The conditional probability and the linear flow graph approaches to analyzing probabilistic causal systems

An alternative theoretical approach to the analysis for dichotomous causal systems that involve probabilistic causation, the conditional probability approach, has recently been explicated. It was shown that there exist various composition and decomposition rules for analyzing various kinds of general causal systems, and an important distinction between pure-“or”-and pure-“and”-causal systems was explicated. In this paper these earlier results are used to analyze a causal system which has been studied by J.A. Davis (1976), who uses his linear flow graph approach to analysis. The results of the conditional probability approach are compared to the linear flow graph, and it is shown that the two approaches lead to strikingly different results.     

Graphics for the Minimal Sufficient Cause Model

Graphical representations of causation have been usedfor at least seventy years, and the modern developmentof directed acyclic graphs to portray causal systemscontinues the trend. It is sometimes difficult tounderstand, however, what it is about these diagramsthat is `causal'. The approach to causal graphicsthat is taken here is to base the development ofgraphics on the concepts of an underlying causaltheory, the minimal sufficient cause model. Thisleads to defining a Boolean `and' for arrows thatrepresent causal pathways, and a Boolean `or' forcollections of pathways. Complementation is a morecomplex operation in the minimal sufficient causetheory than simply inverting the sense of a causalrelationship, and this also is represented in thegraphics. By using diagrams that are more faithful tounderlying causal systems, and using a coherent causaltheory, both the perception of causation and itsanalysis might be enhanced.     

Ethnographic studies of social causation: observing conjunctures of action sequences in particular settings

While the importance of ethnographic research for causal understanding is widely recognized, the logic of this argument has not been well articulated. The purpose of this paper is to specify the conditions under which direct field observations become essential for causal studies. It is argued in this paper that the concept of “human causation” is fundamental for understanding the causal relevance of ethnographic research. Specifically, ethnographic methods are most suited for observing conjunctures of action sequences in particular social settings. In causally fixed settings, ethnographers focus on the recurrent causal pathways leading to a predictable outcome; and in causally open settings, ethnographers study the critical junctures in and important conjunctures of contingent sequences of events that give rise to unpredictable outcomes. In both instances, the immersion of the researcher into the setting to observe the unfolding of the events being studied plays a key role in uncovering and understanding the underlying causal processes. The limitations of ethnographic methods for causal studies are also discussed in the paper.     

Mulaik on atomism, contraposition and causation

Causal inference using statistical models plays a central role in many areas of behavioral science, but the underlying metatheory of causal explanation remains poorly developed. Mulaik’s work on causation offers a useful foray into this topic. Evaluation of two negative arguments applied to a broad range of theories of causation offer overdue critical assessment of this contribution. More broadly, the critical evaluation of Mulaik’s arguments speak to the need for better integration of substantive theories and statistical models in causal research.     

The Dynamic Production Axioms and System Dynamics Behaviors: The Foundation for Future Integration

Realistically, organizational and/or system performance is dynamic and non-linear. However, in the efficiency literature, system performance is frequently evaluated considering linear combinations of the input/output variables and without explicitly taking into account the causes of efficiency behavior nor the dynamic behavior of systems. Policy decisions based on these results may be sub-optimized because the non-linear relationships among variables, causal relationships, and feedback mechanisms are ignored.This research takes the initial step of evaluating system performance in a dynamic environment, by relating the factors that effect system performance to the policies that govern it. To accomplish this, this paper extends the concepts of the static production axioms into a dynamic realm, where inputs are not instantaneously converted into outputs. The relationships of these new dynamic production axioms to the basic behaviors associated with system dynamics structures are explored.     

Black-White Economic Participation in Large U.S. Cities: Some Determinants of Their Differentials

Abstract . In this study some of the causes of the differentials in Black-White economic participation in large U.S. cities are examined. Specific concern is with unemployment and withdrawal from the labor force. The underlying hypothesis is that these differentials are a response to compositional differences between the two populations. The effects of region, economic structure, and segregation are also examined. A causal model is explicated in which these variables are inter related and their direct and indirect effects upon the dependent variables analyzed. The data suggest that, while having similar patterns of causes, labor force withdrawal and unemployment differentials are distinct phenomena. It also appears that a major portion of the variance in these variables is not accounted for by compositional differences between Blacks and Whites on socio-economic variables. However, educational and occupational differentiation have theoretically significant effects. It is argued that a major portion of the residual variance is due to discrimination, although there is no way of directly testing this hypothesis. An important negative finding is the apparently minor impact of residential segregation, suggesting that the physical isolation of Blacks is not a key factor in their limited economic participation. Finally, the data suggest that it is meaningful to regard assimilation as a multidimensional phenomenon whose dimensions are causally interrelated.     

Disproving Causal Relationships Using Observational Data*

Economic theory is replete with causal hypotheses that are scarcely tested because economists are generally constrained to work with observational data. We describe a method for testing a hypothesis that one observed random variable causes another. Contingent on a sufficiently strong correspondence between the two variables, an appropriately related third variable can be employed for the test. The logic of the procedure naturally suggests strong and weak grounds for rejecting the causal hypothesis. Monte Carlo results suggest that weakly grounded rejections are unreliable for small samples, but reasonably reliable for large samples. Strongly grounded rejections are highly reliable, even for small samples.     

Direct and indirect effects of continuous treatments based on generalized propensity score weighting

This paper proposes semi- and nonparametric methods for disentangling the total causal effect of a continuous treatment on an outcome variable into its natural direct effect and the indirect effect that operates through one or several intermediate variables called mediators jointly. Our approach is based on weighting observations by the inverse of two versions of the generalized propensity score (GPS), namely the conditional density of treatment either given observed covariates or given covariates and the mediator. Our effect estimators are shown to be asymptotically normal when the GPS is estimated by either a parametric or a nonparametric kernel-based method. We also provide a simulation study and an empirical illustration based on the Job Corps experimental study.     

Using Single-Equation Models of Function-of-Functions Type in Sociological Research

This paper argues that single-equation explanatorymodels of many types of social phenomena should not be built in accordance with establishedsociological ways of thinking. In sociological research, the focus is often on the causal mechanisms behindphenomena, and it is often interesting to use models that show the hierarchical causal structure,that is, how influences are nested in the causal process. I propose such a model with a form that reflectsa two-step structure. According to this model, the dependent factor is a product of independentfactors that are linear functions of variables. The model, which should be used with the factor product inunexpanded form, can be assumed to have wide application. However, the models used insociological research and discussed in textbooks are generally very different. They do not have afunction-of-functions form, but take a form in which variables are directly entered. Furthermore,even if they take interaction into consideration, they are linear in an extended sense because they construeit as one or more terms that are products of single variables. In comparison with the proposed typeof model, these models are technically simpler. However, this paper argues that the proposed typeof model is superior in many contexts because it better reflects the causal process.     

The political economy of environmental and social security policies: the role of environmental lobbying

Social security provides retirement benefits to the old at the expense of the working young, while environmental investment benefits the future of the young at the expense of the old. This paper presents a model incorporating this intergenerational conflict on public spending and considers the political determination of environmental investment and social security by focusing on the Markovian political equilibria. It is shown that (1) the political equilibria are generally inefficient, and (2) the introduction of environmental lobbying into politics may improve environmental quality but degrade lifetime utility in the long run.      

Modelling short- and long-term effects in the aggregate demand for soft drinks

A dynamic regression model for non-durable commodity demand is specified based on the additive unobserved components seasonal model with causal variables. The seasonal component which includes retailer stock effects is modelled as a seasonal ARMA process with fixed temperature effects. The non-seasonal component incorporates both short-run consumer responses and long-run adaptation to steady-state growth paths. Aggregation effects on per capita demand caused by the increase of new consumers as real income grows are investigated and the resulting growth trends accounted for. It is shown that neglecting these trends in the long-run adaptation process is likely to produce biased predictions and misleading estimates of crucial response parameters.     

Diagnosing Response Style Behavior by Means of a Latent-Class Factor Approach. Socio-Demographic Correlates of Gender Role Attitudes and Perceptions of Ethnic Discrimination Reexamined

It is generally accepted that response style behavior in survey research may seriously distort the measurement of attitudes and subsequent causal models that include attitudinal dimensions. However, there in no single accepted methodological approach in dealing with this issue. This article aims at illustrating the flexibility of a latent class factor approach in diagnosing response style behavior and in adjusting findings from causal models with latent variables. We present a substantive example from the Belgian MHSM research project on integration-related attitudes among ethnic minorities. We argue that an extreme response style can be detected in analyzing two independent sets of Likert-type questions referring to `gender roles' and `feelings of ethnic discrimination'. If the response style is taken into account the effect of covariates on attitudinal dimensions is more adequately estimated.     

Necessary and sufficient factors in employee downsizing? A qualitative comparative analysis of lay‐offs in France and the UK, 2008–2013

Embedded in the literature on financialization and institutional approaches, this study is an examination of the causal factors of employee downsizing in two institutionally dissimilar settings, France and the UK, using the fuzzy sets variant of Qualitative Comparative Analysis. The findings show that the roughly equivalent use of large‐scale lay‐offs in the two countries is coupled with strikingly different causal factors. Our argument suggests the importance of complex causation whereby employee downsizing reflects the growing influence of financial considerations in the governance of companies, but its diffusion across countries is shaped by different configurations of institutional arrangements.     

On the Relationship Between Validity and Power

This article examines the relationship between predictive validity and statistical power. It is shown that higher validity may be associated with scales exhibiting lower power, whereas higher power may be warranted for scales possessing lower validity. High validity is in general neither a necessary nor sufficient condition for high power with scales contemplated for use as dependent variables in social science research. Conversely, high power is neither necessary nor sufficient for high validity. A numerical example illustrates this interrelation.     

Factual and modal notions in social research

The article discusses a distinction between factual and modal notions, and corresponding generalizations, in social research. The discussion starts from the suggestion, made by Charles Ragin, that theoretical statements in social research most often can be formulated as statements about sets of cases and relations between such sets. In contrast to this view, it is argued that theoretical statements in social research often require modal notions referring to possibilities and probabilities which cannot be formulated as statements about sets of cases. In order to show this, the article reformulates Ragin’s set-theoretic approach in the conceptual framework of statistical variables. It is shown that this can be done for both crisp and fuzzy set versions of Ragin’s approach. The article then goes on to argue that social research is often interested in modal generalizations (probabilistic and deterministic rules) which require a fundamentally different conceptual framework. The article shows how such a framework can be defined, and finally indicates its usage for causal interpretations.     

Bounds on average causal effects in studies with a latent response variable

Consider a case where cause-effect relationships between variables can be described as a directed acylic graph and the corresponding recursive factorization of a joint distribution. In order to provide the bounds on average causal effects in studies with a latent response variable, this paper proposes a graphical criterion for selecting covariates and variables caused by the response variable. The result enables us not only to judge from the graph structure whether the bounds on an average causal effect can be expressed through the observed quantities, but also to provide their closed-form expressions in case where its answer is affirmative. The graphical criterion of this paper is helpful to evaluate the bounds on average causal effects when it is difficult to observe a response variable.     

An Empirical Study of Schedule Delay Causes Based on Taiwan's Litigation Cases

Many uncertainties can cause construction projects to be delayed, resulting in conflicts between the two parties to a construction contract. Using case study and content analysis methodologies, this study analyzed 79 litigation cases in Taiwan to identify the main causes of schedule delays in construction projects, which are “change orders,” “changed scope of the work,” “delayed site handover,” and “weather.” In this study, these causes are organized into a causation model to provide a reference for preventing schedule delay. The employed approach can be implemented for assessments of other regions, as schedule delays are common features in most construction projects.     

Conditional independence and log linear models for multi-dimensional contingency tables

Interference about conditional independence in relation to log linear models are discussed for contingency tables. The parameters and likelihood ratios for a log linear model with a dependent variable are shown to be identical to those for a multivariate model. An approximaate method of calculating log likelihood ratios, even when all dimensions of the table have more than two levels (no binary variables) is derived. The implications for sociological “causal” models are discussed.     

Generalized method of moments specification testing

This paper analyzes the asymptotic power properties of specification tests which are based on a finite set of moment conditions. It shows that any such test may fail against general misspecification that causes estimator inconsistency. The mutual asymptotic equivalence of maximal degree of freedom tests is shown and the form of optimal tests against specific forms of misspecification is derived. Applications to testing for exogeneity of a set of instrumental variables are presented.