Replicate and generalize to make urban research coherent

Scientific approaches to urban research are a century in the making but have long remained tangled in debates and disagreements. Here, we reflect on how current enthusiasm for a scientific urbanism resurfaces old conflicts and suggest that a diverse set of arguments revolve around two concepts: replication and generalizability. We argue that addressing these challenges directly could play an important role in advancing empirical research. New empirical methods, including those coming from artificial intelligence, offer means of radically accelerating urban research, but also risk a revisiting of longstanding disagreements. Applying replication and generalizability as a conceptual bridge may be key to some of these new methods contributing to a more integrated and interdisciplinary urbanism. Beyond empirical urbanism, replication and generalizability may enable the development of a more integrated, systematic urban field by supporting the translation of methodologies, findings, and theories between disciplines. We conclude by cautioning that further work is needed to understand how replication and generalizability apply across the urban field (significant areas of which are not addressed in this paper) and by calling for authors to test the value of replication and generalizability in their areas of urban expertise.