Abstract: Despite continued calls for data sharing and replication in management and social science research, there remains a gap between espoused values (sharing/replication is 'good') and revealed preference (we don't share/replicate). Why is this the case, and can anything be done about it? We identify and address incentive issues by adapting and extending algorithms for synthetic data generation for use in management and social science research. Simulation results and application to actual data sets demonstrate the potential of these methods to enable researchers to produce and share synthetic data, thereby promoting replication, extension, and ultimately, knowledge generation, while removing constraints and disincentives of sharing authentic data.

Nigel Melville is an Assistant Professor of Business Information Technology at the Stephen M. Ross School of Business, University of Michigan, and a Special Sworn Status researcher of the US Census Bureau. His research focuses on the organizational performance impacts of IT, IT and innovation, digital commons problems, and epistemology. His professional experience includes new product development and R&D with Motorola, applied semiconductor research at Lawrence Berkeley National Laboratory, and co-founding a CRM software company. The common theme was the application of information and information technology to create new value for organizations, which is the focus of his research.