Data sharing is the provision of a public good. A public good can be used by "everyone" (non-exclusiveness) and does not reduce the availability to others (non-rivalry). To share research data, one has to invest time and resources into data management to ensure reusability of the data (costs). Scientists also partly give up future research ideas as others can work on the same ideas when the data is public. At the same time, the value of shared data (benefit) increases because of new perspectives of looking at the data. This presentation aims to theoretically analyze motivational factors of data sharing by applying game theory. In a basic public goods game, participants choose how much of their private endowment they want to contribute to a public pot. The contributions in the pot are multiplied and evenly shared among the players. Variations of the public goods game, such as punishments and rewards, were experimentally tested to research motivations for providing public goods. Different game scenarios relate to real world regimes like the UK where researchers are "punished" when they do not share their data, or Germany where funders have followed a laissez-fair approach so far.