The Post Hoc Logical Fallacy and the Importance of Control Groups
July 22, 2010
Going to be doing a couple more logical fallacy posts.
One of the more common logical fallacies is the Post hoc ergo propter hoc, or “after this, therefore because of this” fallacy. It’s not difficult to understand this one; You do something, something else happens, therefore a causal relationship is assumed. I drank orange juice and got over my cold, therefore orange juice fights a cold. I shaved, and it came back darker and thicker, therefore shaving darkens and thickens your hair.
This falls nicely in line with the human tendency to search out patterns. Pattern recognition is a large evolutionary advantage; the creature that recognizes patterns, whether it be the look of a poisonous or non-poisonous plant, or recognize the roar of a lion and connect it to the lion being near, is more likely to survive than the creature unable to do so. This accounts for such phenomena as paredolia, as well as the post hoc logical fallacy.
However, just because a pattern is seen doesn’t mean it is always there. Consider the obvious non-causal example of crossing your fingers. It is common practice to cross one’s fingers when hoping for something, however, very few people actually believe crossing your fingers actually improves the likelihood of the desired outcome. However, if the desired outcome is achieved after crossing one’s fingers, it could be erroneously concluded that it did have an effect.
And speaking of errors, those two examples I mentioned at the beginning. Neither of those are actually true. Both common ideas often “confirmed” by post hoc rationale. Studies have shown no statistical link between vitamin C and reduction in severity or length of a cold. The illusion of darker, thicker hair can be attirbuted both to cutting off the hair’s natural taper, thus only showing the thick base, and the fact that people generally begin shaving around the time hair is starting to grow in, often around puberty. As time passes, the hair grows thicker and darker, regardless of whether or not it is shaved off. However, since it was lighter and thinner when most people start, they attribute the shaving to the thickening in darkening, when in fact, there is no link between shaving and hair growth.
Now how do we avoid this error in the future? Control groups. Whenever an experiment is performed to determine the effectiveness of something, there should always be an experimental group that does not receive the treatment or procedure. This group is the control group. A common example is the use of placebos in medical studies. One group is administered the actual experimental treatment, while the other is administered a placebo, an identical looking treatment that does nothing. This could be in the form of a sugar pill as opposed to the actual drug, a saline injection as opposed to an active injection, or any other treatment that appears the same as the actual. This allows the experimenter to compare the results from both the experimental group, group receiving the active treatment and the control group, the group receiving the placebo. If the experimental group shows significant improvement over the placebo group, than there is a definite effect, while if they stay the same, there is no actual effect.
