It’s hard to assess the climate impacts of digital technologies. Impacts can be direct, indirect and systemic – where is the system boundary drawn that defines what’s counted and what’s not? In new work led by Vlad Coroama at the Roegen Centre for Sustainability and involving iDODDLE PI, Charlie Wilson, we use simple statistical techniques on a pooled sample of 104 impact estimates from 9 digital applications to show that certain study design characteristics significantly affect outcomes. For example, we show impact estimates were significantly less beneficial if their corresponding studies included indirect effects (not just direct ICT operational footprints), rebound effects, or induced demand and other systemic effects within study system boundaries. We also found impact estimates were significantly less beneficial if their corresponding studies were conducted by independent authors, not by consultants or by industry actors – emphasising the important role for impartial, rigorous academic research in this field.
Take home = study design matters, so read the methods section before taking impact estimates at face value!
