Monday, January 07, 2008

Lies, damn lies, and younger siblings

I have spoken about the danger of linking the birth of a younger sibling to the onset of stuttering before. I feel obliged to write about this again because I have read it in Barry Guitar's newly update book which is supposed to be a bible among speech therapists.

I have been looking at the empirical numbers. Unfortunately, I could only find the birth spacing of women in developing countries. The younger sibling is born

<2 years after birth of last child with a probability of 25%
2-3 years later with a probability of 25%
3-4 years later with a probability of 25%
>4 years with a probability of 25%.

Lets assume the older sibling starts to stutter at age 3, then in 50% of the cases his younger sibling is born within a year, and in 25% (50% divide by 2) within half a year.

So in 25% of the cases, parents can perceive a strong correlation between onset of stuttering and birth of younger sibling. And it's a complete illusion!!!! It is a beautiful example of the common logical fallacy of confusing correlation with causation.

Note: I am not saying that no relationship exists to exact time of onset (i.e. when it starts now or a few weeks later) or strength of onset (i.e. whether the frustration or general well-being level is higher resp lower). But the coincidence explains very well any observation... and if there is a relationship it is very weak.

No comments: