documents are the most important output type, followed by spreadsheets, databases, multi-dimensional visualisations and models.

Figure 2: Research data types (n=211)

Other data types like audio and video recordings, images, maps or software are less important.

Data storage

Most of the respondents prefer local solutions for the storage of their research data, either the hard disk of their private personal computer (83%), or their professional work-station (49%) (figure 3).

Figure 3: Storage places (n=212)

Only 12% store their data on a campus-based server. 19% conserve their data in the cloud, on a distant server (gratis or not), and 8% store them on another institution’s server, probably in collaborative, multi-site research projects.

55% say that they store their data in two or more different places, which means usually on the personal (private) computer and on the work-station on the campus. 76% say that they also make use of other devices, most often external hard drives (61%) and USB flash drives (46%) while other devices such as CD or DVD are mentioned by few.

How much memory space do they occupy? 38% of the respondents do not know exactly. 51% estimate their storage size at up to 100 gigabytes while only 13 respondents (half of them are psychologists) mentioned 1 terabytes or more. While 28% state that they back up the data on a daily schedule, nearly as many (26%) do it on an irregular basis, each time when needed or from time to time. Nearly all of them (97%) say that they are the only people in charge of the preservation of their research data and that there is no data officer or at least another person specifically designated to perform this task.

Taken together and as a whole, these responses appear to reflect a more or less individual data behaviour, private rather than professional, with available, often private (and cheap) devices but caring for security and backup. We cannot say if the respondents already faced serious problems with their research data (loss of data, security breach etc.); probably they never have because they do not seem really unsatisfied with this situation.

Most respondents (64%) do not share their research data with colleagues or other people (figure 4). Nobody, except themselves, have access to their data files.

Figure 4: Access to research data (n=209)

The survey does not really identify one or two priorities. The answers rather reflect a set of needs, some more important (storage space, general advice for data management), others a little bit less (technical and legal issues, publishing and citing). One third mentioned ethical issues and help for the preparation of data management plans (DMP).

Open access

We also tried to measure the colleagues’ general willingness to deposit and share their data. Are they motivated? Which kind of data would they deposit? Would they publish their data along with articles? Which data repository would they prefer?

About 40% of the respondents express a positive opinion about data sharing. Either they have already deposited their data in a repository (16%) or they intend to do so in the future (25%). 30% admit that they were not aware of this possibility (figure 6).

Figure 6: Deposit of research data in a data repository (n=162)

Another third (29%) clearly say that they never deposited their research data in the past and that they have no intention of doing so in the future. A deposit in a data repository is no option for five reasons: sensitive and confidential data, risk of plagiarism, workload, data illegibility (nobody would understand my raw data) and intellectual property (these are MY data).

This rejection of data sharing significantly decreases when it comes to the question of publishing data along with an article, i.e. when data sharing is incentive or obligatory (figure 7). Here, only 7% declare that they never shared their data in the past and that they will not share them in the future.

Figure 7: Data publishing (n=147)

44% of the respondents state that they have already published data together with an article, and 31% announce that they intend to do so in the future while 18% admit that they simply did not know about this possibility.

We also asked when and which kind of data they would share. Again, only 12% clearly state that they will not deposit or share their research results in this way (figure 8).

Figure 8: Data deposit (n=187)

37% underline that they would share those data asked for by their peers while others say that they would deposit data produced in collaborative research projects (33%) or with public funding (25%). The last question was about the kind of repository they would prefer for the deposit and sharing of their research data (figure 9). Even without a clear preference (many respondents mark more than one answer), we can identify a ranking of preferred options.

Figure 9: Preferred data repository (n=173)

The academic library plays a central part in particular in education, advice and assistance. The library will also be partner to a series of in-depth interviews with a smaller sample of scientists (n=30-50) in order to learn more about data behaviour, best practices and data-related needs. And the academic librarians’ contribution is expected and necessary also for the development of infrastructures and tools.

Acknowledgments

The study was supported by the Department of International Relations of the University of Lille 3 and the European Institute for Social Sciences and Humanities at Lille. We would like to thank all colleagues who contributed to the success of the survey, in particular Peter Schirmbacher, Maxi Kindling and Elena Simukovic (Berlin), Isabelle Westeel, Cécile Malleret and Stéphane Chaudiron (Lille).

References

Averkamp, S., Gu, X., Rogers, B., 2014. Data management at the University of Iowa: A university libraries report on campus research data needs. University of Iowa. http://ir.uiowa.edu/lib_pubs/153/

Bauer, B., Ferus, A., Gorraiz, J., Gumpenberger, C., Gründhammer, V., Maly, N., Mühlegger, J. M., Preza, J. L., Sánchez Solís, B., Schmidt, N., Steineder, C., 2015. Researchers and their data. Results of an Austrian survey - report 2015. e-infrastructures austria, Vienna. http://e-infrastructures.at/das-projekt/deliverables/

Chaudiron, S., Maignant, C., Schöpfel, J., Westeel, I., 2015. Livre blanc sur les données de la recherche dans les thèses de doctorat. Université de Lille 3, Villeneuve d’Ascq. http://hal.univ-lille3.fr/hal-01192930

Herb, U., 2015. Open Science in der Soziologie. Eine interdisziplinäre Bestandsaufnahme zur offenen Wissenschaft und eine Untersuchung ihrer Verbreitung in der Soziologie. vwh Verlag Werner Hülsbusch, Glückstadt.

Kuipers, T., van der Hoeven, J., 2009. Insight into digital preservation of research output in Europe. Survey report. PARSE insight, European Commission, Brussels. http://www.parse-insight.eu/downloads/PARSE-Insight_D3-4_SurveyReport_final_hq.pdf

Neuroth, H., Strathmann, S., Oßwald, A., Ludwig, J. (Eds.), 2013. Digital curation of research data. Experiences of a baseline study in Germany. vwh Verlag Werner Hülsbusch, Glückstadt. http://www.nestor.sub.uni-goettingen.de/bestandsaufnahme/Digital_Curation.pdf

Prost, H., Schöpfel, J., 2015. Les données de la recherche en SHS. Une enquête à l’Université de Lille 3. Rapport final. Université de Lille 3, Villeneuve d’Ascq. http://hal.univ-lille3.fr/hal-01198379v1

Ramjoué, C., 2015. Towards Open Science: The vision of the European Commission. Information Services & Use 35 (3), 167-170. http://doi.org/10.3233/isu-150777

Rebouillat, V., 2015. Archives ouvertes de la connaissance : Valoriser et diffuser les données de recherche. Master’s thesis, ENSSIB, Villeurbanne. http://cataloguebib.enssib.fr/cgi-bin/koha/opac-detail.pl?biblionumber=1903

Rege, A., 2015. Retour sur l’enquête sur les pratiques de publication scientifique et de production de données de la recherche. In: Université de Strasbourg, Projet AOC, COPIL 27 mars 2015.

Reilly, S., Schallier, W., Schrimpf, S., Smit, E., Wilkinson, M., 2011. Report on integration of data and publications. ODE Opportunities for Data Exchange, The Hague. http://www.stm-assoc.org/2011_12_5_ODE_Report_On_Integration_of_Data_and_Publications.pdf

Simukovic, E., Kindling, M., Schirmbacher, P., 2013. Umfrage zum Umgang mit digitalen Forschungsdaten an der Humboldt-Universität zu Berlin. Report, Humboldt-Universität zu Berlin, Zentraleinrichtung Computer- und Medienservice (Rechenzentrum), Berlin. http://edoc.hu-berlin.de/docviews/abstract.php?id=40341

Simukovic, E., Kindling, M., Schirmbacher, P., 2014. Unveiling research data stocks: A case of Humboldt-Universität zu Berlin. In: iConference, 4-7 March 2014, Berlin. pp. 742-748. https://www.ideals.illinois.edu/handle/2142/47259

Tenopir, C., Allard, S., Douglass, K., Aydinoglu, A. U., Wu, L., Read, E., Manoff, M., Frame, M., 2011. Data sharing by scientists: Practices and perceptions. PLoS ONE 6 (6), e21101+. http://doi.org/10.1371/journal.pone.0021101

Tenopir, C., Dalton, E. D., Allard, S., Frame, M., Pjesivac, I., Birch, B., Pollock, D., Dorsett, K., 2015. Changes in data sharing and data reuse practices and perceptions among scientists worldwide. PLoS ONE 10 (8), e0134826+. http://doi.org/10.1371/journal.pone.0134826

Ward, C., Freiman, L., Jones, S., Molloy, L., Snow, K., Jul. 2011. Making sense: Talking data management with researchers. International Journal of Digital Curation 6 (2), 265-273. http://doi.org/10.2218/ijdc.v6i2.202

Research data management in social sciences and humanities: A survey at the University of Lille (France)