To the Ministry's main lobby The Ministry Blog
concert setlists

9 June, 2005

Say it with stats

It's hardly news that statistics can be misquoted to convey any chosen message, but here's an example.  It has been reported in a local free 'newspaper' (really just advertising and 'taster' articles borrowed from a real newspaper produced by the same publisher) that:

Train users want industry bosses to spend money on getting trains to run on time rather than invest in safety improvements.
... 45% of people would rather catch a train on time than worry about how safe it is.  Just 21% said train and rail companies should spend money on safety improvements as a priority.

I suspect those last three words hint at the true interpretation. I can quite imagine that the survey, for IMechE, asked people to choose their single highest priority for investment, and found that 45% value punctuality most whilst 21% rank safety highest.
Yet that's their first priority, not their sole concern. I'd be very surprised if people were specifically asked whether they'd want improved punctuality at the expense of safety. Nor do I believe that people were asked outright whether money should be spent on safety, to which 79% said 'no'.

I may be wrong, but this looks like a lazy journalist has just used statistics entirely out of context to support a sensationalist fiction. Of course this is very rare, and no reputable newspaper would print such a story – right?


I suspect it may be even worse than that. Consider this scenario.

Two different questions are asked:

1) Would you rather catch a train on time or have it be safe (or something along those lines)? - to which 45% answered "on time" and 55% said "safe."

2) Which of the following do you think should be the priority for spending (list including safety, on-time performance, better snacks, new uniforms for the engineers, etc.)? - to which 21% answered safety and undisclosed numbers gave other answers.

With this scenario, it is entirely possible to have safety as a very high priority compared to on-time performance, but we just don't know.

I'm not saying this is how the survey results turned out, of course, just agreeing with your point that it is easy to distort surveys and statistics.

Reminds me of the story about the CEO who was scandalized to find that 40% of all sick days in his company were Mondays or Fridays!

Posted by Jon. at June 9, 2005 11:09 PM
Site Home Tull Tour History Annotated Passion Play
Day in the life... Page design and original graphics © NRT, 2003