Causal Inference

Individuals do not respond uniformly to treatments, events, or interventions. Sociologists routinely partition samples into subgroups to explore how the effects of treatments vary by covariates like race, gender, and socioeconomic status. In so doing, analysts determine the key subpopulations based on theoretical priors. Data-driven discoveries are also routine, yet the analyses by which sociologists typically go about them are problematic and seldom move us beyond our expectations, and biases, to explore new meaningful subgroups. Emerging machine learning methods allow researchers to explore sources of variation that they may not have previously considered, or envisaged. In this paper, we use causal trees to recursively partition the sample and uncover sources of treatment effect heterogeneity. We use honest estimation, splitting the sample into a training sample to grow the tree and an estimation sample to estimate leaf-specific effects. Assessing a central topic in the social ine...

Adaptive control in the face of uncertainty involves making online predictions about events in order to plan actions to reliably achieve desirable outcomes. However, we often face situations in which the effects of our actions are probabilistic (e.g., returns on our investments), and the environment itself is dynamic, - it can change irrespective of any action we decide to take, and in a relatively stable or unstable way (e.g., economic climate). Therefore, the problem of control concerns learning to isolate the effects that are generated directly by our actions, from those that occur independently of them. We face this problem in a host of uncertain dynamic environments: ecological (using fertilizers to increase crop yield), economic (investing in real estate), industrial (operating nuclear power plants), mechanical (driving cars) and medical (stemming outbreaks of disease). Thus far there has been little attempt at a synthesis of the amassing research directed towards understanding control in the face of uncertainty. The aim of this book is to bring together theoretical and empirical research from disparate disciplines spanning Social, Cognitive and Clinical Psychology, Human Factors and Neuroscience to Engineering and Machine learning. It serves as a round up of the different techniques used to examine issues concerning control (prediction, agency, causality, uncertainty) and unifies a substantial amount of disparate research. The general framework (Monitoring and control- MC framework) used to achieve this proposes two central ideas (agency, uncertainty) that have important ramifications across disciplines. In sum, control processes involve self-regulatory mechanisms that enable us to make subjective estimates of success, which helps to anchor our interpretation of uncertain events. In addition, these estimates are integrated with estimations of rates of change of the events in the environment which helps us track the outcomes we aim to control.

An experiment was conducted to investigate the relative contributions of syntactic form and content to conditional reasoning. The content domain chosen was that of causation. Conditional statements that described causal relationships (if (cause>, then (effect>) were embedded in simple arguments whose entailments are governed by the rules-oftruth-functional logic (i.e., modus ponens, modus tollens, denying the antecedent, and affirming the consequent). The causal statements differed in terms ofthe number of alternative causes and disabling conditions that characterized the causal relationship. (A disabling condition is an event that prevents an effect from occurring even though a relevant cause is present.) Subjects were required to judge whether or not each argument's conclusion could be accepted. Judgments were found to vary systematically with the number of alternative causes and disabling conditions. Conclusions of arguments based on conditionals with few alternative causes or disabling conditionswerefoun~d:tobe-rnore accept~ able than cdnclusions based on those with many.

In this paper we examine the ability to perform causal reasoning with recursive equilibrium models. We identify a critical postulate, which we term the Manipulation Postulate, that is required in order to perform causal inference, and we prove that there exists a general class F of recursive equilibrium models that violate the Manipulation Postulate. We relate this class to the existing phenomenon of reversibility and show that all models in F display reversible behavior, thereby providing an explanation for reversibility and suggesting that it is a special case of a more general and perhaps widespread problem. We also show that all models in F possess a set of variables V’ whose manipulation will cause an instability such that no equilibrium model will exist for the system. We define the Structural Stability Principle which provides a graphical criterion for stability in causal models. Our theorems suggest that drastically incorrect inferences may be obtained when applying the Manipulation Postulate to equilibrium models, a result which has implications for current work on causal modeling, especially causal discovery from data.

本報告は，因果関係に関する「人間らしい」推論の1つのモデルを提案する。
従来より，p という事象とq という事象とのあいだにどのような関係があればそれが「因果関係」と認知されるのか，について議論されてきた。われわれは，その認知の中で働いていると見られる「対称性バイアス」および「相互対称性バイアス」という2 つの非論理的な認知的バイアスが，結果として合理的な因果関係の認知・推論・期待形成をもたらしているのではないかという仮説を立てた。
これを検証するため，この2 つのバイアスが柔軟に働く「緩い対称性モデル」を定義し，コンピューター内に構築したエージェントに実装して，このエージェントが「2 本腕バンディット問題」に対して示すパフォーマンスを調べた。その結果，このエージェントの振る舞いは人間による被験者実験のデータときわめてよく合致するだけでなく，機械的な硬直性を超えた柔軟かつ人間らしい挙動を示すことがわかった。
このように，「緩い対称性モデル」は人間の因果推論を非常によく表現するモデルであると言える。と同時に，「対称性バイアス」と「相互排他性バイアス」は人間の推論の基礎にある本質的な契機だと考えられる。

This volume explores abductive cognition, an important but, at least until the third quarter of the last century, neglected topic in cognition. The book aims at increasing knowledge about creative and expert inferences. The study of these high-levelmethods of abductive reasoning is situated at the crossroads of philosophy, logic, epistemology, artificial intelligence, neuroscience, cognitive psychology, animal cognition and evolutionary theories; that is, at the heart of cognitive science. Philosophers of science in the twentieth century have traditionally distinguished between the inferential processes active in the logic of discovery and the ones active in the logic of justification. Most have concluded that no logic of creative processes exists and, moreover, that a rational model of discovery is impossible. In short, scientific creative inferences are irrational and there is no “reasoning” to hypotheses. On the other hand, some research in the area of artificial intelligence has shown that methods for discovery could be found that are computationally adequate for rediscovering – or discovering for the first time – empirical or theoretical laws and theorems.

Content Level » Research

Keywords » Abductive Cognition - Cognitve Systems - Human Cognition - Hypothetical Reasoning

Dieser Beitrag befasst sich mit folgenden Fragen: (1) Wie verwenden Sozialwissenschaftler Begriffe wie „Ursache“ und „Kausalität“? (2) Wie gehen Sozialwissenschaftler vor, wenn Kausalität ermittelt werden soll? (3) Inwieweit sind die verwendeten Begriffe und Verfahrensweisen sinnvoll? Nach der Explikation des Kausalitätsbegriffs wird gezeigt, warum  Experimente in besonderem Maße geeignet sind, Kausalhypothesen zu testen. Sodann steht die Prüfung von
Kausalaussagen durch multivariate statistische Verfahren zur Diskussion. Wir zeigen, dass Kausalmodelle aus Ursachenbehauptungen im vorher explizierten Sinne bestehen. Weiter wird ausführlich die Regressionsanalyse als Beispiel multivariater Verfahren zur Prüfung von Kausalaussagen diskutiert. U. a. wird gezeigt, dass diese als eine Explikation des vorher explizierten Kausalitätsbegriffs angesehen werden kann und in der Lage ist, Kausalhypothesen zu testen. Es können aber keine induktiven Schlüsse gezogen werden. Es wird weiter argumentiert, dass es keine Algorithmen gibt, die aus einem gegebenen Datensatz gültige Kausalmodelle ableiten. Weitere Fragen, die angesprochen werden, sind die erforderliche Anzahl von Fällen für die Prüfung von Kausalhypothesen und Kausalhypothesen in Messmodellen.

Over the last several decades, the theories of emergence and downward causation have become arguably the most popular conceptual tools in scientific and philosophical attempts to explain the nature and character of global organization observed in various biological phenomena, from individual cell organization to ecological systems. The theory of emergence acknowledges the reality of layered strata or levels of systems, which are consequences of the appearance of an interacting range of novel qualities.

A closer analysis of emergentism, however, reveals a number of philosophical problems facing this theory. In Emergence, Mariusz Tabaczek offers a thorough analysis of these problems and a constructive proposal of a new metaphysical foundation for both the classic downward causation–based and the new dynamical depth account of emergence theory developed by Terrence Deacon. Tabaczek suggests ways in which both theoretical models of emergentism can be grounded in the classical and the new (dispositionalist) versions of Aristotelianism. This book will have an eager audience in metaphysicians working both in the analytic and the Thomistic traditions, as well as philosophers of science and biology interested in emergence theory and causation.

Terrorism studies has an enduring puzzle: how do terrorist organisations grow, persist and survive? Using the Japanese new religion Aum Shinrikyo as a case study, this thesis posits that three causal mechanisms—cultural transmission, social learning, and folklore—affect terrorist organisations’ survivability, success, or failure. Original contributions to knowledge include a new analytical theory of strategic subcultures in terrorist organisations, new tests using the qualitative methodology of process tracing, and a new causal analysis of Aum Shinrikyo’s initiatory, religious sub-system (which facilitated founder Shoko Asahara’s paranoid delusions, leadership elite deviance, and follower indoctrinability). A new research agenda is also identified.

We assemble a rich dataset that maps geocoded microdata on various measures of living standards, point-specific geographical attributes, and remotely-sensed pollution data onto gridded population data to determine whether urbanization yields greater benefits or costs. We focus on developing Asia where the world’s largest proportion of urban population is projected to come from. We find evidence that population density, our chosen measure of urban scale, provides greater benefits relative to costs, indicating that urbanization brings net positive gains to the developing world. In particular, our causal estimates show that higher population density leads to greater household wealth, improved hygiene and sanitation, and better access to utilities, health services, clean energy, and private goods and services. On the other hand, our results suggest that density either reduces or does not significantly affect urban costs such as indoor pollution, space congestion, diseases, and crime. The only exception is outdoor pollution which deteriorates with population density.

A central aim of medical research is causal inference. Does this drug have harmful side-effects? Is this medical intervention effective? Does this chemical cause cancer? To provide evidence that bears on these important questions, many sorts of measurements are made in a variety of types studies. These measurements generate a plethora of data, and this data must be quantitatively summarized so that the data is rendered relevant to causal hypotheses. That is, to render measurements made in medical research into evidence for a causal hypothesis, those measurements must be transformed into summary quantifications, called " outcome measures. " This paper has two aims. First, we argue for the superiority of one particular form of outcome measure, called absolute measures. Second, we argue against a widely held myth in epidemiology. The myth is that in observational methods like case-control studies only a particular relative outcome measure called the odds ratio can be calculated, and we argue that there is no justification for this myth.

Statistical mediation and moderation have most prominently been distinguished by Baron and Kenny (1986). More complex models that combine both of these effects have recently received increased attention, namely mediated moderation and moderated mediation (e.g., Muller, Judd, & Yzerbyt, 2005). Presently the focus is on a three variable model that is often claimed to represent an instance of moderated mediation or mediated moderation. More specifically, in this model a single variable is considered to simultaneously mediate and moderate the same effect.
We show that this specific model however cannot be considered either an instance of mediated moderation nor of moderated mediation. Also, we argue that this particular model is a priori misspecified. A data pattern that seems to agree with this model is recognized as plausible, but it indicates that the model must be modified in one of two ways to be methodologically sound. We conclude by recommending to not use this three variable model and to consider evidence that seemingly agrees with it as evidence that the three variable model is inadequate.

In this paper hope to break some new ground and help make Spinoza's key doctrine of Scientia Intuitiva more readily understandable. I will proceed in the following order (keep in mind the word ‘proceed’). I will first provide a close preliminary analysis of the content and development of Spinoza’s discussion of scientia intuitiva in the Treatise on the Emendation of the Intellect and the Ethics. In the second part, I will address two closely related questions which have drawn the attention of many readers: (i) Is scientia intuitiva a cognition that is inferential? and (ii) Is scientia intuitiva a kind of cognition that is closely tied to causation? To the surprise of many, I will argue that Spinoza’s intuitive knowledge is both inferential and causal (in a certain rigid sense of ‘inferential’ and ‘causal’ to be explained below). In the third and final part, I will explain Spinoza’s preference for scientia intuitiva and his frequent use of mathematical examples, and then attempt to map the major remaining questions and problems surrounding the doctrine.

Hypotheses about psychological processes are often tested using traditional mediation analysis. This analysis relies onmeasurement of a transmitting variable. Conducting this analysis has become almost synonymous with examining processhypotheses. An alternative strategy to mediation analysis (the Testing-a-Process-hypothesis-by-an-Interaction Strategy, TPIS) isillustrated here. TPIS is based on a fully experimental design whereby a hypothesized process is tested by an interaction betweenthe hypothesized cause of an effect and a contextual variable. In the interaction term, the contextual variable allows comparisonof the causal effect observed when the process is uninterrupted to the effect observed when the process is interrupted. Thus, TPIS translatesatheoreticalprocesshypothesis intoastatisticalinteractionhypothesisthatusesafully experimental designtodirectlyexamine the hypothesized process.

... Free Will: a historical and philosophical introduction jej autor, Ilham Dilman, przenikają karty najstarszych dzieł literatury greckiej [Dilman, 1999]. ... świadomej woli: „plany działania nie wydają się wymuszać zachowa-nia w tym sensie, iż jeśli pojawi się dany plan, dochodzi do ...

PC is a prototypical constraint-based algorithm for learning Bayesian networks, a special case of directed acyclic graphs. An existing variant of it, in the R package pcalg, was developed to make the skeleton phase order independent. In return, it has notably increased execution time. In this paper we clarify that the PC algorithm the skeleton phase of PC is indeed order independent. The modification we propose outperforms pcalg's variant of the PC in terms of returning correct networks of better quality as is less prone to errors and in some cases it is a lot more computationally cheaper. In addition, we show that pcalg's variant does not return valid acyclic graphs.

This article considers whether employment is a "turning point" in the lives of criminal offenders. We present a brief overview of research on the relationship between work and crime. Our review suggests that work programs appear to be more effective for adult offenders than for adolescents and young adults. The quality of employment also appears to be important for reducing

Zusammenfassung Exploratives Experimentieren schafft vielfältige Bedingungen und beobachtet deren Folgen zur systematischen Untersuchung von Ereignismengen hinsichtlich Bedingung-Folge-Verhältnissen. Dabei müssen bestimmende notwendige Bedingungen von begleitenden notwendigen Bedingungen sowie von begleitenden zufälligen Umständen unterschieden werden. Diese werden in konkreten Experimenten vermöge Variation und Kontrolle (Konstanthaltung, Minimierung) von Bedingungen/Umständen empirisch untersucht. Dafür stehen die von J. S. Mill entwickelten Methoden zur Verfügung: Übereinstimmungsmethode, Differenzenmethode, indirekte Differenzenmethode, Residuenmethode, Variationsmethode. Sie führen zwar nicht zu endgültiger Sicherheit bezüglich eindeutiger Bestimmtheit der Komponenten von Bedingung-Folge-Verhältnissen, geben aber eine solide Grundlage ab für sachgemäße Ideenbildungen im Hinblick auf die Entdeckung von Gesetzen der anorganischen Natur. Das einzelne ideale naturwissenschaftliche ...