9. Post hoc ergo propter hoc

Definition

Latin for "after this therefore because of this.” This is when we infer A caused B simply because B happened after A.

Examples

1) I ate chili fries in March and had stomach cancer in July. Therefore, the chili fries caused my stomach cancer (Thanks for this example, Steve Hansen).

2) Caveman Bob beat the wall of the cave and the sun reappeared. Therefore, beating the wall of the cave caused the sun to reappear.

3) My wife must be driving the car incorrectly because we never had transmission problems until she drove the car.

4) We raised the tax rate and crime went down, so the tax rate caused crime to go down.

5) We prayed for a Mercedes and then got one. Prayer works!

6) Since I began eating my boogers last month, I haven't been sick. Therefore, eating boogers is like getting a flu shot.

7) Henry received a vaccine and became ill. Therefore, the vaccine caused his illness.

Discussion

It is important to understand that there really might be a causal connection between these events, but we cannot infer there is one merely because one event chronologically preceded another.

The post hoc fallacy is very similar to the axiom that correlation does not imply causation, which is the idea that related variables are not necessarily causally related in a direct way. The crime rate may increase as people watch more violent television, but we cannot infer violent television caused the increase in crime from the correlation alone. Further research is needed.

How to avoid

Create a scientific experiment to test for a causal connection. You can also use the inductive techniques created by philosopher John Stuart Mill to test for a causal connection.

Most importantly, do not assume causation simply because one event precedes another or simply because one event is correlated with another.

Exercise

1. List some superstitions that seem to be based on the post hoc fallacy.
2. Explain what might be the problem with the following reasoning of some scientists in 2009: “HDL Cholesterol is negatively correlated with the incidence of heart attacks. Therefore, you should take HDL-raising medications to decrease your risk of heart attack.”

Answers

1. Break a mirror and have seven years bad luck. It is bad luck to walk under a ladder. Also, getting x after praying for x may be a post hoc fallacy leading to a superstition about the effectiveness of prayer.

2. This example is from Wikipedia’s article, Correlation does not imply Causation (2012). The problem is other underlying factors like genes, exercise, and diet may be causing both HDL levels and the likelihood of having a heart attack. That is, an HDL medication may simply treat the symptom, not the cause. This is one negative consequence of believing correlation implies causation.