Oxford University Press's
Academic Insights for the Thinking World

Complexities of causation

Imagine the thrill of discovering a new species of frog in a remote part of the Amazon. Scientists are motivated by the opportunity to make new discoveries like this, but also by a desire to understand how things work. It’s one thing to describe the communities of microorganisms in our guts, but quite another to learn what causes these communities to change and how these changes influence health. A big part of science, with important practical implications, is understanding how things work, i.e., causation.

Sometimes it’s simple to understand causation. Do you read murder mysteries? One part of solving the crime is figuring out how the victim died; if there’s a smoking gun, the detective can be pretty sure of the cause of death. Scientists occasionally have definitive evidence to test an hypothesis about a cause-effect relationship. Often, however, they don’t have such a smoking gun, and have to rely instead on multiple lines of individually inconclusive evidence.

Several factors contribute to complexity in causation. Behaviors of animals are influenced by internal physiological processes and external environmental factors as well as the evolutionary history of the lineage to which the animals belong. Causes may be necessary but not sufficient, or sufficient but not necessary, or “Insufficient but Necessary parts of a condition which is itself Unnecessary but Sufficient” (INUS conditions, Mackie 1965). Doing science is challenging because of these complexities in understanding causation, but there is a more fundamental source of complexity that is challenging for everyone trying to understand results of research.

Kintigh Generating Station, a coal-fired power plant in Somerset, New York. Image Credit: Matthew D. Wilson, CC BY-SA 2.5 via Wikimedia Commons.
Kintigh Generating Station, a coal-fired power plant in Somerset, New York. Burning coal is one example of an anthropogenic cause of global warming. Photo by Matthew D. Wilson, CC BY-SA 2.5 via Wikimedia Commons.

Humans often fail to recognize that events can have multiple interacting causes and instead attribute events to single, specific causes. Two examples in the news lately illustrate these features of human thinking, one involving climate change and another involving vaccination against common diseases. Many factors influence our changing climate: volcanic activity, changes in solar radiation, and other natural factors as well as human activities that release greenhouse gases into the atmosphere. Climatologists account for both natural and anthropogenic factors in modeling global temperature change since the beginning of the Industrial Revolution; only models that include both types of factors are consistent with the observed pattern of temperature change since 1860. Yet climate contrarians seem to believe that the cause of climate change must be either natural factors or human influences, not both. Since we know that volcanic activity and changes in solar radiation influence climate, they argue, we must ignore human influences. In fact, effects of both natural and anthropogenic factors on global average temperature are obvious before 1950, but virtually all change in temperature since 1950 can be attributed to human activity. If anything, recent history supports the opposite of the claim of climate contrarians.

Large-scale vaccination programs have greatly reduced the incidence of diseases such as measles and whooping cough in developed countries. Some individuals in any population can’t be vaccinated because they are too young, too old, or have compromised immune systems, so the success of vaccination programs depends on vaccinating a large enough percentage of the whole population that the disease won’t spread to those who can’t tolerate the vaccine. This is herd immunity, and the necessary fraction that must be vaccinated depends on the basic reproduction rate of the organism causing the disease. Since measles and whooping cough are easily spread, they have high reproduction rates, and more than 90% of people must be vaccinated to prevent epidemics. There have been several outbreaks of whooping cough in the United States in recent years, coincident with two factors: (1) a decline in vaccination due partly to an active and energetic anti-vaccination movement and (2) replacement of the traditional vaccine using killed pertussis cells (DTP, which includes vaccines for diphtheria and tetanus as well as whooping cough, also known as pertussis) with an acellular preparation that has fewer side effects but isn’t as effective as the traditional vaccine (DTaP). Members of the anti-vaccination movement reject the importance of herd immunity and argue that the sole cause of recent outbreaks of whooping cough is reduced effectiveness of the current acellular vaccine. In fact, this cause is synergistic with the recent decline in vaccination. Herd immunity is even more important now that physicians use a less effective vaccine against whooping cough. Since some people will have been vaccinated with DTaP long enough ago that their protection against whooping cough has lapsed, an even larger percentage of the population needs to be vaccinated to prevent spread of the disease.

You may think that you are too sophisticated to assume that events have single, specific causes as illustrated by these simple examples. If so, I urge you to read T. C. Chamberlin’s classic 1890 paper “The method of multiple working hypotheses.” Chamberlin summarizes the risk of narrow-mindedness for practicing scientists thinking about causation as follows: “The moment one has offered an original explanation for a phenomenon which seems satisfactory, that moment affection for his intellectual child springs into existence; and . . . his intellectual offspring . . . grows more and more dear to him.”

Feature Image: Eruption of Kanaga Volcano on Adak Island, Alaska. Volcanic eruptions are one example of a non-anthropogenic cause of global warming. Photo by  E. Klett, U.S. Fish and Wildlife Service. Public Domain via Wikimedia Commons.

Recent Comments

  1. Stephen Jenkins

    This extension of my posting on causation was inspired by “Vaccines: Calling the shots”, a PBS NOVA documentary that aired on August 26, 2015. Causes must precede their effects, but this doesn’t mean that X caused Y just because X happened before Y. Yet some parents won’t vaccinate their children against measles because of reports of initial symptoms of autism following vaccination. A British physician named Andrew Wakefield fed this fear of a connection between vaccination and the development of autism with a small study published in 1998 that was later shown to be fraudulent, and this fear continues to be stoked by celebrities such as Jenny McCarthy and Robert Kennedy, Jr.

    There certainly are cases in which a child develops autism following vaccination against measles. These cases might provide evidence of a causal relationship, but they might also be coincidental. How can we evaluate these possibilities? We need to think about the fact that a very large number of children are vaccinated against measles at about the age when signs of autism first appear. This being the case, it shouldn’t be surprising that some kids show these signs after vaccination even if there is no connection between vaccination and autism. Individual reports by parents of a child who develops autism shortly after vaccination are only weak evidence of a causal relationship because the connection between vaccination and autism could easily be coincidental, especially because the vast majority of vaccinated children don’t develop autism. Anjali Jain and five other researchers recently examined the prevalence of autism in 95,727 children. Some of these children had older siblings with autism spectrum disorder (ASD); these children with autistic siblings were less likely to receive MMR (measles, mumps, rubella) vaccine, but more likely to develop autism than children with unaffected siblings. In general, there was no difference in likelihood of developing autism between children who were vaccinated and not vaccinated, regardless of whether older siblings had autism or not.

    As of September 2, 2015, the NOVA documentary on vaccination is still available for viewing online. It includes an example of an Australian child who started having seizures shortly after vaccination at age 6 months. Luke’s parents thought the vaccination caused the seizures, but eventually discovered a research team in Melbourne that studies cases like these. The researchers found that Luke had a form of epilepsy caused by a rare genetic mutation. The onset of Luke’s epilepsy may have been triggered by vaccination, but if it hadn’t been, it likely would have been triggered by some other stress before long.

Comments are closed.