The monthly blog from Andrew Sharp, Director General of IARO - January issue

• There is a famous quotation, “There are lies, damned lies and statistics”. Its originator is difficult to track down, but it could have been Benjamin Disraeli. Whoever it was had the right idea – the book, “How to lie with statistics” is one regularly commended to students!

  Not all misused statistics are attempts to deceive – just attempts to place one’s own viewpoint in the most flattering light. This is a waste of time if people then find an alternative set of data which undermine your own!

  Take, for example, the aviation industry’s comments about CO2 emissions – effectively, they are so small that they are not worth bothering about. Indeed, if you look at the latest “Transport Statistics Great Britain”, you will find that domestic civil aviation accounts for just 0.4% of total national CO2 emissions. However, if you look in more detail at the same table, you’ll find that domestic rail also accounts for just 0.4% of national CO2 emissions. And, of course, domestic rail does significantly more work!

  Another fact which always irks me is the one IATA use – that air is the safest means of transport. So it is – if you use one measure of risk. Again using domestic UK figures, air transport has 0.05 fatalities/billion passenger kilometres: rail has 0.7 and car 3.0. However, change the exposure measure to hours and the ranking changes. Rail has 30 fatalities/billion passenger hours, air 31 and car 130. Change the metric to journeys and the ranking changes again – rail at 20 fatalities/billion passenger journeys, car at 40 and air at 117! Not the safest!

  It’s obvious when you think about it. Landing and take-off are the most dangerous part of a flight, so if you use a measure of risk – passenger-km – which minimises this, you get the most favourable answer!

  And, of course, airport workers are notoriously prone to accidents – something IATA is tackling, for excellent reasons.

  Moral of the story? Don’t use statistics as a drunk uses a lamp-post – for support rather than illumination! Get to the bottom of any statistics you use and explain clearly what you are using. If you don’t, the opposition will find alternative statistics or will pick holes in yours. The sound-bite is less good because you have to add an explanation – but it does demonstrate you are taking the trouble to check your data!

  To end with another un-attributed quote – “Figures can’t lie, but liars can figure”! I hope you enjoyed reading this: I look forward to your feedback.

  Andrew  

  

  5 Jan 2010

  Posted by: Andrew Sharp