Counterfactual Framework Research Articles

Methodological challenges in causal research on racial and ethnic patterns of cognitive trajectories: measurement, selection, and bias.

Research focused on understanding how and why cognitive trajectories differ across racial and ethnic groups can be compromised by several possible methodological challenges. These difficulties are especially relevant in research on racial and ethnic disparities and neuropsychological outcomes because of the particular influence of selection and measurement in these contexts. In this article, we review the counterfactual framework for thinking about causal effects versus statistical associations. We emphasize that causal inferences are key to predicting the likely consequences of possible interventions, for example in clinical settings. We summarize a number of common biases that can obscure causal relationships, including confounding, measurement ceilings/floors, baseline adjustment bias, practice or retest effects, differential measurement error, conditioning on common effects in direct and indirect effects decompositions, and differential survival. For each, we describe how to recognize when such biases may be relevant and some possible analytic or design approaches to remediating these biases.

Neighborhood Poverty and American Indian Infant Death: Are The Effects Identifiable?

Poor living conditions are posited as an underlying cause of American Indian (AI) infant mortality, which is unusually high in the postneonatal period. We explore whether the effects of neighborhood poverty on AI infant death are identifiable by using observational data. Vital records for infants born between 1990 and 1999 to AI women in a metropolitan area (n = 4751) are linked with tract-level poverty data. A counterfactual framework, an explicit causal contrast study design, and propensity score matching methods were used. For each comparison, we created exchangeable groups by matching infants with the same probability of exposure to poverty when one was exposed and the other was not. Our results suggest that neighborhood poverty has little effect on AI infant death outcomes. Importantly, the study design makes transparent the challenge of identifying appropriate analytic comparison groups in studies of neighborhood poverty and health. Collecting additional data will likely not overcome the fact that AIs with a high probability of living in poverty rarely reside in low-poverty neighborhoods. Yet, some of them must if a meaningful counterfactual comparison is to be made and the effects of neighborhood poverty on AI infant death are to be identified.

Dimension reduction summaries for balanced contrasts

We discuss the covariate dimension reduction properties of conditional density ratios in the estimation of balanced contrasts of expectations. Conditional density ratios, as well as related sufficient summaries, can be used to replace the covariates with a smaller number of variables. For example, for comparisons among k populations the covariates can be replaced with k - 1 conditional density ratios. The dimension reduction properties of conditional density ratios are directly connected with sufficiency, the dimension reduction concepts considered in regression theory, and propensity theory. The theory presented here extends the ideas in propensity theory to situations in which propensities do not exist and develops an approach to dimension reduction outside of the potential outcomes or counterfactual framework. Under general conditions, we show that a principal components transformation of the estimated conditional density ratios can be used to investigate whether a sufficient summary of dimension lower than k - 1 exists and to identify such a lower dimensional summary.

Taking a new look at empirical models of adoption: average treatment effect estimation of adoption rates and their determinants

AbstractThis article shows that the observed sample adoption rate does not consistently estimate the population adoption rate even if the sample is random. It is proved that instead the sample adoption rate is a consistent estimate of the population joint exposure and adoption rate, which does not inform about adoption per se. Likewise, it is shown that a model of adoption with observed adoption outcome as a dependent variable and where exposure to the technology is not observed and controlled for cannot yield consistent estimates of the determinants of adoption. The article uses the counterfactual outcomes framework to show that the true population adoption rate corresponds to what is defined in the modern policy evaluation literature as the average treatment effect (ATE), which measures the effect or impact of a “treatment” on a person randomly selected in the population. In the adoption context, a “treatment” corresponds to exposure to the technology. The article uses the ATE estimation framework to derive consistent nonparametric and parametric estimators of population adoption rates and their determinants and applies the results to consistently estimate the population adoption rates and determinants of the NERICA (New Rice for Africa) rice varieties in Côte d'Ivoire. The ATE methodological approach developed in the article has significant policy implications with respect to judging the intrinsic merit of a new technology in terms of its potential demand by the target population independently of issues related to its accessibility and in terms of the decision to invest or not in its wide‐scale dissemination.

From counterfactuals to sufficient component causes and vice versa

The sufficient component cause model and the counterfactual model have now become cornerstones for causal thinking in epidemiology. Dana Flanders' paper [I] provides a nice exposition of certain rela tionships that hold between these two forms of causal models. The sufficient component cause model con siders sets of actions, events or states of nature which together inevitably bring about the outcome under consideration. The model gives an account of the causes of a particular effect. It addresses the question, "Given a particular effect, what are the various events which might have been its cause?" The potential outcomes or counterfactual model focuses on one particular cause or intervention and gives an account of the various effects of that cause. In contrast to the sufficient component cause framework, the potential outcomes framework addresses the question, "What would have occurred if a particular factor were intervened upon and thus set to a different level than it in fact was?" Unlike the sufficient component cause framework, the counterfactual framework does not require a detailed knowledge of the mechanisms by which the factor affects the outcome. This feature of the counterfactual framework makes it well suited to formalize causal effects in randomized experiments in which only data on exposure and outcome are available. Both the sufficient component cause and the counterfactual models have a long history in the philosophical and in the epidemiologic-statistical lit erature. The philosophical ideas behind counterfac tual models can be traced back at least as far as Hume [2], who defined a cause as a relation between two objects "where, if the first object had not been, the second never [would have] existed."' The philos ophy of counterfactuals has been developed exten sively by Lewis [3, 4] and has received considerable

Strengthening Causal Inference in Studies using Non-experimental Data: An Application of Propensity Score and Instrumental Variable Methods

This study attempts to show how studies using non-experimental data can strengthen causal inferences by applying propensity score and instrumental variable methods based on the counterfactual framework. For illustrative purposes, we examine the effect of having private health insurance on the probability of experiencing at least one hospital admission in the previous year. Using data from the 4th wave of the Korea Labor and Income Panel Study, we compared the results obtained using propensity score and instrumental variable methods with those from conventional logistic and linear regression models, respectively. While conventional multiple regression analyses fail to identify the effect, the results estimated using propensity score and instrumental variable methods suggest that having private health insurance has positive and statistically significant effects on hospital admission. This study demonstrates that propensity score and instrumental variable methods provide potentially useful alternatives to conventional regression approaches in making causal inferences using non-experimental data.

Estimation of Direct Causal Effects

Many common problems in epidemiologic and clinical research involve estimating the effect of an exposure on an outcome while blocking the exposure's effect on an intermediate variable. Effects of this kind are termed direct effects. Estimation of direct effects is typically the goal of research aimed at understanding mechanistic pathways by which an exposure acts to cause or prevent disease, as well as in many other settings. Although multivariable regression is commonly used to estimate direct effects, this approach requires assumptions beyond those required for the estimation of total causal effects. In addition, when the exposure and intermediate variables interact to cause disease, multivariable regression estimates a particular type of direct effect-the effect of an exposure on an outcome when the intermediate is fixed at a specified level. Using the counterfactual framework, we distinguish this definition of a direct effect (controlled direct effect) from an alternative definition, in which the effect of the exposure on the intermediate is blocked, but the intermediate is otherwise allowed to vary as it would in the absence of exposure (natural direct effect). We illustrate the difference between controlled and natural direct effects using several examples. We present an estimation approach for natural direct effects that can be implemented using standard statistical software, and we review the assumptions underlying our approach (which are less restrictive than those proposed by previous authors).

Counterfactual Models of Neighborhood Effects: The Effect of Neighborhood Poverty on Dropping Out and Teenage Pregnancy

This article investigates the causal effects of neighborhood on high school dropping out and teenage pregnancy within a counterfactual framework. It shows that when two groups of children, identical at age 10 on observed factors, experience different neighborhoods during adolescence, those in high‐poverty neighborhoods are more likely to drop out of high school and have a teenage pregnancy than those in low‐poverty neighborhoods. Causal inferences from such associations have been plagued by the possibility of selection bias. Using a new method for sensitivity analysis, these effects are shown to be robust to selection bias. Unobserved factors would have to be unreasonably strong to account for the associations between neighborhood and the outcomes.

The (mis)estimation of neighborhood effects: causal inference for a practicable social epidemiology

The resurgence of interest in the effect of neighborhood contexts on health outcomes, motivated by advances in social epidemiology, multilevel theories and sophisticated statistical models, too often fails to confront the enormous methodological problems associated with causal inference. This paper employs the counterfactual causal framework to illuminate fundamental obstacles in the identification, explanation, and usefulness of multilevel neighborhood effect studies. We show that identifying useful independent neighborhood effect parameters, as currently conceptualized with observational data, to be impossible. Along with the development of a dependency-based methodology and theories of social interaction, randomized community trials are advocated as a superior research strategy, one that may help social epidemiology answer the causal questions necessary for remediating disparities and otherwise improving the public's health.

THE ESTIMATION OF CAUSAL EFFECTS FROM OBSERVATIONAL DATA

▪ Abstract When experimental designs are infeasible, researchers must resort to the use of observational data from surveys, censuses, and administrative records. Because assignment to the independent variables of observational data is usually nonrandom, the challenge of estimating causal effects with observational data can be formidable. In this chapter, we review the large literature produced primarily by statisticians and econometricians in the past two decades on the estimation of causal effects from observational data. We first review the now widely accepted counterfactual framework for the modeling of causal effects. After examining estimators, both old and new, that can be used to estimate causal effects from cross-sectional data, we present estimators that exploit the additional information furnished by longitudinal data. Because of the size and technical nature of the literature, we cannot offer a fully detailed and comprehensive presentation. Instead, we present only the main features of methods that are accessible and potentially of use to quantitatively oriented sociologists.

A Note on Nonlocality, Causation, and Lorentz Invariance

The status of a causal approach to EPR-Bell nonlocal correlations in terms of a counterfactual framework for causation is considered. It is argued that when the relativistic spacetime structure of the events is taken into due account, the adoption of this approach is best motivated by the assumption of a preferred frame of reference, an assumption that seems even more in need of justification than the causal theory itself.