Many people, including government officials and members of the judiciary, commonly confuse correlation with causation, although they are critically different. The ramifications can be drastic, especially in legal matters, because culpability in civil or criminal litigation can turn on a demonstration of causation.
Correlation, which may be the result of serendipity rather than cause-and-effect, should not be dispositive evidence that could deprive a defendant of funds or freedom. Examples are discussed below.
Correlation means only that there is a relationship or pattern in time or space between the values of two variables. Causation means that one event or activity actually causes another event to occur. Causation can only be definitively determined by means of an appropriately designed experiment.
The plot below illustrates how an erroneous conclusion could be reached. It shows a correlation in time between organic food sales and the increase in autism, but nobody — nobody who is rational, at least — would think that one causes the other. We would need to do a careful observational study — or, preferably, a randomized clinical trial — in order to conclude that there was causation.
How do people come up with correlations that turn out to be merely coincidental but do not involve causation? One way that often ends up reported in the scientific literature and media is via food frequency questionnaires, or FFQs. They ask people a large number of questions about things they do and eat, and also about their symptoms, illnesses and life events, and then they look for correlations. And, lo and behold, if you ask enough questions, statistical anomalies arise — you get statistical false positive correlations similar to the figure above.
For example, a 2008 article in the Proceedings of the Royal Society used a FFQ with 131 questions, which were asked at two different time points, giving a total of 262 questions. The British academic researchers reported an association between women eating breakfast cereal and increased odds of having a boy baby.[1] Of course, the sex of a zygote is actually determined by whether the male's sperm contributes an X or a Y chromosome.
Similarly, a 2017 article in the journal Heart by an international group of researchers, based on a FFQ with 192 food questions, found a decrease in atrial fibrillation, an abnormal heart rhythm, associated with chocolate consumption.[2] These kinds of studies fill the medical literature with findings that are likely to be spurious.
The distinction between correlation and causation is also often critically important in the law, because a demonstration of causation is necessary to hold a person responsible — legally or morally — for an injury or a harm.
The federal courts provided a reminder of the importance of demonstrating causation in a case called Daniels-Feasel v. Forest Pharmaceuticals Inc. The plaintiff alleged that Lexapro, a prescription antidepressant, caused her child to develop autism as a result of her having taken the medication during pregnancy.[3] The plaintiff attempted to prove causation through the presentation of expert evidence. Her attorneys presented an alleged expert who relied upon a "weight of the evidence" analysis to assert that the medication caused autism.
But the U.S. District Court for the Southern District of New York determined that the alleged expert had not adopted a methodology that used a scientific method or explained how "weighing the evidence" was other than expressing a largely arbitrary opinion.
Therefore, the plaintiff was unable to prove general causation — that is, that prenatal exposure to Lexapro can cause autism. The lower court granted summary judgment for the pharmaceutical company in 2021, and the U.S. Court of Appeals for the Second Circuit concurred in July 2023.
Courts have an obligation as gatekeepers of the admission of evidence in a trial to exclude speculative or idiosyncratic conclusions that claim to be scientific evidence that support causation. In the Lexapro case, both the district court and the Second Circuit acted by prohibiting the introduction of subjective opinion that purported to be scientific evidence of causation.
Unfortunately, not all courts have exercised their gatekeeper role as judiciously as the courts did in the Daniels-Feasel case. Medical products company Johnson & Johnson is facing tens of thousands of lawsuits alleging that the company's talc products, including its once-ubiquitous baby powder, caused cancers — including ovarian cancer and mesothelioma — in women who used it.
Johnson & Johnson is attempting to resolve those lawsuits, as well as any future talc lawsuits, but In May 2024, a bankruptcy judge rejected the company's proposed $8 billion settlement to resolve all talc-related ovarian cancer lawsuits. Tens of thousands of suits remain outstanding.
The pivotal question is, of course, does talc actually cause cancer? The evidence for that is, at best, dubious.
According to a systematic review of talc and cancer published in Frontiers in Toxicology:
Although the paucity of mechanistic information remains a limitation, the balance of
evidence, especially the negative animal studies and body of higher-quality cohort studies demonstrating no increased cancer risks, provides the best scientific evidence of lack of carcinogenicity, at least at "real world" human-relevant exposures.[5]
A caveat, however: Three of the 11 authors of the review "have been retained as expert witnesses on behalf of defendants in litigation matters in which it has been alleged that products containing talc caused mesothelioma or other cancers."
Another unfolding example with potentially important consequences is the avalanche of lawsuits — with many thousands of plaintiffs — alleging that the ingestion of acetaminophen during pregnancy has caused attention deficit hyperactivity disorder or autism in offspring.[6] The outcome remains to be seen.
The gatekeeper role of the courts in such civil litigation alleging torts is critical. Aside from the requirement to demonstrate causation in order to hold a person legally responsible for an injury or harm, there are several other reasons:
- To protect the scientific method and the integrity of science;
- To prevent the removal of safe and efficacious products from the marketplace due to
litigation costs and unjustly imposed economic liability, when causation does not exist; - To promote scientific literacy and to avoid propagandized scares and fears among the public; and
- To shield the legal system from ideologically driven lawsuits where the object is not justice and fairness, but anti-science campaigns of economic intimidation.
Judges as gatekeepers are especially needed — and must be especially careful — when the lawsuit, often a class action, is ideologically motivated. Judges should be alert to speculative causation claims in such class actions, and be willing to dismiss the suit, as in the Daniels-Feasel litigation.
Fundamentally, the imperative for getting determinations of causation right is simple: It ensures that, as the wheels of justice turn, reputations, lives and livelihoods are protected. On one hand, individuals should be appropriately compensated for damages to them. On the other, businesses that are blameless — i.e., not the cause of the injury or harm — should be protected from both the risk of, and the actual imposition of, unjust and unfair damages that could threaten their very existence.
By helping to distinguish correlation from causation, science can help to adjudicate that balance.
Drew L. Kershen is a professor emeritus of agricultural law at the University of Oklahoma College of Law. Henry I. Miller, a physician and molecular biologist, is the Glenn Swogger Distinguished Fellow at the American Council on Science and Health. He was the founding director of the FDA's Office of Biotechnology.
[1] https://royalsocietypublishing.org/doi/abs/10.1098/rspb.2008.0105.
[2] https://heart.bmj.com/content/103/15/1163.
[3] Daniels-Feasel v. Forest Pharmaceuticals Inc.
[4] https://www.lawsuit-information-center.com/2-billion-verdict-in-missouri-motivates-jj-to-settle-talcum-powder-lawsuits.html
[5] https://www.frontiersin.org/articles/10.3389/ftox.2023.1157761/full.
[6] https://www.lawsuit-information-center.com/tylenol-autism-lawsuit.html.
Henry I. Miller, MS, MD
Henry I. Miller, MS, MD, is the Glenn Swogger Distinguished Fellow at the American Council on Science and Health. His research focuses on public policy toward science, technology, and medicine, encompassing a number of areas, including pharmaceutical development, genetic engineering, models for regulatory reform, precision medicine, and the emergence of new viral diseases. Dr. Miller served for fifteen years at the US Food and Drug Administration (FDA) in a number of posts, including as the founding director of the Office of Biotechnology.
