Friday 18 January 2013

Vital Statistics

New year, new venue and a fair few new faces too.  After a break for Christmas we held our third event on Thursday.  There were worries that the cold weather and the two month break between meetings might mean a lower than usual turnout.  We need not have fretted.  Local Skeptics did us proud as nearly 40 turned up to The White Horse to hear Dr. Kevin McConway talk about the use and misuse of Statistics in the media.

Kevin explained that, whilst some may yawn when they hear the word 'statistics', in reality everyone is interested in them.  To prove his point he showed us a variety of newspaper reports all reliant on stats.  But whilst evidence based reports ought to be better than those with no evidence at all we must examine exactly who is supplying the data and their motivations for doing so.  This is especially important when such data is everywhere but the skills to analyse it are not so abundant.

For example, Kevin showed a newspaper report highlighting the increasing threat of expensive baby buggies being stolen.  However, the company commissioning such research was the PR representative of an insurance firm launching cover for expensive buggies. Of course, this isn't to say the report wasn't true.  But being aware of the reasons for the commissioning of the study in the first place adds valuable context.

In another example, Kevin showed a report under the headline 'Careless Pork Costs Lives' that claimed eating bacon increased the chances of getting cancer by 20%.  In fact, the bare stats are that the chances go from 5% to 6%  - yes, a rise of 20% but a risk level that many fry-up aficionados may consider acceptable.

The overall message from Kevin is that we should enjoy the use of statistics in the media as often such reports can be very informative.  In fact, Kevin thought that the tabloid press was better at getting messages across than the more upmarket papers. But we should always ask -

  • "Who is producing these statistics?"
  • "What is their agenda?"
  • "What is the sample size?"
  • "Is it a single study?"   
  • "Is it just observing or is it experimental?"

Kevin works with journalists (including the BBC programme 'More Or Less') and stressed that the vast majority of journalists hate to get their facts wrong and that in recent years their use of statistical data has improved greatly. Yet journalists and statisticians have an equally low opinion of each other which is a shame as working together would be mutually beneficial.

After a break to refresh our glasses (20p off a pint of ale for CAMRA members, by the way!) Chris fielded various questions on different types of statistical analysis and political opinions polls. Luckily, Bedford's own local 'Stats Celeb' Mike Smithson (of the UK's most read political blog politicalbetting.com) was on hand to add his expert opinion on, well, opinion! Since Mike backed Obama at 50-1 a whole three years before his 2008 Presidential victory then he clearly knows his stuff. And so another stimulating evening drew to a close.

How about some useful links should you want to know more stats stuff?


Our next event is on Thursday 21st February when Stephen Law, Senior Lecturer in Philosophy at Heythrop College, University of London, will be telling us how to avoid 'Believing Bullshit'.  It certainly isn't bull to advise you to get there in good time, for good company, good food, good ale, good debate and....er...a seat!