It is sometimes thought that randomized study group allocation is uniquely proficient at producing comparison groups that are evenly balanced for all confounding causes. Philosophers have argued that in real randomized controlled trials this balance assumption typically fails. But is the balance assumption an important ideal? I run a thought experiment, the CONFOUND study, to answer this question. I then suggest a new account of causal inference in ideal and real comparative group studies that helps clarify the roles of confounding (...) variables and randomization. 1Confounders and Causes2The Balance Assumption3The CONFOUND Study 3.1CONFOUND 13.2CONFOUND 24Disjunction C and the Ideal Study 4.1The ultimate ‘other cause’: C4.2The ideal comparative group study4.3Required conditions for causal inference5Confounders as Causes, Confounders as Correlates6Summary. (shrink)

The experimental approach in economics is a driving force behind some of the most exciting developments in the field. The 'experimental revolution' was based on a series of bold philosophical premises which have remained until now mostly unexplored. This book provides the first comprehensive analysis and critical discussion of the methodology of experimental economics, written by a philosopher of science with expertise in the field. It outlines the fundamental principles of experimental inference in order to investigate their power, scope and (...) limitations. The author demonstrates that experimental economists have a lot to gain by discussing openly the philosophical principles that guide their work, and that philosophers of science have a lot to learn from their ingenious techniques devised by experimenters in order to tackle difficult scientific problems. (shrink)

This article shows how the MISS account of models—as isolations and surrogate systems—accommodates and elaborates Sugden’s account of models as credible worlds and Hausman’s account of models as explorations. Theoretical models typically isolate by means of idealization, and they are representatives of some target system, which prompts issues of resemblance between the two to arise. Models as representations are constrained both ontologically (by their targets) and pragmatically (by the purposes and audiences of the modeller), and these relations are coordinated by (...) a model commentary. Surrogate models are often about single mechanisms. They are distinguishable from substitute models, which are examined without any concern about their connections with the target. Models as credible worlds are surrogate models that are believed to provide access to their targets on account of their credibility (of which a few senses are distinguished). (shrink)

Bas C. van Fraassen presents an original exploration of how we represent the world.

Experiments are commonly thought to have epistemic privilege over simulations. Two ideas underpin this belief: first, experiments generate greater inferential power than simulations, and second, simulations cannot surprise us the way experiments can. In this article I argue that neither of these claims is true of experiments versus simulations in general. We should give up the common practice of resting in-principle judgments about the epistemic value of cases of scientific inquiry on whether we classify those cases as experiments or simulations, (...) per se. To the extent that either methodology puts researchers in a privileged epistemic position, this is context sensitive. (shrink)

one takes to be the most salient, any pair could be judged more similar to each other than to the third. Goodman uses this second problem to showthat there can be no context-free similarity metric, either in the trivial case or in a scientifically ...

Inferences like these are known as extrapolations.

In order to make scientific results relevant to practical decision making, it is often necessary to transfer a result obtained in one set of circumstances—an animal model, a computer simulation, an economic experiment—to another that may differ in relevant respects—for example, to humans, the global climate, or an auction. Such inferences, which we can call extrapolations, are a type of argument by analogy. This essay sketches a new approach to analogical inference that utilizes chain graphs, which resemble directed acyclic graphs (...) (DAGs) except in allowing that nodes may be connected by lines as well as arrows. This chain graph approach generalizes the account of extrapolation I provided in my (2008) book and leads to new insights that integrate the contributions of the other participants of this symposium. More specifically, this approach explicates the role of “fingerprints,” or distinctive markers, as a strategy for avoiding an underdetermination problem having to do with spurious analogies. Moreover, it shows how the extrapolator’s circle, one of the central challenges for extrapolation highlighted in my book, is closely tied to distinctive markers and the Markov condition as it applies to chain graphs. Finally, the approach suggests additional ways in which investigations of a model can provide information about a target that are illustrated by examples concerning nanomaterials in sunscreens and Wendy Parker’s discussion of fingerprints in climate science. (shrink)

Simulations (both digital and analog) and experiments share many features. But what essential features distinguish them? I discuss two proposals in the literature. On one proposal, experiments investigate nature directly, while simulations merely investigate models. On another proposal, simulations differ from experiments in that simulationists manipulate objects that bear only a formal (rather than material) similarity to the targets of their investigations. Both of these proposals are rejected. I argue that simulations fundamentally differ from experiments with regard to the background (...) knowledge that is invoked to argue for the “external validity” of the investigation. (shrink)

I argue for an intentional conception of representation in science that requires bringing scientific agents and their intentions into the picture. So the formula is: Agents (1) intend; (2) to use model, M; (3) to represent a part of the world, W; (4) for some purpose, P. This conception legitimates using similarity as the basic relationship between models and the world. Moreover, since just about anything can be used to represent anything else, there can be no unified ontology of models. (...) This whole approach is further supported by a brief exposition of some recent work in cognitive, or usage-based, linguistics. Finally, with all the above as background, I criticize the recently much discussed idea that claims involving scientific models are really fictions. (shrink)

ABSTRACTExperimentation is traditionally considered a privileged means of confirmation. However, why and how experiments form a better confirmatory source relative to other strategies is unclear, and recent discussions have identified experiments with various modeling strategies on the one hand, and with ‘natural’ experiments on the other hand. We argue that experiments aiming to test theories are best understood as controlled investigations of specimens. ‘Control’ involves repeated, fine-grained causal manipulation of focal properties. This capacity generates rich knowledge of the object investigated. (...) ‘Specimenhood’ involves possessing relevant properties given the investigative target and the hypothesis in question. Specimens are thus representative members of a class of systems, to which a hypothesis refers. It is in virtue of both control and specimenhood that experiments provide powerful confirmatory evidence. This explains the distinctive power of experiments: although modelers exert extensive control, they do not exert this control over specimens; although natural experiments utilize specimens, control is diminished. (shrink)

I consider recent strategies proposed by econometricians for extrapolating causal effects from experimental to target populations. I argue that these strategies fall prey to the extrapolator’s circle: they require so much knowledge about the target population that the causal effects to be extrapolated can be identified from information about the target alone. I then consider comparative process tracing as a potential remedy. Although specifically designed to evade the extrapolator’s circle, I argue that CPT is unlikely to facilitate extrapolation in typical (...) econometrics and evidence-based policy applications. To argue this, I offer a distinction between two kinds of extrapolation, attributive and predictive, the latter being prevalent in econometrics and evidence-based policy. I argue that CPT is not helpful for predictive extrapolation when using the kinds of evidence that econometricians and evidence-based policy researchers prefer. I suggest that econometricians may need to consider qualitative evidence to overcome this problem. (shrink)

In the human sciences, experimental research is used to establish causal relationships. However, the extrapolation of these results to the target population can be problematic. To facilitate extrapolation, we propose to use the statistical technique Latent Class Regression Analysis in combination with the analogical reasoning theory for extrapolation. This statistical technique can identify latent classes that differ in the effect of X on Y. In order to extrapolate by means of analogical reasoning, one can characterize the latent classes by a (...) combination of features, and then compare these features to features of the target. (shrink)

A number of recent discussions comparing computer simulation and traditional experimentation have focused on the significance of “materiality.” I challenge several claims emerging from this work and suggest that computer simulation studies are material experiments in a straightforward sense. After discussing some of the implications of this material status for the epistemology of computer simulation, I consider the extent to which materiality (in a particular sense) is important when it comes to making justified inferences about target systems on the basis (...) of experimental results. (shrink)

There is a takeover movement fast gaining influence in development economics, a movement that demands that predictions about development outcomes be based on randomized controlled trials. The problem it takes up—of using evidence of efficacy from good studies to predict whether a policy will be effective if we implement it—is a general one, and affects us all. My discussion is the result of a long struggle to develop the right concepts to deal with the problem of warranting effectiveness predictions. Whether (...) I have it right or not, these are questions of vast social importance that philosophers of science can, and should, help answer. (shrink)

Why do we represent the world around us using causal generalizations, rather than, say, purely statistical generalizations? Do causal representations contain useful additional information, or are they merely more efficient for inferential purposes? This paper considers the second kind of answer: it investigates some ways in which causal cognition might aid us not because of its expressive power, but because of its organizational power. Three styles of explanation are considered. The first, building on the work of Reichenbach in "The Direction (...) of Time", points to causal representation as especially efficient for predictive purposes in a world containing certain pervasive patterns of conditional independence. The second, inspired by work of Woodward and others, finds causal representation to be an excellent vehicle for representing all-important relations of manipulability. The third, based in part on my own work, locates the importance of causal cognition in the special role it reserves for information about underlying mechanisms. All three varieties of explanation show promise, but particular emphasis is placed on the third. (shrink)

Comparative process tracing is the best analysis of extrapolation inferences in the philosophical and scientific literature so far. In this essay I examine some similarities and differences between comparative process tracing and former attempts to capture the logic of extrapolation, such as the analogical approach. I show that these accounts are not different in spirit, although comparative process tracing supersedes previous proposals in terms of analytical detail. I also examine some qualms about the possibility of drawing extrapolation inferences in the (...) social sciences and conclude by suggesting that there may be cases of extrapolation without process tracing. (shrink)

I argue against theories that attempt to reduce scientific representation to similarity or isomorphism. These reductive theories aim to radically naturalize the notion of representation, since they treat scientist's purposes and intentions as non-essential to representation. I distinguish between the means and the constituents of representation, and I argue that similarity and isomorphism are common but not universal means of representation. I then present four other arguments to show that similarity and isomorphism are not the constituents of scientific representation. I (...) finish by looking at the prospects for weakened versions of these theories, and I argue that only those that abandon the aim to radically naturalize scientific representation are likely to be successful. (shrink)

In this article we defend the inferential view of scientific models and idealisation. Models are seen as “inferential prostheses” (instruments for surrogative reasoning) construed by means of an idealisation-concretisation process, which we essentially understand as a kind of counterfactual deformation procedure (also analysed in inferential terms). The value of scientific representation is understood in terms not only of the success of the inferential outcomes arrived at with its help, but also of the heuristic power of representation and their capacity to (...) correct and improve our models. This provides us with an argument against Sugden’s account of credible models: the likelihood or realisticness (their “credibility”) is not always a good measure of their acceptability. As opposed to “credibility” we propose the notion of “enlightening”, which is the capacity of giving us understanding in the sense of an inferential ability. (shrink)

The assessment of models in an experiment depends on their material nature and their function in the experiment. Models that are used to make the phenomenon under investigation visible - sensors - are assessed by calibration. However, calibration strategies assume material intervention. The experiment discssed in this paper is an experiment in economics to measure the influence of technology shocks on business cycles. It uses immaterial, mathematical instruments. It appears that calibration did not work for these kinds of models, it (...) did not provide reliable evidence for the facts of the business cycle. (shrink)

Francesco Guala once wrote that ‘The problem of extrapolation is a minor scandal in the philosophy of science’. This paper agrees with the statement, but for reasons different from Guala’s. The scandal is not, or not any longer, that the problem has been ignored in the philosophy of science. The scandal is that framing the problem as one of external validity encourages poor evidential reasoning. The aim of this paper is to propose an alternative—an alternative which constitutes much better evidential (...) reasoning about target systems of interest, and which makes do without consideration of external validity. (shrink)

Philosophers of science have had little to say about ‘middle-range theory’ although much of what is done in science and of what drives its successes falls under that label. These lectures aim to spark an interest in the topic and to lay groundwork for further research on it. ‘Middle’ in ‘middle range’ is with respect to the level both of abstraction and generality. Much middle-range theory is about things that come under the label ‘mechanism’. The lectures explore three different kinds (...) of mechanism: struc tural mechanisms or underlying systems that afford causal pathways; causal-chain mechanisms that are represented in what in policy contexts are called ‘theories of change’ and for which I give an extended account following the causal process theory of Wesley Salmon; and middle-range-law mechanisms like those discussed by Jon Elster, which I claim are —and rightly are— rampant throughout the social sciences. The theory of the democratic peace, that democracies do not go to war with democracies, serves as a running example. The discussions build up to the start of, first, an argument that reliability in social (and natural) science depends not so much on evidence as it does on the support of a virtuous tangle of practices (without which there couldn’t even be evidence), and second, a defence of a community-practice centred instrumentalist understanding of many of the central basic principles that we use (often successfully) in social (and in natural) science for explanation, prediction and evaluation. (shrink)

In Simulation and Similarity, Michael Weisberg offers a similarity-based account of the model–world relation, which is the relation in virtue of which successful models are successful. Weisberg’s main idea is that models are similar to targets in virtue of sharing features. An important concern about Weisberg’s account is that it remains silent on what it means for models and targets to share features, and consequently on how feature-sharing contributes to models’ epistemic success. I consider three potential ways of concretizing the (...) concept of shared features: as identical, quantitatively sufficiently close, and sufficiently similar features. I argue that each of these concretizations faces significant challenges, leaving unclear how Weisberg’s account substantially contributes to elucidating the relation in virtue of which successful models are successful. Against this background, I outline a pluralistic revision and argue that this revision may not only help Weisberg's account evade several of the problems that I raise, but also offers a novel perspective on the model–world relation more generally. 1Introduction 2Weisberg’s Feature-Sharing Account 3What Is a Shared Feature? 3.1Identity 3.2Sufficient closeness 3.3Sufficient similarity 4Turning Weisberg’s Account ‘Upside Down’ 5Conclusion. (shrink)