This is about fears of developing breast cancer, not Scottish politics, but it’s the best example I’ve seen for some time of how easy it is to generate anxiety with a badly-scaled graph and a comment from a supposed expert. I feel sure it’s worth remembering for the next time the graph is about Scotland’s economy or crime or drug deaths or whatever.
The BBC’s Head of Statistics (really?) presented the above graph and then, astonishingly, said:
‘But if you take the most commonly prescribed oestregen and progestogen it goes from 63 all the way up to 83.’
Because he doesn’t remind us that this is an increase from 63 out of a 1 000 to 83 out of a thousand or from 0.63% to 0.83% and thus an increase of only 0.2%, his ‘all the way up’ and the scaling of the graph make it seem particularly worrying.
The scaling of the graph is a classic, done perhaps for innocent reasons of appearance or visibility, but it is wrong. The vertical scale should go up to 1 000 so that the bars can be in accurate proportion, like this:
Not so scary?
And, what about context? All good journalists claim to do it. If you compare the risk of developing breast cancer when you smoke with the risk when you have HRT might that be useful information for the viewer? It’s a 0.2% increased risk if you have HRT and 14% if you smoke. That’s a SEVENTY times greater increase in risk. In the case of a woman with a family history of breast cancer it’s 35% and that’s a 175 times greater risk.
If you know of any comparable distortions about Scotland, please share.