• 감성과학
• /
• 제26권1호
• /
• pp.55-64
• /
• 2023

본 연구는 범주속성들이 공통원인 혹은 공통효과 인과 네트워크로 연결되었을 때 인과강도에 따른 속성추론을 검증했다. 인과범주에서 속성추론을 검증한 기존 연구들은 인과관계의 방향, 연결된 속성의 개수, 원인 혹은 결과의 여부 등에 따라 고유한 추론 패턴이 나타남을 보여주었다. 다만 기존 연구들은 인과관계에 따른 추론패턴을 주로 탐색했으며 인과관계의 효과가 인과강도에 따라 어떤 변화를 보이는지 확인한 연구는 찾아보기 어렵다. 본 연구에서는 공통원인(실험 1), 공통효과(실험 2) 네트워크에서 인과강도에 따른 속성추론을 검증했다. 이를 위해 참가자들에게 속성들이 인과적 관련성을 가지는 범주를 학습하게 한 다음 속성추론 과제를 실시하도록 했다. 실험 결과 인과관계 뿐만 아니라 인과강도 역시 속성추론에 중요한 영향을 미쳤다. 인과강도가 강할 떄 공통원인 속성에 대해서는 추론이 약해진 반면 공통효과 속성에 대해서는 추론이 강해졌다. 또한 인과강도가 강할 때 공통원인이 존재하는 경우 결과속성들에 대한 추론이 강해진 반면 공통효과에서는 반대의 결과가 나타났다. 특히 공통효과에서는 인과강도가 강할 때 인과적 절감이 더 뚜렷하게 나타났다. 이 결과들은 인과적 범주에서의 속성추론에서 참가자들은 인과관계 뿐만 아니라 인과강도를 고려한다는 것을 일관성있게 보여준다.

• Communications for Statistical Applications and Methods
• /
• 제29권2호
• /
• pp.177-191
• /
• 2022

This paper addresses the use of machine learning methods for causal estimation of treatment effects from observational data. Even though conducting randomized experimental trials is a gold standard to reveal potential causal relationships, observational study is another rich source for investigation of exposure effects, for example, in the research of comparative effectiveness and safety of treatments, where the causal effect can be identified if covariates contain all confounding variables. In this context, statistical regression models for the expected outcome and the probability of treatment are often imposed, which can be combined in a clever way to yield more efficient and robust causal estimators. Recently, targeted maximum likelihood estimation and causal random forest is proposed and extensively studied for the use of data-adaptive regression in estimation of causal inference parameters. Machine learning methods are a natural choice in these settings to improve the quality of the final estimate of the treatment effect. We explore how we can adapt the design and training of several machine learning algorithms for causal inference and study their finite-sample performance through simulation experiments under various scenarios. Application to the percutaneous coronary intervention (PCI) data shows that these adaptations can improve simple linear regression-based methods.

• 한국시스템다이내믹스연구
• /
• 제6권1호
• /
• pp.5-15
• /
• 2005

This paper investigates psychological differences between constructors and interpreters of causal maps. This paper argues that dissipation effects and dilution effects applies to those who are to interpret causal maps not to those who construct them. Dissipation effects are psychological tendency that people perceive causal effect as weak as the number of causal links increases. Dilution effects occur when people undervalue the strength of causal relation as the number of causal variables increases. Experimental results show that concentration effects opposite to the dissipation effects and dilution effects explain more correctly the perception of constructors of causal maps. This paper points out that this asymmetric psychological tendencies between constructors and interpreters of causal maps is the psychological source of the communication problems between systems thinkers and their clients.

• 응용통계연구
• /
• 제27권7호
• /
• pp.1077-1095
• /
• 2014

관찰연구를 이용하여 인과관계를 추론할 경우 무작위 통제시험과는 달리 교란변수로 인한 편향을 제어하기 위한 통계적 전략이 필요하다. 최근에는 성향점수(propensity score) 를 이용한 짝짓기나 원인변수의 역확률을 가중치로 사용하는 주변구조모형이 제안되어 사용되고 있다. 이러한 인과관계 추론은 처치(treatment)가 명확히 주어진 경우에 교란변수를 통제하고 그 처치가 결과에 미치는 영향을 평가하는 방법에 초점이 맞추어져 있다. 하지만 기존의 방법의 경우 원인변수인 처치가 직접관측이 가능한 범주형 변수이고 결과변수 또한 직접관측이 가능한 변수인 경우에만 사용할 수 있는 한계를 갖고 있다. 본 연구에서는 원인변수인 처치와 결과변수의 결괏값의 직접적인 관측이 어려운 경우, 측정오차를 고려한 잠재범주모형(latent class analysis)의 변수로 모형화 함으로써 잠재범주 간의 원인적 영향력을 추정하는 방법을 제시하고자 한다. 그리고 미국의 The National Longitudinal Study of Adolescent Health 자료를 이용하여, 약물사용의 잠재범주에 대한 청소년기의 비행(delinquency)이라는 잠재범주의 원인적 영향력을 추정하였다.

• 논리연구
• /
• 제5권2호
• /
• pp.115-131
• /
• 2002

The problem of disjunctive causal factors is generalized as follows. Suppose that there are no factors of the kind considered so far that need to be held fixed in background contexts. Nevertheless, it is still possible that within the background contexts, each disjunct of a disjunctive causal factor X v W confers a different probability on an effect factor in Question. So a problem arises of how we identify a single causally significant probability of the effect factor in the presence of the disjunctive causal factor, assuming that each disjunct of the disjunctive causal factor confers a different probability on the effect factor. In this paper, I first introduce an experiment in which disjunctive causal factors seem to pose a problem for the theory of probabilistic causation. Second, I show how Eells' solution to the problem of disjunctive causal factors meets the problem that arises in the experiment. Third, I examine Hitchcock's arguments against Eells' solution, arguing that Hitchcock misconstrues Eells' solution, and disregards the feature of the theory of probabilistic causation such that a factor is a causal factor for another factor relative to a population P of a population type Q.

• Journal of Preventive Medicine and Public Health
• /
• 제56권4호
• /
• pp.303-311
• /
• 2023

Traditional mediation analysis, which relies on linear regression models, has faced criticism due to its limited suitability for cases involving different types of variables and complex covariates, such as interactions. This can result in unclear definitions of direct and indirect effects. As an alternative, causal mediation analysis using the counterfactual framework has been introduced to provide clearer definitions of direct and indirect effects while allowing for more flexible modeling methods. However, the conceptual understanding of this approach based on the counterfactual framework remains challenging for applied researchers. To address this issue, the present article was written to highlight and illustrate the definitions of causal estimands, including controlled direct effect, natural direct effect, and natural indirect effect, based on the key concept of nested counterfactuals. Furthermore, we recommend using 2 R packages, 'medflex' and 'mediation', to perform causal mediation analysis and provide public health examples. The article also offers caveats and guidelines for accurate interpretation of the results.

• 가정과삶의질연구
• /
• 제26권5호
• /
• pp.333-354
• /
• 2008

This study aimed to investigate the role of domain-specific causal mechanism information and domain-general conditional probability in young children's causal reasoning on psychology and biology. Participants were 121 3-year-olds and 121 4-year-olds recruited from seven childcare centers in Seoul, Kyonggi Province, and Busan. After participants watched moving pictures on psychological and biological phenomena, they were asked to choose appropriate cause and justify their choices. Results of this study were as follows: First, young children made different inferences according to domain-specific causal mechanisms. Second, the developmental level of causal mechanisms has a gap between psychology and biology, and biological knowledge was proved to be separate from psychological knowledge during the preschool period. Third, young children's causal reasoning was different depending on the interaction effect of domain-specific mechanisms and domain-general conditional probability: children could make more inferences based on domain-specific causal mechanisms if conditional probability between domain-appropriate cause and effect was evident. To conclude, it can be inferred that the role of domain-specific causal mechanisms and domain-general conditional probability is not competitive but complementary in young children's causal reasoning.

• 응용통계연구
• /
• 제32권2호
• /
• pp.173-185
• /
• 2019

과학적 연구에서 핵심적인 연구 주제 또는 가설은 대부분 인과적 질문(causal question)을 포함한다. 예를 들어, 전염병 예방을 위한 치료법의 효과 연구, 특정 정책의 시행으로 인한 효용(utility)의 평가에 대한 연구, 특정 사용자를 대상으로 노출된 광고의 종류에 따른 광고의 효과성에 대한 연구는 모두 인과 관계(causal relationship)의 추론이 요구된다. 이러한 인과 관계를 다루는 통계적 인과 추론(statistical causal inference)의 주요 관심사 중 하나는 모집단에 일종의 개입(정책 혹은 처치)을 적용한 후 개입의 효과를 정확하게 추정하는 것이다. 인과 추론은 임상실험과 정책결정에서 주로 이용되었으나, 이른바 빅데이터 시대의 도래로 가용한 관측자료가 폭발적으로 증가하였고 이로 인하여 인과 추론에 대한 잠재적 응용가치와 수요가 지속적으로 증가하고 있다. 하지만 가용한 대부분의 자료는 임의실험 기반의 자료와 달리 개입이 임의로 분배되지 않은 비실험 관측자료이다. 따라서, 본 논문은 비실험 관측자료로부터 개입의 효과를 추정하기 위한 인과 추론의 핵심 개념과 최근의 연구동향을 소개하고자 한다. 이를 위하여 본문에서는 먼저 개입의 효과를 Neyman-Rubin의 잠재 결과(potential outcome) 모형으로 나타내고, 개입의 효과를 추정하는 여러 접근법 중 특히 성향점수(propensity score) 기반 추정법과 회귀모형 기반 추정법을 중점적으로 소개한다. 최근 연구동향으로는 (1) 평균 효과 크기 추정을 넘어선 개인별 효과 크기의 추정, (2) 효과크기 추정에 있어서 자료 규모의 증대로 인한 차원의 저주가 야기하는 난제들과 이에 대한 해결방안들, (3) 복합적 인과관계를 반영하기 위한 Pearl의 구조적 인과 모형(structural causal model) 및 잠재 결과 모형과의 비교의 3가지 주제로 구분하여 소개한다.

• 보건행정학회지
• /
• 제8권1호
• /
• pp.52-96
• /
• 1998

To find causal factors and improvement plans of organizational commitment of public servants in urban health centers, a hypothetical causal model, which included 2 endogenous variables(organizational commitment & organizational satisfaction) and 15 exogenous variables, was constructed. Exogenous variables consisted of individual factors (sex, age, education, job-grade, and annual salary), psychological variables(pride for organization, extrinsic motivation, intrinsic motivation and support of supervisor) ad structural variables(formalization, centralization, communication, job-conflict, job-decision, and workload). In the hypothetical causal model, organizational commitment was supposed to be effect variable, and organizational satisfaction was presumed to be intervening variable to mediate between organizational commitment and exogenous variables. For data collection, cross-sectional self-administered questionnaire survey was conducted to 1,295 public servants from 32 urban health centers nationwide. The survey responses were from 934, 72.1% of subjects. But 756 responses(58.4%) were analyzed because of excluding ones with missing values. The hypothetical causal model was fitted by covariance structural analysis with maximum likelihood method. Main results were as follows: (1) The fitted causal model accounted for 33 and 55 percent of total variance of organizational commitment and organizational satisfaction of public servants, respectively. (2) In order of effect size, pride for organization, supervisor support, communication, extrinsic motivation and centralization had an indirect effect effect on organizational commitment through organizational satisfaction. However, the effect of centralization was negative. (3) Pride for organiztion, intrinsic motivation, organizational satisfaction, job-conflict, supervisor support, communication, age, centralization, annual salar and extrinsic motivation had indirect or direct effects on organizational commitment in order of effect size. Among them, effects of job-conflict and centraldization were negative. In conclusion, these results suggested that organizational commitment of public servants in urban health centers could be enhanced by pride for organization, intrinsic and extrinsic motivations, prevention of job-conflict and excess centralization, supervisor support and active communication. Especially, pride for organization and intrinsic motivation were expected to play the most important role.

• 아동학회지
• /
• 제29권5호
• /
• pp.243-269
• /
• 2008

The role of domain-specific causal mechanism information and domain-general conditional probability in young children's causal reasoning on physics and psychology was investigated with the participation of 121 3-year-olds and 121 4-year-olds recruited from seven child care centers in Seoul, Kyonggi Province, and Busan. Children watched moving pictures on physical and psychological phenomena, and were asked to choose an appropriate cause and justify their choice. Results showed that young children's causal reasoning differed depending on domain-specific mechanism. In addition, their causal reasoning on physics and psychology differed by the developmental level of causal mechanism. The interaction of domain-specific mechanism and domain-general conditional probability influenced children's causal reasoning : evident conditional probability between domain-appropriate cause and effect helped children make more inferences based on domain-specific causal mechanism.