Monday, June 23, 2008

Effect and cause

I got to read the Sunday Express [sigh] front page headlines "Suicides 'Linked to mobile phone masts'". Apparently some Doctor has discovered that of the 22 suicides he looked at they lived closer than average to a mobile phone mast. Thus at an instant we can know that mobile phone masts cause suicide or at least severe depression - ta da.

Except of course that might be complete nonsense. Let me offer what may be a more plausible solution to the statistics.

1) Currently people don't want mobile phone masts put up next to their houses.
2) So those who live next to them either had them there for ages or couldn't fight the planners.
3a) From point 1 the price of houses next to masts is likely to fall
3b) From point 1 those who can afford to move away from the mast will do so

Therefore it could be said that those who live closer to a mobile mast are either 'poor' or may have been forced into negative equity.

I think that either of those two may be enough to cause or exasperate depression. Ironically in this hypothesis it's the media's own trumpeting of the 'health risks' caused by mobile masts that would have caused this; in that instance the headline could truthfully read "Suicides 'Linked to media scare stories'" though somehow I doubt that'd ever get printed.

Update - I note that Bad Science has found this story too.

3 comments:

Don B said...

There was a short piece on Radio 4 yesterday morning along the same lines as yourself. The comments there were along the lines that the observation by the GP was statistical rubbish and the Sunday Express was irresponsibly feeding the public's gullability and fears with pseudo-medical science.

FlipC said...

The Express being irresponsible? Never!

I think one of the things people forget with statistics is that the data that goes into the 'mix' is selected by the statistician. So why was the good doctor looking at mobile phone mast locations?

Statistically I could point out that of the 22 suicides he looked at 100% of them lived in Wales, which in itself could be classed as a reason to be depressed :-P Likewise 100% of them lived in Bridgend, with much the same conclusion. IOW he's looking at something that has already drawn attention to itself and trying to link it to something else.

Now if he'd taken a nationwide survey of suicides and plotted that against mobile phone mast locations and their strengths and found a correlation then I'd be paying attention. Even in that case though my alternative hypothesis may still be valid.

Ah for shame the online version of the story doesn't allow feedback.

Anonymous said...

Lies!

Damned lies!!

Statistics! >:-D

quod enim mavult homo verum esse, id potius credit.