{"title":"Using Counterfactual Worlds to Triangulate Evidence in the Real World","authors":"Jeremy A. Labrecque, Sonja A. Swanson","doi":"10.1007/s40471-023-00340-0","DOIUrl":null,"url":null,"abstract":"<h3 data-test=\"abstract-sub-heading\">Purpose of Review</h3><p>When a causal question can be reasonably approached using more than one set of causal assumptions (i.e., using different identification strategies), triangulation can be used to compare estimates relying on these different assumption sets to gain insight into the validity of the causal assumptions used. This review covers the current understanding of triangulation from a counterfactual causal inference perspective.</p><h3 data-test=\"abstract-sub-heading\">Recent Findings</h3><p>We use counterfactuals to clarify and supplement the current understanding of triangulation. We propose a counterfactual definition of triangulation, propose assumptions on which triangulation relies, and discuss important practical issues such as triangulation with different estimands and the role of random error. Lastly, we examine two published examples of triangulation to illustrate these points.</p><h3 data-test=\"abstract-sub-heading\">Summary</h3><p>Triangulation, by leveraging causal inference reasoning and substantive knowledge, can potentially allow us to gain more insight into the validity of causal assumptions underlying many study designs than we would by considering each study design in isolation.</p>","PeriodicalId":48527,"journal":{"name":"Current Epidemiology Reports","volume":"1 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2023-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Current Epidemiology Reports","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1007/s40471-023-00340-0","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Purpose of Review
When a causal question can be reasonably approached using more than one set of causal assumptions (i.e., using different identification strategies), triangulation can be used to compare estimates relying on these different assumption sets to gain insight into the validity of the causal assumptions used. This review covers the current understanding of triangulation from a counterfactual causal inference perspective.
Recent Findings
We use counterfactuals to clarify and supplement the current understanding of triangulation. We propose a counterfactual definition of triangulation, propose assumptions on which triangulation relies, and discuss important practical issues such as triangulation with different estimands and the role of random error. Lastly, we examine two published examples of triangulation to illustrate these points.
Summary
Triangulation, by leveraging causal inference reasoning and substantive knowledge, can potentially allow us to gain more insight into the validity of causal assumptions underlying many study designs than we would by considering each study design in isolation.