Cause and Effect

There are many limits to what we can demonstrate in research, and one of the hardest is causation.  To demonstrate that A causes B, we need to ensure 3 things:

1. That A and B are related

2. That A occurs before B

3. That no other variable explains their relationship.

I beat this drum so much that I’m pretty sure my students are good and sick of hearing it.  But I am combatting this misconception nearly every day from popular media.  While on a trip to the UK, I found this mini article in their newspaper.

IMG_1264

 

Ok, so what I’m guessing is that they gathered these people together, asked about their history of head injury and tested their cognitive abilities.  So they may have met the first criteria.  But that’s it.

It is unlikely they tested these people BEFORE their head injuries because it would be difficult to predict who would suffer such an injury.

The problem with this conclusion is that it could be that B causes A – perhaps poor cognitive functioning increased the likelihood that these people would injure themselves.  Alternatively, it could be that some other variable accounts for both A and B. Perhaps some neurological difference makes the injured people more impulsive, and also impacts on their attention and concentration thereby affecting their cognitive performance.

Now, I know it makes a kind of logical sense that hurting your head, and therefore possibly your brain, would impact on what your brain does, but this research CANNOT actually draw this conclusion.