Counterfactual Model Research Articles

Causal superseding

When agents violate norms, they are typically judged to be more of a cause of resulting outcomes. In this paper, we suggest that norm violations also affect the causality attributed to other agents, a phenomenon we refer to as “causal superseding.” We propose and test a counterfactual reasoning model of this phenomenon in four experiments. Experiments 1 and 2 provide an initial demonstration of the causal superseding effect and distinguish it from previously studied effects. Experiment 3 shows that this causal superseding effect is dependent on a particular event structure, following a prediction of our counterfactual model. Experiment 4 demonstrates that causal superseding can occur with violations of non-moral norms. We propose a model of the superseding effect based on the idea of counterfactual sufficiency.

Constructing Full Adult Life-cycles from Short Panels

In this paper we discuss two alternative approaches to constructing complete adult life-cycles using data from an 18-year panel. The first of these is a splicing approach - closely related to imputation - that involves stitching together individuals observed at different ages. The second is a microsimulation approach that uses panel data to estimate transition probabilities between different states at adjacent ages and then simulates a large number of individuals with different initial values. Our aim throughout is to construct life-cycle profiles of employment, earnings and family circumstances that are representative of UK individuals born between 1945 and 1954. On balance, we find the microsimulation approach is to be preferred because it allows us to correct for observable differences across cohorts, and it is more amenable to counterfactual modelling.

Do Alcohol Excise Taxes Affect Traffic Accidents? Evidence From Estonia

Objective: This article examines the association between alcohol excise tax rates and alcohol-related traffic accidents in Estonia.Methods: Monthly time series of traffic accidents involving drunken motor vehicle drivers from 1998 through 2013 were regressed on real average alcohol excise tax rates while controlling for changes in economic conditions and the traffic environment. Specifically, regression models with autoregressive integrated moving average (ARIMA) errors were estimated in order to deal with serial correlation in residuals. Counterfactual models were also estimated in order to check the robustness of the results, using the level of non-alcohol-related traffic accidents as a dependent variable.Results: A statistically significant (P <.01) strong negative relationship between the real average alcohol excise tax rate and alcohol-related traffic accidents was disclosed under alternative model specifications. For instance, the regression model with ARIMA (0, 1, 1)(0, 1, 1) errors revealed that a 1-unit increase in the tax rate is associated with a 1.6% decrease in the level of accidents per 100,000 population involving drunk motor vehicle drivers. No similar association was found in the cases of counterfactual models for non-alcohol-related traffic accidents.Conclusions: This article indicates that the level of alcohol-related traffic accidents in Estonia has been affected by changes in real average alcohol excise taxes during the period 1998–2013. Therefore, in addition to other measures, the use of alcohol taxation is warranted as a policy instrument in tackling alcohol-related traffic accidents.

Hospital Volume and Patient Outcomes in Hepato-Pancreatico-Biliary Surgery: Is Assessing Differences in Mortality Enough?

BackgroundThe impact of regionalization on morbidity, failure to rescue (FTR), length of stay (LOS), and readmission remains unclear. We sought to examine hospital-volume-related differences in outcomes following complex hepato-pancreatico-biliary (HPB) surgery and define potential benefits of regionalization across quality metrics. MethodsPatients undergoing HPB surgery in the Surveillance, Epidemiology and End Results (SEER)-Medicare linked data from 1986 to 2002 were identified. Hospital volume was stratified into tertiles (low volume [LV] <4 cases/year; intermediate volume [IV] 4–10 cases/year; high volume [HV] ≥11 cases/year). The incidence of complications, FTR (mortality following a complication), and LOS was compared across hospital-volume strata. A counterfactual model examined hypothetical outcomes assuming all patients had been treated at HV centers. ResultsTen thousand two hundred eight patients underwent pancreatic (46.1 %), hepatic (36.2 %), or biliary (17.8 %) procedures. Overall mean age ranged from 72.7 years at HV centers to 73.4 at LV centers (P < 0.001), and patients at HV centers (75.4 %) were more likely to have ≥3 comorbidities versus IV (70.0 %) or LV (64.7 %) centers (P < 0.001). The incidence of post-operative complications was lower at HV (39.1 %) compared with IV (41.9 %) or LV (44.8 %) centers. Major complications included hemorrhagic anemia (7.3 %), failure to thrive (5.1 %), and respiratory infection/failure (3.5 %); each was less common in HV hospitals (P < 0.05). FTR after major complication tended to be higher at LV (36.7 %) and IV (37.3 %) hospitals compared with HV hospitals (29.7 %) (P = 0.10). Mortality was higher at LV (10.5 %) and IV (8.1 %) hospitals versus HV centers (5.4 %) (P < 0.001). HV hospital patients had shorter median LOS (10 days) compared with IV (12 days) or LV (12 days) hospital patients (P < 0.001). Readmission varied across centers (HV 19.1 % vs. IV 19.2 % vs. 16.7 %; P = 0.02). In a counterfactual model with all patients treated at a HV center, 6.4 % fewer complications and a 26.0 % increase in post-complication rescue would be expected, along with a 32.0 % reduction in index mortality and an 8.1 % reduction in total patient-days. A minor increase in readmissions (7.1 %) would be anticipated with 13.3 % fewer deaths during readmission. ConclusionAlthough patients treated at HV hospitals had more medical comorbidities, outcomes across a wide spectrum of quality metrics were better than at IV or LV hospital following complex HPB surgery. A 20–30 % reduction in morbidity and mortality and an 8 % reduction in hospital patient-days could be anticipated had all patients been treated at HV hospitals.

Universal Service Obligations and Public Payphone Use: Is Regulation Still Necessary in the Era of Mobile Telephony?

In this paper, we analyse empirically whether universal service obligations for the provision of public payphones is still relevant in the European electronic communications market. We use micro-level data on 106,989 households from 27 EU countries, during the period 2005 to 2009. We first estimate the impact of universal service obligations on public payphone use. Then, we set up a counterfactual scenario to predict the average use of public payphones if no universal service obligations were imposed in the EU countries. We find that universal service obligations do not have any significant impact on public payphone use. Only coverage obligations, i.e., when the law imposes a minimum number of payphones depending on the density of population, have a positive, though low, impact on payphone use. Our counterfactual model predicts that if universal service obligations had been absent, between 2005 and 2009 public payphone use would have been 15 percent lower.

Counter-factual mathematics of counterfactual predictive models.

Counter-factual mathematics of counterfactual predictive models.

Promoting clean energy investment: An empirical analysis of property assessed clean energy

From 2008 to 2010 a handful of Property-Assessed Clean Energy (PACE) programs offered property-secured loans to homeowners for residential clean energy investments. This analysis uses difference-in-differences models and synthetic counterfactual models to estimate the effect of three California PACE programs on residential photovoltaic installations. While PACE programs do not offer superior terms to other solar financing options, we find that PACE financing increases solar installations by approximately 3.8watts per owner-occupied household per quarter, a 108% increase over the mean watts per owner-occupied household.

The Relationship Between Cycle Tourism and Sustainable Transport in the UK

The use of both sustainable transport outcomes and tourism impacts as a mixed justification for investment in cycling infrastructure has led to the two often being seen as synonymous. The environmentally friendly credentials of cycle tourism are predicated on a conceptualization of cycle tourism in which cycling as a form of transport supplants other energy-consuming and -polluting forms of transport within the tourism trip. However, using a recent meta-analysis of UK data, this research note shows that even when the environmental costs of major cycling events are excluded, in absolute terms recreational cycle tourism across its full range of forms in the UK still generates considerable motorized transport use. But, the use of counterfactual models shows that in relative terms, on average across all its forms, recreational cycle tourism in the UK reduces the use of motorized transport to get to and from destinations by 12.2% and reduces motorized transport use at destinations by 7.6%. Consequently, recreational cycle tourism in the UK does have a positive sustainable transport outcome, but this is far smaller than is often visualized.

Evidence-Based Practices in a Changing World

Experimental and quasi-experimental designs are used in educational research to establish causality and develop effective practices. These research designs rely on a counterfactual model that, in simple form, calls for a comparison between a treatment group and a control group. Developers of educational practices often assume that the population from which control groups are drawn is unchanging in its behavior or performance. This is not always the case. Populations and study samples can change over time—sometimes dramatically so. We illustrate this important point by presenting data from 5 randomized control trials of the efficacy of Kindergarten Peer-Assisted Learning Strategies, a supplemental, peer-mediated reading program. The studies were conducted across 9 years and involved 2,591 students. Findings demonstrate a dramatic increase in the performance of control students over time, and suggest the need for a more nuanced understanding of the counterfactual model and its role in establishing evidence-based practices.

The juvenile crime dilemma

I develop a dynamic model of behavior to analyze juvenile crime. Forward-looking youths consistently decide between crime and legal activities depending on their endowment of work- and crime-specific human capital, which in turn is shaped by their history of past choices. The model explicitly recognizes the contrasting levels of punishment of the juvenile and adult criminal systems. In order to evaluate whether the model explains the evolution of crime, I calibrate it and test whether it can account for the observed variations in juvenile crime levels across changes in economic and legal conditions. The model is able to reproduce 91 percent of the recent increase in juvenile crime in Uruguay by affecting key model parameters in line with observed facts (a decrease in the relative returns of legal activities and the introduction of a lenient juvenile crime regulation and enforcement strategy). Counterfactual model results predict that a reduction in the age of criminal majority would significantly lower juvenile crime involvement. However, if the transmission of crime-related skills in correctional facilities were strong enough, harsher punishments to juveniles would increase the likelihood of criminal involvement later in life.

Introduction: Legislation, Legal Episteme, and Empirical Knowledge

This is the introduction to a special issue of The Theory and Practice of Legislation [formerly Legisprudence] I edited on Empirical Knowledge, Common Sense, and Legal Cognition. It puts the contributions in the special issue in context by drawing attention to some current trends in lawmaking practices as well as academic research. On the one hand, reference to empirical research in e.g. behavioural economics is increasingly frequent in both policymaking and legal reforms. On the other hand, the use of scientific empirical knowledge in legislation is neither systematic nor uncontroversial. It is suggested that law resists empirical knowledge for three main reasons: epistemic (related to how empirical knowledge is generated), institutional (how this knowledge is channelled and transferred into legislation or other legal procedures) and normative (related to law’s commitment to counterfactual models of human behaviour and non-instrumental goals).

Introduction: Legislation, Legal Episteme, and Empirical Knowledge

This is the introduction to a special issue of The Theory and Practice of Legislation [formerly Legisprudence] I edited on Empirical Knowledge, Common Sense, and Legal Cognition. It puts the contributions in the special issue in context by drawing attention to some current trends in lawmaking practices as well as academic research. On the one hand, reference to empirical research in e.g. behavioural economics is increasingly frequent in both policymaking and legal reforms. On the other hand, the use of scientific empirical knowledge in legislation is neither systematic nor uncontroversial. It is suggested that law resists empirical knowledge for three main reasons: epistemic (related to how empirical knowledge is generated), institutional (how this knowledge is channelled and transferred into legislation or other legal procedures) and normative (related to law’s commitment to counterfactual models of human behaviour and non-instrumental goals).

Regional innovation policy and the performance of small enterprises

Purpose of the study . The paper aims at disentangling the role of a R&D place based policy implemented in the Italian province of Trento, during the period 2002-2007 in spurring R&D investment of private firms and in enhancing the firm performances. Methodology . This paper proposes an investigation based on an original and unique database about the population of limited liabilities firm operating in Trentino. The econometric analysis is based on counterfactual models. In particular, we make use of the Conditional-Difference-in-Differences estimator. Findings . We find that R&D incentives positively affect investments in intangible assets and human capital, while they have no effect on firms’ turnover, labor productivity and profitability. Research limits . Two issues are not fully studied in the present paper: the role of micro-complementarities and the role of localized knowledge spillovers. which may well be the main advantage of a place-based policy Practical implications . The results on additionality allow to shed light about the degree of success of the direct measures of public interventions (subsidies). Moreover, the findings can be used to fine tuning the industrial policies. Originality of the study . The context of analysis allows us to disaggregate the different components of R&D spending of firms, namely labor force and its quality, intangibles, fixed assets. Indeed, The study is based on exhaustive information about financial R&D subsidies the population of the firms subsidized during the years 2002-2007 in Trentino matched with the population of limited liabilities firms operating in Trentino during the same time window.

Designing observational biologging studies to assess the causal effect of instrumentation

Summary Biologging has improved ecological knowledge on an increasing number of species for more than 2 decades. Most studies looking at the incidence of tags on behavioural, physiological or demographic parameters rely on ‘control’ individuals chosen randomly within the population, assuming that they will be comparable with equipped individuals. This assumption is usually untestable and untenable since biologging studies are more observational than experimental, and often involve small sample sizes. Notably, background characteristics of wild animals are, most of the time, unknown. Consequently, investigating any causal effect of instrumentation is a difficult task, subjected to hidden biases. We describe the counterfactual model to causal inference which was implicit in early biologging studies. We adopted methods developed in social and political sciences to construct a posteriori an appropriate control group. Using biologging data collected on Scopoli's shearwaters (Calonectris diomedea) from a small Mediterranean island, we used this method to achieve objective causal inference on the effect of instrumentation on breeding performance and divorce. Our method revealed that the sample of instrumented birds was nonrandom. After identification of a relevant control group, we found no carry‐over effects of instrumentation on breeding performance (taking into account imperfect detection probability) or divorce rate in Scopoli's shearwaters. Randomly chosen control groups can be both counterproductive and ethically dubious via unnecessary additional disturbance of populations. The counterfactual approach, which can correct for selection bias, has wide applicability to biologging within long‐term studies.

STRUCTURAL EQUATIONS AND BEYOND

AbstractRecent accounts of actual causation are stated in terms of extended causal models. These extended causal models contain two elements representing two seemingly distinct modalities. The first element are structural equations which represent the “(causal) laws” or mechanisms of the model, just as ordinary causal models do. The second element are ranking functions which represent normality or typicality. The aim of this paper is to show that these two modalities can be unified. I do so by formulating two constraints under which extended causal models with their two modalities can be subsumed under so called “counterfactual models” which contain just one modality. These two constraints will be formally precise versions of Lewis’ (1979) familiar “system of weights or priorities” governing overall similarity between possible worlds.

Strategies to Improve Child Immunization via Antenatal Care Visits in India: A Propensity Score Matching Analysis.

Numerous studies have examined the empirical evidence concerning the influence of demographic and socio-economic factors influencing child immunization, but no documentation is available which shows the actual impact of antenatal care (ANC) visits on subsequent child immunization. Therefore, this paper aims to examine the net impact of ANC visits on subsequent utilization of child immunization after removing the presence of selection bias. Nationwide data from India’s latest National Family Health Survey conducted during 2005–06 is used for the present study. The analysis has been carried out in the two separate models, in the first model 1–2 ANC visit and in the second model three or more ANC visits has been compared with no visit. We have used propensity score matching method with a counterfactual model that assesses the actual ANC visits effect on treated (ANC visits) and untreated groups (no ANC visit), and have employed Mantel-Haenszel bounds to examine whether result would be free from hidden bias or not. Using matched sample analysis result shows that child immunization among the groups of women who have completed 1–2 ANC visits and those who had more than two visits was about 13 percent and 19 percent respectively, higher than the group of women who have not made any ANC visit. Findings of nearest neighbor matching with replacement method, which completely eliminated the bias, indicate that selection bias present in data set leads to overestimates the positive effects of ANC visits on child immunization. Result based on Mantel-Haenszel bounds method suggest that if around 19 percent bias would be involved in the result then also we could observe the true positive effect of 1–2 ANC visits on child immunization. This also indicates that antenatal clinics are the conventional platforms for educating pregnant women on the benefits of child immunization.

Improving Causal Inferences in Risk Analysis

Recent headlines and scientific articles projecting significant human health benefits from changes in exposures too often depend on unvalidated subjective expert judgments and modeling assumptions, especially about the causal interpretation of statistical associations. Some of these assessments are demonstrably biased toward false positives and inflated effects estimates. More objective, data-driven methods of causal analysis are available to risk analysts. These can help to reduce bias and increase the credibility and realism of health effects risk assessments and causal claims. For example, quasi-experimental designs and analysis allow alternative (noncausal) explanations for associations to be tested, and refuted if appropriate. Panel data studies examine empirical relations between changes in hypothesized causes and effects. Intervention and change-point analyses identify effects (e.g., significant changes in health effects time series) and estimate their sizes. Granger causality tests, conditional independence tests, and counterfactual causality models test whether a hypothesized cause helps to predict its presumed effects, and quantify exposure-specific contributions to response rates in differently exposed groups, even in the presence of confounders. Causal graph models let causal mechanistic hypotheses be tested and refined using biomarker data. These methods can potentially revolutionize the study of exposure-induced health effects, helping to overcome pervasive false-positive biases and move the health risk assessment scientific community toward more accurate assessments of the impacts of exposures and interventions on public health.

Counterfactuals, Autonomy and Downward Causation: Reply to Zhong

In recent papers, Lei Zhong argues that the autonomy solution to the causal exclusion problem is unavailable to anyone that endorses the counterfactual model of causation. The linchpin of his argument is that the counterfactual theory entails the downward causation principle, which conflicts with the autonomy solution. In this note I argue that the counterfactual theory does not entail the downward causation principle, so it is possible to advocate for the autonomy solution to the causal exclusion problem from within the counterfactual theory of causation.

Selection Effects and Prevention Program Outcomes

A primary goal of the paper is to provide an example of an evaluation design and analytic method that can be used to strengthen causal inference in nonexperimental prevention research. We used this method in a nonexperimental multisite study to evaluate short-term outcomes of a preventive intervention, and we accounted for effects of two types of selection bias: self-selection into the program and differential dropout. To provide context for our analytic approach, we present an overview of the counterfactual model (also known as Rubin's causal model or the potential outcomes model) and several methods derived from that model, including propensity score matching, the Heckman two-step approach, and full information maximum likelihood based on a bivariate probit model and its trivariate generalization. We provide an example using evaluation data from a community-based family intervention and a nonexperimental control group constructed from the Washington State biennial Healthy Youth Survey (HYS) risk behavior data (HYS n = 68,846; intervention n = 1,502). We identified significant effects of participant, program, and community attributes in self-selection into the program and program completion. Identification of specific selection effects is useful for developing recruitment and retention strategies, and failure to identify selection may lead to inaccurate estimation of outcomes and their public health impact. Counterfactual models allow us to evaluate interventions in uncontrolled settings and still maintain some confidence in the internal validity of our inferences; their application holds great promise for the field of prevention science as we scale up to community dissemination of preventive interventions.

Social sector expenditure and child mortality in India: a state-level analysis from 1997 to 2009.

BackgroundIndia is unlikely to meet the Millennium Development Goal for child mortality. As public policy impacts child mortality, we assessed the association of social sector expenditure with child mortality in India.Methods and FindingsMixed-effects regression models were used to assess the relationship of state-level overall social sector expenditure and its major components (health, health-related, education, and other) with mortality by sex among infants and children aged 1–4 years from 1997 to 2009, adjusting for potential confounders. Counterfactual models were constructed to estimate deaths averted due to overall social sector increases since 1997. Increases in per capita overall social sector expenditure were slightly higher in less developed than in more developed states from 1997 to 2009 (2.4-fold versus 2-fold), but the level of expenditure remained 36% lower in the former in 2009. Increase in public expenditure on health was not significantly associated with mortality reduction in infants or at ages 1–4 years, but a 10% increase in health-related public expenditure was associated with a 3.6% mortality reduction (95% confidence interval 0.2–6.9%) in 1–4 years old boys. A 10% increase in overall social sector expenditure was associated with a mortality reduction in both boys (6.8%, 3.5–10.0%) and girls (4.1%, 0.8–7.5%) aged 1–4 years. We estimated 119,807 (95% uncertainty interval 53,409 – 214,662) averted deaths in boys aged 1–4 years and 94,037 (14,725 – 206,684) in girls in India in 2009 that could be attributed to increases in overall social sector expenditure since 1997.ConclusionsFurther reduction in child mortality in India would be facilitated if policymakers give high priority to the social sector as a whole for resource allocation in the country’s 5-year plan for 2012–2017, as public expenditure on health alone has not had major impact on reducing child mortality.