Kirby: Believe 3 percent of campaign rhetoric 1 percent of the time
Earlier this month, presidential candidate Mitt Romney made the claim that 47 percent of Americans pay no income tax, believing they are victims entitled to all sorts of government handouts.
Although the comment was made in a private meeting, fact-checking wolves immediately appeared and ate him.
OK, that "wolves" remark was 60 percent satirical, 18 percent mean-spirited, 15 percent insult, and 7 percent I have no idea because it's 4 o'clock in the morning and I'm just pulling this stuff out of my, um, back pocket.
The point is that Mitt immediately relented and came up with more statistics indicating that he was 71 percent right about the 47 percent claim even though 91.3 percent of Americans think he got those stats from the same place I get mine.
NOTE: All of the above statistics have a 211 percent plus or minus margin for error. Like I said, it's early.
Here's the thing. Statistics lie. So says a quote generally (and almost certainly incorrectly) attributed originally to Mark Twain, "There are three kinds of lies: lies, damned lies, and statistics."
ANOTHER NOTE: Spare me the peevish emails about Mark Twain being the "real" author of this quote. Check it out yourself: http://www.york.ac.uk/depts/maths/histstat/lies.htm.
Statistics by themselves do not lie. They're merely inert numbers with no real need or capability of lying. Rather it's the spin people choose to put on them.
For example, if I claimed that 87.2 percent of men would rather watch someone set off a badger hole stuffed with dynamite than sit through an opera, I could just be spinning this statistic for my own purposes even if it were true. (It's not, but ought to be.)
The stat may have come from genuine research indicating that more than 87 percent of men snore during an opera, after which 62 percent of their wives won't talk to them for 100 percent of the ride home.
Most statistical claims don't involve research. They are in fact invented on the spot to support some half-baked claim or lay the foundation for a nonsensical point.
MY WIFE: "Where's the hood for your truck?"
MY BRAIN: "Quick, say something scientific."
ME: "According to a study released by Slumbolt University, women are 78 percent more likely than men to drive under a cow."
That was not an actual fact. Don't quote me.
When it comes to putting your faith in any form of statistics, it's always best to consider the source first. Check the following claims and see if you can determine which one is a valid statistical claim.
DAUGHTER: "Like way over 99 percent of the kids at school totally have iPods."
SALES GUY: "Best of all, the 2013 Brat Hauler will outrun 89 percent of all police vehicles."
POLITICIAN: "Forty-seven percent of Americans expect the government to fund cosmetic butt lifts."
CLERGY PERSON: "God likes us 61 percent more than he likes those other people."
NEWSPAPER COLUMNIST: "At least 25 percent of my facts are completely accurate."
The answer of course is none of them. All five statistical claims are outright fabrications. The only one that even comes close is the one about butt lifts.
You'll hear a lot of percentages tossed around during the countdown to Election Day. At best, you can only believe 3 percent of all statistical claims, and then only 1 percent of the time.
Robert Kirby can be reached at firstname.lastname@example.org or facebook.com/stillnotpatbagley.
See more about comments here.