Last modified: 2018-05-25
Abstract
queries concern the assessment of causal relationship in individual cases by evaluating the probability of causation. However is not always clear how and whether, to usefully employ scientific
data for this pourpose.
Given even a randomized sample we can typically only provide bounds for the
probability of causationÂ
if some fundamental conditions, namely exogeneity, comparability and sufficiency are satisfied. In this work we make the fundamental
conditions operative by means of a Bayesian model
selection procedure.
References
Chen, J. and Chen, Z., Extended Bayesian Information Criteria for Model Selection with large Model Spaces. Biometrika, 95, 3, 759-771, 2008.
Dawid, A.P., The role of scientific and statistical evidence in assessing
causality. In Perspectives on Causation}, (ed. R.Goldberg), 133-147. Hart Publishing, Oxford, 2011.
Dawid, A.P., Faigman, D.L., and Fienberg, S.E., Fitting science into legal contexts: Assessing effects of causes or causes of effects?\ (with {D}iscussion and authors' rejoinder).Sociological Methods and Research}, 43 359--421, 2014.
Dawid, A.P., Musio, M. and Fienberg, S.E.,
From Statistical Evidence to Evidence of Causality. \emph{Bayesian Analysis}, 11, 725-752, 2016.
Dawid, A.P., Musio, M. and Murtas, R., The Probability of Causation, Law, Probability and Risk, 16, 4, 163-179, 2017.
Holland, P.W., Statistics and Causal Inference Journal of the American Statistical Association 81, 396, 945-960, 1986.
Rubin, D.B., Estimating Causal Effects of Treatments in Randomized and Nonrandomized Studies. Journal of Educational Psychology, 66, 688--701, 1974.