The following excerpt about the correlation-causation fallacy is from Professor Douglas Walton’s Informal Logic: A Pragmatic Approach.
A study published in the journal Nature on May 13, 1999, found that babies who slept with a night-light on had an increased chance of developing myopia (nearsightedness) later in life. A subsequent study, co-authored by Professor Karla Zadnik of the College of Optometry at Ohio State University, found other factors that explained the correlation between babies sleeping with a night-light on and later myopia. These other factors are cited in example 8.14.
The previous study, conducted by researchers at the University of Pennsylvania, found that “ambient light exposure during sleep at night in the ﬁrst two years” of a child’s life greatly increase that child’s chances of developing myopia. This earlier study showed that nearly half of the children who had slept in a fully lit room had become myopic later in childhood. But the same study did not take into consideration whether or not parents were near- sighted, according to Zadnik. Her study took into account parental myopia. The researchers noticed that nearsighted parents were more likely to use a night-light in their child’s room. “We think this may be due to the parents’ own poor eyesight,” Zadnik said. Also Zadnik said her study found that genetics plays a signiﬁcant role in causing myopia.
In this case, the earlier study showed an impressive correlation between children who had slept in a fully lit room and those who became myopic later in life. As the later study showed, other factors were at work. Parents who were more likely to use a night-light were also more likely to be myopic, and parents who are myopic are more likely, for genetic reason, to have children that also tend to be myopic.
To deal with this kind of case, we need to recognize that the initial correlation between night-lights and myopia sets up a causal hypothesis between these two factors that needs to be explored further by asking critical questions. Could there be other factors, for example, linked to both myopia and the use of night-lights with babies? Further studies needed to be made, and in this case were made, to answer this critical question.
What cases like this one bring out is that a dialogue or inquiry can proceed by asking questions, or suggesting possible causal relationships at the opening stages of the inquiry. Later on, as the inquiry proceeds, and more evidence comes in, these relationships may be ﬁrmed up, repudiated, questioned further, or even made more complex by the discovery of other factors that had not previously been identiﬁed. To leap ahead, and ignore or pre-empt the natural and reasonable order of dialogue by drawing a causal conclusion too ﬁrmly or too quickly could be a post hoc error.
On the other hand, it may be natural and reasonable, at the early stages of the investigation, to advance causal connections as hypotheses that may later on turn out to be refuted. This in itself is not inherently fallacious or incorrect, provided the hypothesis is corrected once further evidence contradicts it.
Informal Logic, 2nd Edition, by Douglas Walton
Copyright © 2008 Douglas Walton. Reprinted with the permission of Cambridge University Press.