I just saw an article on the BBC website (here), discussing some research showing that left-handed people tend to die nine years earlier than right-handed people. This seems an extraordinary finding. Can it really be true that tools such as knives and scissors are mostly designed for right-handed people, and they are dangerous for left-handed people to use, with the result that left-handers injure themselves more often and die off earlier? Surely not!
In fact, as pointed out in the BBC article, the research is completely flawed. But it is interesting to consider why, and also to see if you can spot the flaws.
In the original report, the researchers traced the families of 2000 people who had recently died and asked whether they had been left-handed or right-handed. And they discovered that the average age of the left-handers when they died was nine years less than the right-handers. It seems straightforward, doesn't it? Can you see the flaw? I admit that I couldn't, and I had to read the explanation.
In fact, until quite recently, it was the practice to encourage left-handers to conform to the norm, so there was substantial pressure in the home and at school for them to change. This means that many of the older people who were reported by their family to be right-handed would actually have been left-handed if there had not been that social pressure. In contrast, the same pressures would not have existed for the younger people to conform, so the proportion of left-handed and right-handed younger people who died would have been more natural.
What is stunning is that this research was published in top medical journals and nobody spotted the flaw. It seems that we are sometimes not very good at quite elementary logical and statistical analysis.