Data are expensive and sometimes irreplaceable. And yet, the tradition in much of science is that data are not broadly shared. As a result, these data are not used to their full potential. Through global public data archiving, we can verify existing findings, reduce duplicate effort, facilitate meta-analysis, and answer new scientific questions. Funders and journals are now implementing stronger data sharing policies, and these sometimes present challenges to individual researchers. To allay concerns, it is crucial that we be able to document the benefits and costs of data-sharing. What are scientists' concerns about sharing data? How does limited distribution of datasets impede scientific progress? What opportunities for transparency, discovery, broadened participation, and training do we forgo as a scientific community? Are these benefits hypothetical or demonstrable? Can we realize the community benefits without sacrificing the interests of investigators who collect data? This session will address the costs and benefits of public data archiving through examination of both personal experience and large-scale quantitative analysis. Several domains will be considered, with particular highlights on the implications of data sharing and withholding on cancer research and the Gulf of Mexico oil spill.