Margin of Error

Sorry, pet peeve time.

Reason, discussing Chris Christie’s candidacy:

Not only did Christie not get the traditional “bump” in polling that candidates get when they first officially announced, he’s dropped from 4 percent to 2 percent in a Monmouth University poll over that time period. With a 5 percent margin of error, it means Christie’s support could be at 0 percent.

He could be at 0 percent? No, that’s not what that means. I suppose I should give credit for not saying he could be at -3 percent, but clearly he can’t be at 0 percent unless the 2 percent in the Monmouth University poll were just flat lying.

He could be at nearly 0 percent, yes. Those 2 percent of a very small poll could be the only people in America who support Chris Christie’s presidency. But that’s still non-zero.

It’s like a “statistical tie”. If one candidate is polling at 30% and another at 20% in a poll with a 5% margin of error, that doesn’t mean the two are tied. It means that there is a VERY SMALL probability that they are tied, but that there is a very high probability that one candidate has a sizable lead.

I understand that in America’s public education system and Kardashian-driven pop culture, we’ve reached a state where most people think numbers are hard. But they’re not this hard, people!

  • Barking Dog

    “Have you actually met people?”