The Skeptic’s Repertoire
The skeptic’s repertoire – correlation vs causation

The skeptic’s repertoire – correlation vs causation

Summary:

  • There are numerous logical and rhetorical fallacies which can influence public debates or private opinions.
  • Logical fallacies are mistakes in reasoning which make arguments and opinions false.
  • Both correlation and causation are used in research to describe the relationship between variables but can be mistaken for one another.
  • Correlation is a connection between two things while causation is a cause-and-effect relation.

This series of articles discusses common logical fallacies and its goal is to inform the skeptical reader as to how to spot and avoid them. This article discusses the common fallacy of mistaking correlation for causation. In other words, assuming that one event is the result of another even when such a conclusion is unjustified.

Correlation is a connection or relationship between two or more things that are not caused by chance [1]. When one factor changes, the other tends to change as well. In a positive correlation, two things are likely to exist together  (more studying can lead to better test results), while in a negative correlation, they do not (higher stress levels can decrease productivity). Causation, on the other hand, is a cause-and-effect relationship. One thing directly leads to another; an event causes a consequence.

The fallacy of false cause, also known in its Latin form – non causa pro causa – identifies the cause of one phenomenon in another where the relation is of a different nature [2]. For example, a fallacy-prone reader could come across the article “Gay Men Earn Degrees at Highest Rate, Study Finds” and easily form a false conclusion that education makes people gay [3]. However, this is correlation only and other factors are at work here. There is a hypothesis, for example, that “gay men respond to homophobia by overcompensating in achievement-related domains like education” [3].

A common version of this fallacy is called post hoc ergo propter hoc (“after it, so it was caused by it”) and it mistakes a sequence of events for causation. Carl Sagan, a skeptical popularizer of science whose work was discussed in one of our previous articles, provides an amusing example in his guide to spotting nonsense called baloney detection kit: “Before women got the vote, there were no nuclear weapons” [4]. Such a statement may imply causation to a reader unfamiliar with fallacies but fails even to explain how the two are related.

Another fallacy seeing causation where it doesn’t exist is called cum hoc, ergo propter hoc (“with it, therefore because of it”). This time, the fallacy implies causation between events occurring at the same time; for example statistics showing a positive correlation between gun ownership and violent crime could lead to the conclusion that the higher number of guns owned, the higher the rate of violent crime [5]. However, “the causal relationship may be the exact reverse” [5]. It may also be true that people buy more guns because they fear violent crimes. In fact, gun violence may be influenced by the right to bear arms, gang violence, income inequality, universal background checks, and many other factors.

When discussing correlation, it is prudent to also mention co-occurrence, also referred to as concurrence. It is “a situation in which two or more things happen at the same time” [6]. In contrast to correlation and causation, this word typically doesn’t imply a relation, suggesting mere coincidence instead. The websites tylervigen.com [6] and reddit.com//r/SpuriousCorrelations/ [7] provide many amusing examples of concurrence, including graphs of similar trends in divorce rates in Maine and per capita margarine consumption.

Unfortunately, false cause fallacy is rather prevalent in our culture and, by hiding in plain sight, not so easy to spot. It doesn’t surprise audiences when, in almost any football match, they see footballers – who spend tremendous amounts of time training – score goals and immediately attribute the athletic success to their gods. Similarly, a black cat crossing the street might be blamed for causing someone an unpleasant day even though there is no evidence for a causal relationship. In fact, determining proper causation is a difficult scientific process which we explained in one of our previous articles. Still, a skeptical reader can avoid fallacies by always asking questions, demanding evidence, and remembering that quick assumptions should be challenged.

References:

  1. Correlation. Retrieved on 24.06.2023 from https://www.macmillandictionary.com/dictionary/british/correlation
  2. Fallacy. Retrieved on 24.06.2023 from https://www.britannica.com/topic/fallacy#ref1102389
  3. Carrasco, Maria. 2023 Inside Higher Ed. Gay Men Earn Degrees at Highest Rate, Study Finds. Retrieved on 24.06.2023 from https://www.insidehighered.com/news/2021/11/30/gay-men-earn-degrees-highest-rate-us
  4. Sagan, Carl (1996). The Fine Art of Baloney Detection. Page 212. The Demon-Haunted World: Science As a Candle in the Dark (Paperback ed.). Ballantine Books. ISBN 978-0-345-40946-1.
  5. Cum Hoc, Ergo Propter Hoc. Retrieved on 24.06.2023 from https://www.fallacyfiles.org/cumhocfa.html
  6. Retrieved on 24.06.2023 from https://tylervigen.com/spurious-correlations
  7. Retrieved on 24.06.2023 from https://www.reddit.com/r/SpuriousCorrelations/