← Back to News List

Nerdy Pet Peeves

by David Hoffman

I'm kind of a nerd, and there's no shortage of evidence. I can tell you which episode of the original Star Trek series aired for the first time on the day I was born (it was this one, thanks for asking). I became a baseball fan largely out of fascination with the statistics on the backs of baseball cards. I have devoted considerable time to thinking about how I would explain 21st-century U.S. society to Abraham Lincoln, should one of us transcend the space-time continuum so we could meet. Seriously.

And it is specifically as a nerd that I am bothered by the following logical fallacies and misuses of data and language:
  • TV weather forecasters referring to "normal" temperatures (as in, "Today's high of 74 was ten degrees above normal") when they mean "average." There are two problems here. First, the average high temperature for a given date encompasses many past instances of higher-than-average temperatures for that day, as well as many instances of lower-than-average temperatures for that day. There is nothing abnormal about a day on which the temperature does not exactly equal the long-term average. Second, average temperatures are increasing over time due to climate change--so more and more, the average temperature is not "normal" at all. A hundred years from now, when the polar ice caps have melted and coastal cities are under water, will the weather forecasters be saying "Well, today's 120 degree high temperature seemed hot, but that's perfectly OK, because 120 degrees is normal for this date"?
  • "Nelson Cruz led the American League in six offensive categories." "218 high temperature records were broken today." Those sentences seem to be telling you something very specific. But do you see the problem? I think of it as the "statistical numerators for which there are no agreed-upon denominators" problem, though there is probably a  better name for it. Here's the issue: Exactly how many offensive categories are there in baseball? (There is no answer, because there are infinite ways to keep track of information about players' performance). Exactly how many places are keeping track of temperature records, and do those places overlap? If you don't have answers to those questions, the initial pieces of information are essentially meaningless.
  • The use of the phrase "statistical dead heat" or "statistical tie" to describe a polling result where the difference in support for two candidates is within the poll's margin of error: "Candidate A leads Candidate B by 52%-48% in the latest poll, but the margin of error is +/- 4%, so it's a statistical dead heat." Wrong! This is not a tie. A more appropriate interpretation of this poll result is: "While we don't know for sure that Candidate A leads Candidate B by 52%-48% among the population sampled, that's the most likely scenario. And while there's a small chance that support for the two candidates among the population sampled is 50%-50%, there's an equal (small) chance that Candidate A's lead is twice as large as the poll indicates."
And don't even get me started on time travel stories that are internally inconsistent regarding the rules of time travel. Either you can alter the present (or create an alternate timeline) by changing the past or not. You can't have it both ways!

I'm guessing I'm not the only one at UMBC with a few nerdy pet peeves, right? Right??

Co-Create UMBC is a blog for and about UMBC, written by David Hoffman and Craig Berger from UMBC Student Life. Join the Co-Create UMBC group on MyUMBC. Like Co-Create UMBC on Facebook. And follow David and Craig on Twitter. 

Tags:

Posted: October 19, 2014, 7:04 PM