Causality Study Fortnight

We address the issue of causation in demography, focusing. particularly onchange through time in aggregate demographic phenomena. We regard the interventionframework associated with the Neyman-Rubin-Holland model of causal estimation as toorarely applicable to be useful as a general solution to causal inference in demography.Discussion centres on a series of examples of demographic change ranging frommechanical demographic processes, where causation is known to occur, through torelatively straightforward cases where causation is not, or only rarely, contested to morecomplex demographic phenomena. We develop and propose ten criteria, or guidelines,intended to assist in the practical evaluation of causation in the context of demographicchange, analogous to those in use in epidemiology. Machine learning and causality (David Corfield & Alex Freitas)