Table of Contents:
In November 2014, an online survey on research data management was conducted at Philipps-Universität Marburg. The aim of the survey is to provide an overview of existing data, experiences and practices of handling data, data cultures, and service needs. The survey is an important milestone of the project "Kompetenzzentrum Forschungsdatenmanagement und -archivierung", in which the University Computer Centre and the University Library have been working together since 2013. The target group of the survey included all professors and scientific staff of all 16 faculties and all academic centres. The overall response rate was 18.3% (professors: 32.3%, scientific staff: 15.8%). For analytical reasons, faculties were grouped into three clusters called "Humanities, Social Sciences and Economics", "Natural Sciences", and "Medicine". Central service facilities and academic centres were categorised as "Centres and Facilities". The distribution of survey participants among the clusters is representative of the university's employment structure. Results clearly demonstrate that research data are an important topic for scientists. They have recognizable experience in research data management and expect an increase in importance in the near future. Members of scientific staff rate the importance of research data and their own experience in handling research data slightly higher than professors. Already, scientists spend a considerable amount of their working time on managing research data. It is also evident that research data management is an important issue in graduate education, as more than half of the respondents undertake graduate studies. All scientific disciplines work with research data as texts, tables, and images. Looking at the formats used, it can be assumed that the majority of data can be kept and re-used for at least 10 years. However, in terms of storage and re-use, there is a need for action in research fields relying on experimental devices, i. e. natural sciences and medicine. Results also show that research data currently held at the university can be considered "small data": The majority of researchers require up to 20 GB of storage space for ongoing as well as for completed projects. Predominantly, data are stored and secured locally as well as according to individual specifications. Existing central server storage and regular backup services are used to a considerably lesser extent. Use of commercial cloud services indicates that easy ways of accessing and exchanging data are much appreciated. Rights management control by users may render these solutions even more attractive, since control over one's own research data is of central importance: With increasing openness of a system, the willingness to use it for research data decreases. Likewise, there is a difference in the readiness to use an institutional repository at Philipps-Universität Marburg, depending on whether data will only be stored or actually shared. Nonetheless, 40% of scientists participating in the survey would consider practising the latter. Notably, the survey demonstrated that there is no clear dichotomy between scientific disciplines in the sense that scientists working in humanities, social sciences and economics have less experience in handling research data or articulate other interests than scientists working in natural sciences or medicine. On the contrary, coinciding needs were expressed across scientific disciplines, both concerning requested research data management services and important functions of a research data repository. In detail, a majority of respondents would like advice and support for technical and legal questions and for general aspects of handling research data; features of a research data repository should include secure storage for at least 10 years, a quality check of data, and protection mechanisms for sensitive data. Beyond that, interests among clusters differ, which is what can be expected due to the nature of Philipps-Universität Marburg as a university with a broad range of scientific disciplines.