A causal relationship between two events is where the first event causes the next event to occur. A correlative relationship is not a relationship at all – it is two events coinciding in a situation that can be mistaken for causal. People frequently mistake a noted correlation of two events as a causative relationship between the two events.
Relationships between two events come in four flavours:
* No relationship at all = randomness / coincidence
– For example, my eyes are hazel and the stranger I just passed is eating a sandwich
* An apparent relationship, with no causation = correlation
– For example, a survey of contents from stomach pumps at the local justice centre found the presence of carrots in 100% of inmates.
* A complex causal relationship, causation is established, but the exact mechanism is not = contributing factor
– Levels of bowel cancer and the presence or absence of roughage in diet
– Increased levels of CO2 in the atmosphere and the increased temperature of the world
* A standard causal relationship, causation is not only established but the mechanism is known and understood = causal relationship
– Placing a transparent vessel containing both hydrogen and oxygen in ultraviolet light forms water and heat… explosively
– If I lose 6 litres of blood from my body, I will die
We humans like to see patterns in things. Patterns are the first part of predicting future events, which can allow us to change our behaviour now to alter the expected outcome. An accurate pattern is a useful tool. However if the pattern is false, changing our behaviour now will not have the desired outcome, so we should recognise this error and let the pattern go.
The temptation is to mistake a perceived pattern for a tool that can accurately predict the future with a deficit of evidence, or to hold onto such a mistaken tool in the face of suitable contradictory evidence.
In the example I gave regarding the presence of carrots in people incarcerated in correctional facilities, it would be a mistake to consider that carrots lead to criminal activity, or that criminal activity leads to the presence of carrots in the stomach. Another example would be to look at the statistics of belief that criminals have and criminal behaviour. In the USA, the vast majority of criminals state that they are Christian while almost no inmates state that they are Athiest. One could conclude from these two bits of data that Christianity is a contributing factor in criminal behaviour, or that being Athiest minimises the likelihood of criminal behaviour. This conclusion is clearly in error. Criminal behaviour causes criminal behaviour, your belief system, or lack of it, is irrelevant.
Another common error is to attempt to make a mechanism for a correlation to support causation. For example, you may correlate your behaviour with an observation of the fullness of the moon. You may feel odd or bizarre and noticed that this happened last time the moon was full, just like it is this time. This seems to be a pattern, correlating the two events together. To explain this pattern you create a causative link between the moon and your behaviour. To explain that causation requires a mechanism, so you suggest the ionisation of the atmosphere increases due to the reflected light of the moon, or the increase in gravity during the full moon and the new moon affects the water in your body, or some other mechanism. Perhaps one of these is correct, but probably not.
The error is *not* in speculating about the mechanism, but in ascertaining confidence in the speculation without testing, and in the correlation in the first place. In this case the ionisation of the atmosphere due to reflected light is negligible in the face of other factors such as solar storms. If you were to react to ionisation shifts, then the change in levels due to sun light would create massive mood swings in comparison to any shift from the reflection of the light from the moon. As for gravity, you get a stronger affect from being near a mountain than from the combination of the sun and moon, so your mood would shift every time you got near a mountain – yet it doesn’t. Also, step back a bit – is there correlation of your mood and the moon real? Take down a chart of your mood for a year. Then compare it to the calendar and see if there is a definite correlation with the phase of the moon. The odds are against such a correlation existing. If it doesn’t, then your observed pattern is false. If it does, then perhaps there is a pattern to test.
Of course another consideration is that perhaps your mood causes the full moon… yet your mood changes frequently while the timing of the full moon does not. And let us not get stuck in the various definitions of what constitutes a “full moon”.
The Logical Fallacy of mistaking correlation with causation is using a correlation or random coincidence to substantiate a conclusion in the mistaken belief that there is a causal relationship between the first event and the second.