Lies, Damn Lies and Statistics

I am, or at least I once was, an accountant, and I always had a slightly jaundiced view of the figures we produced which ‘proved’ a case.  I always felt that you could make figures argue almost any way you chose.  As the following story illustrates, I am not alone in this feeling:-

There once was a business owner who was interviewing people for a division manager position. He decided to select the individual that could answer the question “How much is 2+2?”

The engineer pulled out his slide rule and shuffled it back and forth, and finally announced, “It lies between 3.98 and 4.02“.
The mathematician said, “In two hours I can demonstrate it equals 4 with the following short proof.”
The physicist declared, “It’s in the magnitude of 1×101.”
The logician paused for a long while and then said, “This problem is solvable.”
The social worker said, “I don’t know the answer, but I a glad that we discussed this important question.”
The attorney stated, “In the case of Svenson vs. the State, 2+2 was declared to be 4.”
The trader asked, “Are you buying or selling?”
The accountant looked at the business owner, then got out of his chair, went to see if anyone was listening at the door and shut the curtains. Then he returned to the business owner, leaned across the desk and said in a low voice, “What would you like it to be?”

So that and the fact that Maths was never my strongest suit meant that I am a little sceptical of statistics, but don’t they sound SO convincing..?!  “9 out of 10 women agree that Magicreem made them look 10 years younger” the advert boldly proclaims, Then down below in tiny writing you read that they asked 127 people.

All this preamble to shed a little mathematical backup to my native doubt.  A common example involves the mythical paedophile-spotting device, it hunts down and kills people that it identifies as paedophiles with a 99% success rate and a 99% chance at properly identifying a innocent person correctly. One would assume that if, out of population of 1 million people, 100 of whom are paedophiles the box identifies a person as a paedophile, there is a 99% chance it’s correct. In reality, it’s a lot closer to 1%. The reason being that the box falsely killed 1% of non-paedophile (9,999 people), as well as correctly killing for 99% of real terrorists (99 people).

Facts and data a very important in helping us make good decisions, but there is always room for good old fashion common sense and instinct. 

Tags: ,

Leave a Reply

This blog is kept spam free by WP-SpamFree.