Introduction The introduction of Bayesian networks (Pearl 1986b) and associated local computation algorithms (Lauritzen and Spiegelhalter 1988, Shenoy and Shafer 1990, Jensen, Lauritzen and Olesen 1990) has initiated a renewed interest for understanding causal concepts in connection with modelling complex stochastic systems. It has become clear that graphical models, in particular those based upon directed acyclic graphs, have natural causal interpretations and thus form a base for a language in which causal concepts can be discussed and analysed in precise terms. As a consequence there has been an explosion of writings, not primarily within mainstream statistical literature, concerned with the exploitation of this language to clarify and extend causal concepts. Among these we mention in particular books by Spirtes, Glymour and Scheines (1993), Shafer (1996), and Pearl (2000) as well as the collection of papers in Glymour and Cooper (1999). Very briefly, but fundamentally,

Evaluation of counterfactual queries (e.g., "If A were true, would C have been true?") is important to fault diagnosis, planning, and determination of liability. In this paper we present methods for computing the probabilities of such queries using the formulation proposed in [Balke and Pearl, 1994], where the antecedent of the query is interpreted as an external action that forces the proposition A to be true. When a prior probability is available on the causal mechanisms governing the domain, counterfactual probabilities can be evaluated precisely. However, when causal knowledge is specified as conditional probabilities on the observables, only bounds can computed. This paper develops techniques for evaluating these bounds, and demonstrates their use in two applications: (1) the determination of treatment efficacy from studies in which subjects may choose their own treatment, and (2) the determination of liability in product-safety litigation. 1 INTRODUCTION A counterfactual sente...

Evaluation of counterfactual queries (e.g., "If A were true, would C have been true?") is important to fault diagnosis, planning, and determination of liability. We present a formalism that uses probabilistic causal networks to evaluate one's belief that the counterfactual consequent, C, would have been true if the antecedent, A, were true. The antecedent of the query is interpreted as an external action that forces the proposition A to be true, which is consistent with Lewis' Miraculous Analysis. This formalism offers a concrete embodiment of the "closest world" approach which (1) properly reflects common understanding of causal influences, (2) deals with the uncertainties inherent in the world, and (3) is amenable to machine representation. Introduction A counterfactual sentence has the form If A were true, then C would have been true where A, the counterfactual antecedent, specifies an event that is contrary to one's real-world observations, and C, the counterfactual consequen...

Structural equation modeling (SEM) has dominated causal analysis in the social and behavioral sciences since the 1960s. Currently, many SEM practitioners are having difficulty articulating the causal content of SEM and are seeking foundational answers.

Abstract not found

This paper does not attempt to correct for shortcomings of the traditional IV method but, rather, to develop a complementary method which can provide unbiased estimates under conditions where the IV method fails. The method relies on finding an auxiliary variable Z

Indirect experiments are studies in which randomized control is replaced by randomized encouragement, that is, subjects are encouraged, rather than forced to receive treatment programs. The purpose of this paper is to bring to the attention of experimental researchers simple mathematical results that enable us to assess, from indirect experiments, the strength with which causal influences operate among variables of interest. The results reveal that despite the laxity of the encouraging instrument, indirect experimentation can yield significant and sometimes accurate information on the impact of a program on the population as a whole, as well as on the particular individuals who participated in the program. Keywords: Causal reasoning, treatment evaluation, noncompliance, graphical models 1 Introduction Standard experimental studies in the biological, medical, and behavioral sciences invariably invoke the instrument of randomized control, that is, subjects are assigned at random to va...

Some of the main users of statistical methods -- economists, social scientists, and epidemiologists -- are discovering that their fields rest not on statistical but on causal foundations. The blurring of these foundations over the years follows from the lack of mathematical notation capable of distinguishing causal from equational relationships. By providing formal and natural explication of such relations, graphical methods have the potential to revolutionize how statistics is used in knowledgerich applications. Statisticians, in response, are beginning to realize that causality is not a metaphysical deadend but a meaningful concept with clear mathematical underpinning. The paper surveys these developments and outlines future challenges.

In this paper we study the identifiability of nonparametric models, that is, models in which both the functional forms of the equations and the probability distributions of the disturbances remain unspecified. Identifiability in such models does not mean uniqueness of parameters but rather uniqueness of the set of predictions of interest to the investigator. For example, predicting the effects of changes, interventions, and control. We provide sufficient and necessary conditions for identifying a set of causal predictions of the type: "Find the distribution of Y , assuming that X is controlled by external intervention", where Y and X are arbitrary variables of interest. Whenever identifiable, such predictions can be expressed in closed algebraic form, in terms of observed distributions. We also show how the identifying criteria can be verified qualitatively, by inspection, using the graphical representation of the structural model. When compared to standard identifiability tests of lin...

) (9.75%) No Cancer 57 19 2 342 59 361 Two opposing interpretations have been offered for these data. The advocates of antismoking legislation argue that the table proves the harmful effect of smoking. They point to the fact that about 85% of the smokers have developed lung cancer, compared to only 9.75% of the non-smokers. Moreover, within each of two subgroups, tar and no tar, smokers show a much higher percentage of cancer than non-smokers. However, the tobacco industry argues that the table tells an entirely different story -- that smoking would actually decrease, not increase, one's risk of lung cancer. Their argument goes as follows: If you choose to smoke, then your chances of building up tar deposits are 95%, compared to 5% if you choose not to smoke (380/400 versus 20/400). To evaluate the effect of tar deposits, we look separately at two groups, smokers and non-smokers, as done in the table below. All numbers are given in thousands. On Bottles and Drugs<F30