Manipulationism holds that information about the results of interventions is of utmost importance for scientific practices such as causal assessment or explanation. Specifically, manipulation provides information about the stability, or invariance, of the relationship between X and Y: were we to wiggle the cause X, the effect Y would accordingly wiggle and, additionally, the relation between the two will not be disrupted. This sort of relationship between variables are called 'invariant empirical generalisations'. The paper focuses on questions about causal assessment (...) and analyses the status of manipulation. It is argued that manipulationism is trapped in a dilemma. If manipulationism is read as providing a conceptual analysis of causation, then it fails to provide a story about the methods for causal assessment. If, instead, manipulationism is read as providing a method for causal assessment, then it is at an impasse concerning causal assessment in areas where manipulations are not performed. Empirical generalisations are then reassessed, in such a way that manipulation is not taken as methodologically fundamental. The paper concludes that manipulation is the appropriate tool for some scientific contexts, but not for all. (shrink)

The purpose of this essay is to defend the deductive-nomological model of explanation against a number of criticisms that have been made of it. It has traditionally been thought that scientific explanations were causal and that scientific explanations involved deduction from laws. In recent years, however, this three-fold identity has been challenged: there are, it is argued, causal explanations that are not scientific, scientific explanations that are not deductive, deductions from laws that are neither causal explanations nor scientific explanations, and (...) causal explanations that involve no deductions from laws. The aim of the present essay is to defend the traditional identities, and to show that the more recent attempts at invalidating them fail in their object. More specifically, this essay argues that a Humean version of the deductive-nomological model of explanation can be defended as the correct account of scientific explanation of individual facts and processes, and as the correct account of causal explanations of individual facts and processes. The deductive-nomological model holds that to explain an event E, say that a is G, one must find some initial conditions C, say that a is F, and a law or theory T such that T and C jointly entail E, and both are essential to the deduction. (shrink)

One way social scientists explain phenomena is by building structural models. These models are explanatory insofar as they manage to perform a recursive decomposition on an initial multivariate probability distribution, which can be interpreted as a mechanism. Explanations in social sciences share important aspects that have been highlighted in the mechanisms literature. Notably, spelling out the functioning the mechanism gives it explanatory power. Thus social scientists should choose the variables to include in the model on the basis of their function (...) in the mechanism. This paper examines the notion of ‘function’ within structural modelling. We argue that ‘functions’ ought to be understood as the theoretical underpinnings of the causes, namely as the role that causes play in the functioning of the mechanism. (shrink)

This paper is an assessment of an attempt, by James Greeno, to measure the explanatory power of statistical theories by means of the notion of transmitted information (It). It is argued that It has certain features that are inappropriate in a measure of explanatory power. In particular, given a statistical theory T with explanans variables St and explanandum variables Mj, it is argued that no plausible measure of explanatory power should depend on the probability P(Si) of occurrence of initial conditions (...) in the systems to which T applies or the magnitudes of the conditional probabilities P(Mj/Si), in the manner in which Ir does. (shrink)

This paper is a realist argument for the existence of “social objects”. Social objects, I argue, are the outcome states of a contingent causal process and in turn posses causal properties. This argument has consequences for what we can mean by realism and consequences for the development of a realist methodology. Realism should abandon the notion of natural necessity in favour of a view that the “real” nature of the social world is contingent and necessity is only revealed in outcome (...) states. This, I argue, has both theoretical and methodological implications and I develop my argument through two case studies, of homelessness and ethnicity. (shrink)

In recent years, realism?particularly critical realism?has become an important philosophical and methodological foundation for social science. A key feature is that of natural necessity, but this coexists alongside an acceptance of contingency in the social world. I argue in this paper that there cannot be any natural necessity in the social world, but rather the real nature of the social world is that it is contingent. This need not lead to an abandonment of realism, and indeed I argue that a (...) contingent form of realism allows us to bring ontological probability back into realism. (shrink)

The computer’s effect on our understanding of causation has been enormous. By the mid‐1980s, philosophical and social‐scientific work on the topic had left us with (1) no reasonable reductive account of causation and (2) a class of statistical causal models tied to linear regression. At this time, computer scientists were attacking the problem of equipping robots with models of the external that included probabilistic portrayals of uncertainty. To solve the problem of efficiently storing such knowledge, they introduced Bayes Networks and (...) directed graphs. By attaching a causal interpretation to Bayes Networks, the philosophy of causation changed dramatically. We are now able to be extremely general about how causal structure connects to data, and systematic about when causal structures are empirically indistinguishable. In this essay I try to motivate and describe this synthesis. (shrink)

A shared problem across the sciences is to make sense of correlational data coming from observations and/or from experiments. Arguably, this means establishing when correlations are causal and when they are not. This is an old problem in philosophy. This paper, narrowing down the scope to quantitative causal analysis in social science, reformulates the problem in terms of the validity of statistical models. Two strategies to make sense of correlational data are presented: first, a 'structural strategy', the goal of which (...) is to model and test causal structures that explain correlational data; second, a 'manipulationist or interventionist strategy', that hinges upon the notion of invariance under intervention. It is argued that while the former can offer a solution the latter cannot. (shrink)

Causation is a macroscopic phenomenon. The temporal asymmetry displayed by causation must somehow emerge along with other asymmetric macroscopic phenomena like entropy increase and the arrow of radiation. I shall approach this issue by considering ‘causal inference’ techniques that allow causal relations to be inferred from sets of observed correlations. I shall show that these techniques are best explained by a reduction of causation to structures of equations with probabilistically independent exogenous terms. This exogenous probabilistic independence imposes a recursive order (...) on these equations and a consequent distinction between dependent and independent variables that lines up with the temporal asymmetry of causation. (shrink)

Social research, from economics to demography and epidemiology, makes extensive use of statistical models in order to establish causal relations. The question arises as to what guarantees the causal interpretation of such models. In this paper we focus on econometrics and advance the view that causal models are ‘augmented’ statistical models that incorporate important causal information which contributes to their causal interpretation. The primary objective of this paper is to argue that causal claims are established on the basis of a (...) plurality of evidence. We discuss the consequences of ‘evidential pluralism’ in the context of econometric modelling. (shrink)

Causal analysis in the social sciences takes advantage of a variety of methods and of a multi-fold source of information and evidence. This pluralistic methodology and source of information raises the question of whether we should accordingly have a pluralistic metaphysics and epistemology. This paper focuses on epistemology and argues that a pluralistic methodology and evidence don’t entail a pluralistic epistemology. It will be shown that causal models employ a single rationale of testing, based on the notion of variation. Further, (...) I shall argue that this monistic epistemology is also involved in alternative philosophical theories of causation. (shrink)

The experimental revolution in the social sciences is one of the most significant methodological shifts undergone by the field since the ‘quantitative revolution’ in the nineteenth century. One of the often valued features of social science experimentation is precisely the fact that there are clear methodological rules regarding hypothesis testing that come from the methods of the natural sciences and from the methodology of RCTs in the biomedical sciences, and that allow for the adjudication among contentious causal claims. We examine (...) critically this claim and argue that some current understandings of the practices that surround social science experimentation overestimate the degree to which experiments can actually fulfil this role as “objective” adjudicators, by neglecting the importance of shared background knowledge or assumptions and of consensus regarding the validity of the constructs involved in an experiment. We take issue with the way the distinction between internal and external validity is often used to comment on the inferential import of experiments, used both among practitioners and among philosophers of science. We describe the ways in which the more common use of the internal/external distinction differs from Cook and Campbell’s original methodological project, in which construct validity and the four-fold validity typology were all important in assessing the inferential import of experiments. We argue that the current uses of the labels internal and external, as applied to experimental validity, help to encroach a simplistic view on the inferential import of experiments that, in turn, misrepresents their capacity to provide objective knowledge about the causal relations between variables. (shrink)

This paper argues against Papineau's claim that causal relations can be reduced to correlations and defends Cartwright's thesis that they can be nevertheless boot-strapped from them, given sufficiently rich causal background knowledge.

Recent studies on probabilistic causation and statistical explanation (Cartwright 1979; Salmon 1984), I believe, have opened up the possibility of a genuine unification between philosophical approaches and causal modeling (CM) in the social, behavioral and biological sciences (Wright 1934; Blalock 1964; Asher 1976). This unification rests on the statistical tools employed, the principle of common cause, the irreducibility of causation to probability or statistics, and the idea of causal process as a suitable framework for understanding causal relationships. The aim of (...) this paper is to draw attention to these four areas of contact by focusing on the relevant aspects of CM.Causal analysis in the social sciences is based on two fundamental notions: model and method. A causal model is an idealized picture of the causal relationships in the world. Method, on the other hand, refers to certain statistical techniques that are used to evaluate and test a causal model using data which consist of joint observations on the model variables. (shrink)

Causal modeling methods such as path analysis, used in the social and natural sciences, are also highly relevant to philosophical problems of probabilistic causation and statistical explanation. We show how these methods can be effectively used (1) to improve and extend Salmon's S-R basis for statistical explanation, and (2) to repair Cartwright's resolution of Simpson's paradox, clarifying the relationship between statistical and causal claims.

This paper makes explicit and takes issue with the bizarre view, which is unfortunately prevalent among social scientists, that causal relations are features of models only. There are some good reasons to represent causal factors with independent variables. But the association between causes and independent variables is only a desideratum in model construction. It is not a criterion for judging which things are causes and which are effects.

Nancy Cartwright's recent criticisms of efforts and methods to obtain causal information from sample data using automated search are considered. In addition to reviewing that effort, I argue that almost all of her criticisms are false and rest on misreading, overgeneralization, or neglect of the relevant literature.

Drawing substantive conclusions from linear causal models that perform acceptably on statistical tests is unreasonable if it is not known how alternatives fare on these same tests. We describe a computer program, TETRAD, that helps to search rapidly for plausible alternatives to a given causal structure. The program is based on principles from statistics, graph theory, philosophy of science, and artificial intelligence. We describe these principles, discuss how TETRAD employs them, and argue that these principles make TETRAD an effective tool. (...) Finally, we illustrate TETRAD's effectiveness by applying it to a multiple indicator model of Political and Industrial development. A pilot version of the TETRAD program is described in this paper. The current version is described in our forthcoming Discovering Causal Structure: Artificial Intelligence for Statistical Modeling. (shrink)

Two kinds of causal inference rules which are widely used by social scientists are investigated. Two conceptions of causation also widely used are explicated — the INUS and probabilistic conceptions of causation. It is shown that the causal inference rules which link correlation, a kind of partial correlation, and a conception of causation areinvalid. It is concluded anew methodology is required for causal inference.

Philosophers and scientists have maintained that causation, correlation, and partial correlation are essentially related. These views give rise to various rules of causal inference. This essay considers the claims of several philosophers and social scientists for causal systems with dichotomous variables. In section 2 important commonalities and differences are explicated among four major conceptions of correlation. In section 3 it is argued that whether correlation can serve as a measure of A's causal influence on B depends upon the conception of (...) causation being used and upon certain background assumptions. In section 4 five major kinds of partial correlation are explicated, and some of the important relations are established among two conceptions of partial correlation, the conception of screening off, the conception of partitioning, and the measures of causal influence which have been suggested by advocates of path analysis or structural equation methods. In section 5 it is argued that whether any of these five conceptions of partial correlation can serve as a measure of causal influence depends upon the conception of causation being used and upon certain background assumptions.The important conclusion is that each of the approaches (considered here) to causal inference for causal systems with dichotomous variables stands in need of important qualifications and revisions if they are to be justified. (shrink)

During the past decades several philosophers of science and social scientists have been interested in the problems of causation. Recently attention has been given to probabilistic causation in dichotomous causal systems. The paper uses the basic features of probabilistic causation to argue that the causal modeling approaches developed by such researchers as Blalock and Duncan can provide, when an additional assumption is added, adequate qualitative measures of one variableś causal influence upon another. Finally, some of the difficulties and issues involved (...) in developing adequate quantitative measures are discussed, and it is concluded that the causal modeling measures cannot provide adequate quantitative measures. (shrink)

The anti-causal prophecies of last century have been disproved. Causality is neither a ‘relic of a bygone’ nor ‘another fetish of modern science’; it still occupies a large part of the current debate in philosophy and the sciences. This investigation into causal modelling presents the rationale of causality, i.e. the notion that guides causal reasoning in causal modelling. It is argued that causal models are regimented by a rationale of variation, nor of regularity neither invariance, thus breaking down the dominant (...) Human paradigm. The notion of variation is shown to be embedded in the scheme of reasoning behind various causal models: e.g. Rubin’s model, contingency tables, and multilevel analysis. It is also shown to be latent – yet fundamental – in many philosophical accounts. Moreover, it has significant consequences for methodological issues: the warranty of the causal interpretation of causal models, the levels of causation, the characterisation of mechanisms, and the interpretation of probability. This book offers a novel philosophical and methodological approach to causal reasoning in causal modelling and provides the reader with the tools to be up to date about various issues causality rises in social science. (shrink)

Causation has always been a philosophically controversial subject matter. While David Hume’s empiricist account of causation has been the dominant influence in analytic philosophy and science during modern times, a minority view has instead connected causation essentially to agency and manipulation. A related approach has for the first time gained widespread popularity in recent years, due to new powerful theories of causal inference in science that are based in a technical notion of intervention, and James Woodward’s closely connected interventionist theory (...) of causation in philosophy. This monograph assesses five manipulationist or interventionist theories of causation, viewed as theories that purport to tell us what causation is by providing us with the meaning of causal claims. It is shown that they cannot do this, as the conditions on causation that they impose are too weak, mainly due to ineliminable circularities in their definitions of causal terms. It is then argued that a subset of Woodward’s theory can nevertheless contribute crucially to an explanation of the unique role that manipulation has in our acquisition of causal knowledge. This explanation differs from the common regularist explanation of the epistemic utility of manipulation and experiment, and it is taken to confirm several important manipulationist intuitions. However, the success of the explanation depends on (this subset of) interventionism not itself being understood as a theory of causation, but as a theory of intervention. (shrink)

The fact that causal laws in the social sciences are most realistically expressed as both multivariate and stochastic has a number of very important implications for indirect measurement and generalizability. It becomes difficult to link theoretical definitions of general constructs in a one-to-one relationship to research operations, with the result that there is conceptual slippage in both experimental and nonexperimental research. It is argued that problems of this nature can be approached by developing specific multivariate causal models that incorporate sources (...) of measurement bias, along with the theoretical variables of interest. Many general concepts are defined in such a way that causal assumptions are built into the definitions themselves. Additionally, in any given piece of research it is necessary to omit many variables from consideration, and this is often done without careful consideration of the assumptions required to justify such omissions. Finally, generalization to more inclusive populations or a diversity of settings ordinarily requires one to replace constants by variables. It is concluded that the criteria of parsimony, generalizability, and precision are incompatible, given the multivariate nature of social causation, and the author expresses his own preference for sacrificing parsimony in favor of the objectives of achieving increased precision and generalizability of social science laws. (shrink)

La relectura del experimento de Milgram y de otras lecturas a las que me ha llevado la invitación a esta colaboración me ha provocado una serie de reflexiones que van más allá de un comentario exclusivamente centrado y escueto en dicho experimento social para extenderme sobre otros aspectos tanto metodológicos como sustantivos. Algunos de estas reflexiones se sitúan en el epicentro de controversias típicas de las ciencias sociales y de la sociología en particular.

In this essay several aspects of the concept ‘indicator’ are criticized, as well as the technical consequences of those misconceptions. The most relevant among them is the pretence of measuring the validity of an indicator, e. g. the degree of semantic correspondence between the indicator and the concept which it is supposes to refer to.