Research on interventions into urban problems has traditionally been organized around the question “What works?” If we are going to make research more useful, our questions need to evolve. In environments that may be low on social capital and resources and high on political conflict, many good ideas will fail because of the hazards of implementation in toxic environments. The question which organizes our thinking has to become something like “How does one navigate the social and political context in such a way that what can work does work?” Research that tries to understand both the outcomes of social interventions and the process of implementing them at a high level goes by different names in different disciplines, but one rapidly-growing approach is called improvement science (Bryk et al. 2015). The central tenet is that research should involve a continuous process of learning how to get better.
Instead of only researchers involved in the design and execution of research, improvement science argues that researchers should work closely at all stages with stakeholders, practitioners, and those affected by the problem under study. The latter groups often have a more nuanced understanding of the context than researchers.
Bryk, Anthony S., Louis M. Gomez, Alicia Grunow, and Paul G. LeMahieu. Learning to improve: How America’s schools can get better at getting better. Harvard Education Press, 2015.