How to lie with statistics

I can't tell you how much I wish the book How to Lie with Statistics was never written.  Sure, it contains a ton of useful information and a lot of tools for the layperson to actually think critically about the way data is presented.  Unfortunately, it has also introduced a meme that somehow translates to "every claim ever made that is backed up by a statistics must be a lie."  Yes, clearly, THAT is what the author was trying to say.  If I had a nickel for every time I've had a basketball argument with somebody that basically said "Your stats are meaningless; watch the game and you'll see the truth," I'd own an NBA team, and I'd hire the Wages of Wins network to run it, and we'd win several championships while having a lower-than-average payroll (thus making me even richer).

Aside: One of the problems is the title.  It's one of those titles that lets people believe that they don't actually need to read the book to get the point.  Sigh.  The book should have been titled "How to present data in a manner that misleads the viewer about what the data is really saying".  That would have sold about 8 copies (all friends and family of the author) and I wouldn't have to put up with this stuff.

Here's the thing.  When I present an argument based on a statistical analysis (such as my claim that Kevin Love is better than LaMarcus Aldridge -- you knew I'd bring that up, right?), if you say something like "The stats cloud the truth" or "You need to get your nose out of the stats and watch the game" or "Anyone can SEE that this not true," you are, let's face it, just making shit up.  You're just saying "that's not true because I know better."  And there's a reason that when you tried that argument in your 8th grade science class, your teacher was unmoved.  If you think that the statistics are lying, you need to demonstrate how or provide some reasoning that a particular piece of data misrepresents the truth.  For instance, when David Berri critiques John Hollinger's PER, he mentions that the metric does not penalize players for missed field goal attempts, and goes on to explain why he believes so.  This is the basis for a debate.  John can refute this claim by presenting some further data if he so chooses.

Claiming that I don't watch the games isn't a good refutation (and, let's be serious here, you are going to claim that you have seen more TIMBERWOLVES games than me?  Oh please.  Nobody else watches them play).  Or claiming that you "see" things that I do not "see" when we have both watched the same amount of basketball is also pretty unlikely to lead anywhere but us shouting "Yes!" and "No!" back-and-forth.

By the way, there's nothing wrong with expressing an opinion.  I'm fine with you saying "I think Aldridge is better than Love and you'll never convince me otherwise." I'm not ok, however, with you dressing up this opinion as fact unless you want to get into a debate.  One that, unfortunately, is going to have to involve stats, because you will forgive me if I don't accept "because everybody can see that" as actual evidence.  And if you want to tell me the statistics are "lying", you have to actually explain why.

Loading...