This paper is based on the President's Lecture, given at the Biennial Meeting of the Italian Statistical Society, Bari, June 2004.
Causality and Causal Models: A Conceptual Perspective†
Article first published online: 8 FEB 2007
DOI: 10.1111/j.1751-5823.2006.tb00298.x
Additional Information
How to Cite
Frosini, B. V. (2006), Causality and Causal Models: A Conceptual Perspective. International Statistical Review, 74: 305–334. doi: 10.1111/j.1751-5823.2006.tb00298.x
- †
Publication History
- Issue published online: 8 FEB 2007
- Article first published online: 8 FEB 2007
- [Received October 2004, accepted June 2006]
- Abstract
- References
- Cited By
Keywords:
- Causality;
- Causal models;
- Directed acyclic graph;
- Confounder;
- Counterfactual
Summary
This paper aims at displaying a synthetic view of the historical development and the current research concerning causal relationships, starting from the Aristotelian doctrine of causes, following with the main philosophical streams until the middle of the twentieth century, and commenting on the present intensive research work in the statistical domain. The philosophical survey dwells upon various concepts of cause, and some attempts towards picking out spurious causes. Concerning statistical modelling, factorial models and directed acyclic graphs are examined and compared. Special attention is devoted to randomization and pseudo-randomization (for observational studies) in view of avoiding the effect of possible confounders. An outline of the most common problems and pitfalls, encountered in modelling empirical data, closes the paper, with a warning to be very cautious in modelling and inferring conditional independence between variables.
Le but de cet article est d'offrir une vue d'ensemble sur le thème des relations causales, à partir de la doctrine philosophique aristotélique, et ensuite étendues et formalisées dans le champ de l'analyse statistique multivarée. Dans la revue philosophique on analyse plusieurs conceptions de cause, et les essais de reconnâtre les causes ”fausses”. La partie centrale du travail s'occupe de modèles causals en forme graphique, qui constituent l'instrument électif de plusieurs recherches causales, et met en evidence la différence entre conditionnement et intervention sur une variable. On a dedié une particulière attention aux procédures de randomization dans le but d'éviter de possible confusions. L'article termine en conseillant d'user de la prudence dans la modelage de l'independence conditionnelle et dans son contrôl empirique.

1751-5823/asset/olbannerleft.gif?v=1&s=c4a445a9d0699bd941ad371812c6a7d9622aed96)
1751-5823/asset/olbannerright.gif?v=1&s=79cf6f118b8497650eddf7fab055f12f63cb785d)
1751-5823/asset/cover.gif?v=1&s=22db79b95bf94617b28b9996a484472e2a49b5c9)