|
Abstract:
|
Accessing and combining large amounts of data is attractive for social scientists. The increasing amount of data also increases privacy risks. There is currently an extreme pressure from society to increase data protection. Several important players in official statistics, academia, and business see differential privacy as the solution. In this opinion piece we put differential privacy in a larger context and discusses from a social science research perspective pros and cons for adapting differential privacy. It becomes clear that social science research work flow must change if differential privacy is implemented. It also becomes clear that common social science data collection will become more costly. However, there are a series of positive side effects in implementing the approach aside from preserving privacy that could solve some issues social scientists grabble with these days. We conclude with an assessment on what seems to be a reasonable approach in the short term, given current technology and point out why we think that collecting data with the promise of using it in a differentially private way will likely not change the participation decision of the public, but might help
|
