Content deleted Content added
Rescuing 2 sources and tagging 0 as dead. #IABot (v1.6.2) (Balon Greyjoy) |
A section titled "Criticisms" had one sentence citing a poorly-informed NYTimes op-ed. This heading was changed to the more accurate "Limitations," and expounded upon slightly. (i.e. the focus remains the same, but the necessary nuance was added.) |
||
Line 61:
# [[Sensitivity (tests)]] and [[Specificity (tests)]]
==
Epidemiological (and other observational) studies typically highlight ''associations'' between exposures and outcomes, rather than causation. While some consider this a limitation of observational research, epidemiological models of causation (e.g. Bradford Hill criteria)<ref>{{Cite journal|last=Fedak|first=Kristen M.|last2=Bernal|first2=Autumn|last3=Capshaw|first3=Zachary A.|last4=Gross|first4=Sherilyn|date=2015-09-30|title=Applying the Bradford Hill criteria in the 21st century: how data integration has changed causal inference in molecular epidemiology|url=https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4589117/|journal=Emerging Themes in Epidemiology|volume=12|doi=10.1186/s12982-015-0037-4|issn=1742-7622|pmc=PMC4589117|pmid=26425136}}</ref> contend that an entire body of evidence is needed before determining if an association is truly causal.<ref>{{Cite web|url=http://sphweb.bumc.bu.edu/otlt/MPH-Modules/EP/EP713_Causality/EP713_Causality_print.html|title=Causal Inference|website=sphweb.bumc.bu.edu|access-date=2018-04-01}}</ref> Moreover, many research questions are impossible to study in experimental settings, due to concerns around ethics and study validity. For example, the link between cigarette smoke and lung cancer was uncovered largely through observational research; however research ethics would certainly prohibit conducting a randomized trial of cigarette smoking once it had already been identified as a potential health threat.
==See also==
| |||