Barker et al 2015 (https://doi.org/10.1145/2676723.2677282) examined what influences CS faculty to adopt new teaching practices. It had a big influence on my thinking when I was working with @thecarpentries ; what I'm looking for now is a similar study of what influences research scientists to adopt better software and data practices - not what those practices are or what their benefits are, but the factors that lead to, or impede, actual adoption. Pointers would be greatly appreciated.