{"doi":"10.1017/s0033291720005127","title":"Causal inference with observational data: the need for triangulation of evidence","abstract":"<jats:title>Abstract</jats:title><jats:p>The goal of much observational research is to identify risk factors that have a causal effect on health and social outcomes. However, observational data are subject to biases from confounding, selection and measurement, which can result in an underestimate or overestimate of the effect of interest. Various advanced statistical approaches exist that offer certain advantages in terms of addressing these potential biases. However, although these statistical approaches have different underlying statistical assumptions, in practice they cannot always completely remove key sources of bias; therefore, using design-based approaches to improve causal inference is also important. Here it is the design of the study that addresses the problem of potential bias – either by ensuring it is not present (under certain assumptions) or by comparing results across methods with different sources and direction of potential bias. The distinction between statistical and design-based approaches is not an absolute one, but it provides a framework for <jats:italic>triangulation</jats:italic> – the thoughtful application of <jats:italic>multiple</jats:italic> approaches (e.g. statistical and design based), each with their own strengths and weaknesses, and in particular sources and directions of bias. It is unlikely that any single method can provide a definite answer to a causal question, but the triangulation of evidence provided by different approaches can provide a stronger basis for causal inference. Triangulation can be considered part of wider efforts to improve the transparency and robustness of scientific research, and the wider scientific infrastructure and system of incentives.</jats:p>","journal":"Psychological Medicine","year":2021,"id":636831,"datarank":0.8329439377382427,"base_score":5.552959584921617,"endowment":5.552959584921617,"self_citation_contribution":0.8329439377382427,"citation_network_contribution":0.0,"self_endowment_contribution":0.8329439377382427,"citer_contribution":0.0,"corpus_percentile":null,"corpus_rank":null,"citation_count":257,"citer_count":0,"citers_with_citation_signal":0,"citers_with_endowment":0,"datacite_reuse_total":0,"is_dataset":false,"is_dataset_confidence":null,"is_data_producer":false,"deposit_databanks":null,"is_oa":false,"file_count":0,"downloads":0,"has_version_chain":false,"published_date":null,"fair_score":null,"fair_percentile":null,"algorithm_id":"datarank_citation_only_1hop_v6","ranking_scope":"data_only","authors":[{"id":6136,"name":"Marcus Robert Munafo","orcid":"0000-0002-4049-993X","position":1,"is_corresponding":false},{"id":570732,"name":"Gemma Hammerton","orcid":"0000-0002-7781-3857","position":0,"is_corresponding":false}],"reference_count":0,"raw_metadata":{"has_enrichment":true,"resolved":true,"title":"Causal inference with observational data: the need for triangulation of evidence","abstract":"<jats:title>Abstract</jats:title><jats:p>The goal of much observational research is to identify risk factors that have a causal effect on health and social outcomes. However, observational data are subject to biases from confounding, selection and measurement, which can result in an underestimate or overestimate of the effect of interest. Various advanced statistical approaches exist that offer certain advantages in terms of addressing these potential biases. However, although these statistical approaches have different underlying statistical assumptions, in practice they cannot always completely remove key sources of bias; therefore, using design-based approaches to improve causal inference is also important. Here it is the design of the study that addresses the problem of potential bias – either by ensuring it is not present (under certain assumptions) or by comparing results across methods with different sources and direction of potential bias. The distinction between statistical and design-based approaches is not an absolute one, but it provides a framework for <jats:italic>triangulation</jats:italic> – the thoughtful application of <jats:italic>multiple</jats:italic> approaches (e.g. statistical and design based), each with their own strengths and weaknesses, and in particular sources and directions of bias. It is unlikely that any single method can provide a definite answer to a causal question, but the triangulation of evidence provided by different approaches can provide a stronger basis for causal inference. Triangulation can be considered part of wider efforts to improve the transparency and robustness of scientific research, and the wider scientific infrastructure and system of incentives.</jats:p>","is_dataset_classified":null,"base_score":5.552959584921617,"endowment":5.552959584921617,"datacite_reuse_total":0,"file_count":0,"downloads":0,"views":0,"has_version_chain":false,"is_dataset":false,"is_oa":false,"pmid":"33682654","pmcid":"PMC8020490","openalex_id":"https://openalex.org/W3135774079","authors":[],"funders":[{"funder_name":"Medical Research Council","grant_id":"MC_UU_00011/7","title":"Causes, Consequences and Modification of Health Behaviour"},{"funder_name":"Wellcome Trust","grant_id":"209138/Z/17/Z","title":null},{"funder_name":"Wellcome Trust","grant_id":"209138","title":"An investigation of the causal pathways from childhood conduct problems to poor health outcomes and criminal behaviour in adulthood"},{"funder_name":"Wellcome Trust","grant_id":"","title":null},{"funder_name":"Wellcome Trust","grant_id":"","title":null}],"total_grants":5,"fwci":29.992,"citation_percentile":0.99846626,"influential_citations":0,"citation_trend":[{"year":2021,"count":5},{"year":2022,"count":21},{"year":2023,"count":46},{"year":2024,"count":67},{"year":2025,"count":76},{"year":2026,"count":42}],"oa_status":"hybrid","license":"cc-by","oa_locations":[{"url":"https://www.cambridge.org/core/services/aop-cambridge-core/content/view/AF5F7918753DF50F26B1D49561F0DF83/S0033291720005127a.pdf/div-class-title-causal-inference-with-observational-data-the-need-for-triangulation-of-evidence-div.pdf","host_type":"journal"},{"url":"https://www.cambridge.org/core/services/aop-cambridge-core/content/view/AF5F7918753DF50F26B1D49561F0DF83/S0033291720005127a.pdf/div-class-title-causal-inference-with-observational-data-the-need-for-triangulation-of-evidence-div.pdf","host_type":"publisher"},{"url":"https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S0033291720005127","host_type":"publisher"},{"url":"https://doi.org/10.1017/s0033291720005127","host_type":"journal"},{"url":"https://pubmed.ncbi.nlm.nih.gov/33682654","host_type":"repository"},{"url":"https://research-information.bris.ac.uk/en/publications/d591c3a5-c0dd-4d45-960e-db868fa18c4c","host_type":"repository"},{"url":"https://www.ncbi.nlm.nih.gov/pmc/articles/8020490","host_type":"repository"},{"url":"https://research-information.bris.ac.uk/ws/files/270757833/causal_inference_with_observational_data_the_need_for_triangulation_of_evidence.pdf","host_type":"repository"},{"url":"https://europepmc.org/articles/PMC8020490","host_type":"Europe_PMC"},{"url":"https://europepmc.org/articles/PMC8020490?pdf=render","host_type":"Europe_PMC"},{"url":"http://dx.doi.org/10.1017/S0033291720005127","host_type":""},{"url":"https://dx.doi.org/10.1017/s0033291720005127","host_type":""},{"url":"https://doi.org/10.1017/S0033291720005127","host_type":""},{"url":"https://hdl.handle.net/1983/d591c3a5-c0dd-4d45-960e-db868fa18c4c","host_type":""},{"url":"https://doi.org/https://doi.org/10.1017/S0033291720005127","host_type":""}],"fields_of_study":["Advanced Causal Inference Techniques","Health Systems, Economic Evaluations, Quality of Life","Statistical Methods and Bayesian Inference","03 medical and health sciences","0302 clinical medicine","Bias","Causality","Humans","Models, Statistical","Observational Studies as Topic","Research Design"],"mesh_terms":["Humans","Research Design","Models, Statistical","Bias","Causality","Observational Studies as Topic"],"keywords":["Causal inference","Observational study","Statistical inference","Robustness (evolution)","Inference","Computer science","Data science","Triangulation","Selection bias","Strengths and weaknesses","Confounding","Econometrics","Observational methods in psychology","Causal structure","Management science","Machine learning","Psychology","Artificial intelligence","Statistics","Mathematics","Social psychology","Engineering","Mental health","epidemiology","Observational Data","Models, Statistical","330","Invited Review Article","610","Causality","Observational Studies as Topic","Bias","Research Design","Humans","/dk/atira/pure/core/keywords/targ; name=TARG"],"sdg_mappings":[{"sdg_number":2,"sdg_label":"2. Zero hunger"},{"sdg_number":16,"sdg_label":"16. Peace & justice"},{"sdg_number":0,"sdg_label":"Industry, innovation and infrastructure"}],"linked_datasets":[],"clinical_trials":[],"software_tools":[],"database_accessions":[{"name":"doi"}],"source":"live","citation_network_status":"fetched"},"created_at":"2026-08-06T17:48:43.596855Z","pmid":null,"pmcid":null,"fwci":null,"citation_percentile":null,"influential_citations":0,"oa_status":null,"license":null,"views":0,"total_file_size_bytes":0,"version_count":0,"fair_f":null,"fair_a":null,"fair_i":null,"fair_r":null,"fair_zscore":null,"fair_rationale":null,"fair_model":null,"fair_agent_version":null,"fair_fulltext_source":null,"fair_has_llm":null,"fair_computed_at":null,"clinical_trials":[],"software_tools":[],"db_accessions":[],"linked_datasets":[],"topics":[]}