“Applied economists have become enthusiastic users of natural experiments to generate exogenous variation in key independent variables although it has been argued that this approach has severe limitations …” (more)
[Geary Behavioural Economics Blog, 26 July]