“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 in “Chapters from My Autobiography”
Statistics are the kind of lie you can tell with a straight face. A defect for the many.
One of the optimal statistical tricks is end-point bias. The condition of choosing the beginning and end of a some data range so as to make your point stronger. It is being used universally by the purveyors of precious metals. “Since 2000, precious metal prices have increased by 800%. How has your stock portfolio done?” Sound familiar. It is true but nonetheless close to being a scam.
An interesting question and one worth studying. From Dec 2000 to Dec 2012 gold appreciated 16.2% annually while The Dow Jones Industrial Average appreciated 0.8%. Compelling, no? A clear choice.
I can make a statistically valid comparison that I might argue has the weight of long history on its side. From December 1974 to December 2012, the Dow increased 8.4% annually. In the same period, gold appreciated 5.8%. And it costs money to hold gold while stocks pay dividends. A clear choice.
Yeah, but current history matters more. Maybe, but not certainly. Lets take the last two years to the end of September 2013. The Dow has appreciated at an annual rate of 18.3% while gold has lost money at the rate of 9%. So there!
Statistics can be useful, but only if you understand where they come from and how to approach them.
The lies in statistics that many people use and find effective are based on two points.
Here are some other examples.
Statistics and probability are not intuitive.
On the probability front, few people think intuitively that if there are 23 randomly selected people in a room, there is a 50% chance that two of them celebrate their birthday on the same day of the year.
You need to understand how a statistic was constructed before you decide it is compelling. If I told you 96.3% of people surveyed favored increasing the monthly welfare benefits would you still take it seriously it if I told you that only people on welfare were surveyed? Some statistics that you see are that misleading.
Because the people who create them know that it works.
Be more cautious. End points matter, relative populations matter, population selection matters, correlation is not causation.
When you “know” something that is wrong you make more mistakes than when you know nothing. When you know nothing, you can at least be right by accident.
Don Shaughnessy is a retired partner in an international accounting firm and is presently with The Protectors Group, a large personal insurance, employee benefits and investment agency in Peterborough Ontario.
email@example.com | Twitter @DonShaughnessy | Follow by email at moneyFYI