Ice creams cause drownings. And there are numbers to prove it. If you take the number of ice creams sold and compare it to the number of drownings there is clearly an upward trend. And it is therefore safe to conclude from this that we should ban ice cream because ice cream is very dangerous. But there is something wrong with this logic. The underlying factor is nice weather. If the weather is nice, more people will go swimming and more people will tragically drown. At the same time, more people will buy ice cream. But it’s not the ice cream that is causing the drowning.
One of the most common mistakes in research and evaluation is jumping to an incorrect conclusion about causality when you see a correlation. Here are some other examples:
Married men live longer than single men
Statistics show that this is true, including in data from Harvard Health Publications. But it’s quite possible that the causal relationship is the other way around. Men who are healthy, rich and well educated and with a higher life expectancy may be more likely to find a wife (perhaps?). And the men with a low life expectancy may be not as likely to get married. So it’s the high life expectancy that is causing the marriage, not the other way around.
Young children who sleep with the lights on are more likely to become short sighted
A study published in the May 13, 1999 edition of Nature found that babies younger than two years old who slept with the light on were at increased risk of developing myopia later in childhood. There was a clear correlation identified, and advice was given to parents to turn the lights off at night. But a later study published in the same journal corrected the findings, and realised that genetics were at play here. Shortsightedness is genetic. Parents who are short sighted are those who are likely to the leave the light on in the bedroom. And they are the ones who are most likely to have short sighted children.
A child with higher self esteem will do better at school
In the 1970s researchers found a link between children who did well at school and those who have high self esteem. This research was presented to parents in a way that suggested to them that if they nurtured high self esteem in their children, good results would follow. Then someone looked at the causal relationship and found that the causal link was happening in the other direction: the good results at school were causing the high self esteem. Children that were raised to have high self esteem yet were not doing particularly well at anything ended up having low self esteem.
There are are number of other incorrect causal relationships, that include the serious misconception about the link between vaccines and autism. It is important to remember that even if there is a correlation between two variables, it is not necessarily enough to provide evidence of a causal relationship. It may provide a very good hint about what may be happening, but before you can conclude something causes something else, you need to understand how it does and why it does.
So when in doubt, just remember the ice cream.
This is a slightly amended transcript from a TEDx talk given by Ionica Smeets, Professor of Science Communication at Leiden University, the Netherlands. Ionica is a mathematician and science journalist with a passion for bring science closer to the general public. You can view her 5-minute TEDx talk here.