Pseudoscience

In the last couple of months, a number of articles have been published in both electronic and print media, regarding the  statictics of cyclists' death in London last year.

No, they were not blaming the London traffic, the Bendy buses or the reckless drivers; Instead they were pointing out to the fact that out of 13 cyclists died in road accidents last year, 10 were women. The articles expressed their surprise to this "disproportionately high" percentage of women cyclists involved in the fatal accidents and some also tried to find out the possible reasons behind it.

Comments poured in from the readers with hundereds of different explanations. While some took no time to dismiss women and their ability to ride cycle and promptly extended their verdict to general uselessness of women, others were very patronising; poor women,the weak and less clever things; it is not fair to compare them with the strong, bold, clever, agile men. This is only to be expected.

One clever comment even went on to explain it from the genetic evolution (his own version, of course); As men had to hunt (and women apparently stayed peacefully in comfy caves, looking after the children, preparing meals and watching cave-soaps in prehistoric telly, perhaps), they are genetically more evolved to scan the surrounding better than women. And this fantastic skill helps them to watch out for the big lorries while cycling in narrow London roads; but women get easily perished under those big wheels due to their lack of "surroundings scanning gene".

 What a beautiful piece of armchair science without any need for proper understanding of the subject or the experimental data! If born in ancient Greece, this man could give stiff competition to Aristotle, who proclaimed women have fewer teeth than men, without bothering to count.

It is truly amazing how so many people just overlooked the sample size. The first thing that should strike anybody is the fact that you can not have a statistics with just 13 data points.The accuracy of any statistical probability depends largely on the sample size; the more data points you have the more accurate is your result.The error in statistical probability is approximately √n, where n is the sample size.

No one would have objected, if 50% of the accidents were involving women, which is 6.5 in this case. But for a sample size of 13, the error is about 3.61. Add this two together (6.5+3.6) and you get 10.1, the number factoring in the error bar. Still a tad more than the actual number of women died. So much for the "disproportionately high" number of women involved in fatal crash.

This is the precise reason why the actual statisticians do not operate with a sample size of 13 while dealing with a population size of London. The error bar here is comparable to the sample size and no one can infer anything from this; no one other than an ignorant, sexist, bigot, hell bent on believing in the male superiority.

The kind of coverage this story got in major mainstream news papers (Guardian, BBC, to name a few) only shows the deep rooted gender bias in the society; You just need to scratch the surface and the eternal male ego will be inflated eclipsing all reasons and facts.

What makes this even more appalling is the use of pseudoscience to make this chauvinistic claim look convincing.Some articles went as far as proclaiming that women cyclists in London are at "considerably greater risk". All based on their sample size of 13 and yet they have not been charged for rumor spreading and scare mongering.

I tossed a coin 3 times  the other day and I got tail in each occasion. Do I see another sensational heading?
 "Always get tail in a toss in Stockholm". Which might open a floodgate of pseudoscientific speculations... proximity to north pole, magnetic effect of earth, effects of aurora borealis & solar wind,long hours of day light, metallic composition of Swedish Krona.....the possibilities are endless.