Why Should Data Sharing (Not) be Rewarded?

In the last decade, there has been an intense scholarly and policy movement to promote (open) data sharing in science to provide or improve scientific rigor, productivity, and transparency. However, there has been a slow uptake of this practice due to mainly a lack of incentives and rewards for researchers. In order to counterbalance the lack of rewards for data sharing, some researchers and research groups are deploying some strategies to compensate for their efforts in sharing data. For instance, researchers publish their data sets in 'data papers' –a scholarly publication describing a particular dataset or collection of data sets without including any analysis. This way, authors increase their list of publications through 'data authorship'; some researchers actively offer their data by imposing different 'knowledge control regimes' (Hilgartner, 2017), which range from citation requests to questionable 'guest authorship' practices (Jabbehdari & Walsh, 2017; Laudel, 2002). I suggest that these practices are not only a consequence of researchers' path dependency on rewards but also a conceptualization of research data as 'products' of research. Considering science's (non)epistemic goals (Potochnik, 2017) and drawing on archival science, we argue that data should not be considered 'products' or 'outputs' but 'by-products' of research.

The goal of this presentation is to stimulate a critical discussion on whether data sharing should, –or should not, be rewarded. First, I will present some empirical findings and hypothetical consequences of researchers' (questionable) current strategies to counterbalance the lack of rewards for data sharing and how these strategies might aggravate Matthew's effect. Second, I will expound my arguments for considering data 'by-products' of research and when and who, if any, should be rewarded for data sharing by considering issues of division of labor in data collection, data management and data curation.