Lies, Damn Lies, Statistics, and Quinnipiac Polls

Let’s say you poll a random sample of 1,000 residents of Boston and New York City, each, about who they think the best football team is. You could expect a few to pick who they think will win the super bowl but, by and large, people in Boston will say “Patriots” while people in NYC will say either “Jets” or “Giants”.  Nothing surprising there, and you really wouldn’t learn anything other than that people root for their home team.

Add up all the choices, and you’ll find that about 50% of the people you asked said Patriots (the 1,000 people you asked who are in Boston), about 30% said Giants, and 20% said Jets(1).  No surprise, right?

So what do you tell people? Why, you tell people that “People in Boston and New York said that the Patriots are their favorite team”!  Because, of course, in your poll, the Patriots came in first.  And what would people say about your analysis? Well, in Boston, they’d say “of course, that’s right!” and you’d get a pat on the back (so to speak).  In New York they would say “bite me” (or similar) and I hate to think what they’d do to you.  Everywhere else they’d just call you an idiot for your analysis.  And, to be fair, you’d actually be an idiot for making such a claim.

So it’s surprising, to say the least, to see that Quinnipiac University has become the source of just this kind of idiocy. They conducted a poll that asked people who the worst post-WW2 president is.  And in this version, Boston, the republicans, overwhelmingly said “Obama”.  And the New Yorkers, aka the democrats, split their worst-president vote between their two least-favorites: George W Bush and Richard Nixon.

And, so, Quinnipiac reports, breathlessly, that “Obama Is First As Worst President Since WWII”.  Just like the Patriots are first as the favorite football team in Boston and New York City.  It’s not really true except in a meaningless way, but it sure sounds shocking, and it is getting them a lot of publicity during an otherwise very quiet week.

It’s sad to see a major university stoop to clickbait headlines that suggest conclusions not supported by the data. But if all you really want is attention, that’s one way to get it…

(1) That 60/40 breakdown is a guestimate based upon some quick googling.  If it’s wrong, it doesn’t affect the analysis.  Nonetheless, I apologize to Jets fans unreservedly.