There's an old book titled, "How to Lie with Statistics," that is well worth the read. The basic rule is, don't believe anything you may infer from them. I have a second rule, never believe ANY numbers generated by a journalist (this after reading that crime had dropped 120% from the previous year. What? Were thieves returning items stolen from previous years?). Third rule, don't draw any conclusions from any percentages given without showing the actual raw numbers (story said, people diagnosed with 'x' jump 50% in one week. At that rate of growth, entire population at risk in just months. (Actual numbers went from two diagnosed cases to three, but wasn't reported.))
Then there was an actual ad that said something like, "65% of smokers say "A" cigaretts are as good or better than "B" cigarettes. Perfectly true, but the numbers were: A better than = 15%; A as good as B = 50%; A worse than B = 45%.
Look, most oeople are uncomfortable with numbers and oercentages, and you know if you are. If you are one, don't conclude anything by someone else's version of what they mean.
Biting the hand that feeds IT
