gotta love statistics

Very funny!

In my other life in a galaxy long ago and far away I used statistics all the time to devise sampling plans and process controls. I attended a lot of statistics courses.

One of the most memorable lines about statistics that I ever heard came from J. M. Juran who was the "Father of Process Quality Control". He said, "If your boss does not understand statistics he will not trust you. If he does not understand statistics you should use pie charts and graphs. If he does not understand pie charts and graphs he will not trust you. If he does not understand pie charts and graphs you should use stick figures. If he does not understand stick figures you probably should not be working for him."
 
Rob, that was pretty funny all right.

"If your boss does not understand statistics he will not trust you. "

The correlary is if he does understand statistics he probably shouldn't trust you (or at least your data) either. Having even a vague understanding (and in my case the vaugeness perhaps yields even less trust of most reported statistics than usual) the possible types of independent variables and other confounding factors leads to a whole lot of yeahsurebutwhatabout's and haveyoucontrolledforthe's in your life.

This isn't to say that stats aren't useful, quite the contrary its just useful to understand how the reported got to where they did and to keep a very open mind that you might only be finding correlation or indirect causation.

I also find it pretty amazing what some people CAN reliably determine with statistics. An excellent example that sticks in my head for some odd reason is what some fellow pulled up from http://research.google.com/archive/disk_failures.pdf (please feel free to ignore that, its long tedious and only really of interest to people with way to many computers) yielded this little gem: http://slashdot.org/comments.pl?sid=222978&cid=18058500 which was widely lauded in certain circles as being a rather ingenious use of statistics to get at what was presumably intended to be hidden data.

Also todays xkcd is astoundingly relevant: http://xkcd.com/1478/
If you peversely enjoy this sort of thing (and it appears a few of you do): http://www.badscience.net/category/statistics/


In this case it appears that the underlying data is reasonably solid and the results are a bit amusing.
 
Figures often beguile me, particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: "There are three kinds of lies: lies, damned lies and statistics."
- Mark Twain's Own Autobiography: The Chapters from the North American Review
 
Top