News you can use

Polls are hogwash, but boy, are they fun

I ask you, "What makes a presidential election year more fun than eating roasted crickets?" Well, sure, there's watching millions being blown. Money better spent, say, reducing the federal budget deficit — but I'm just being silly. Candidates believe practice makes perfect — just in case, you know, they're elected. Dollars aside, the main attraction is election polls.

It's impossible to pick up a newspaper with a nuclear sub parked on it. Go ahead, try. Yet inside papers everywhere, reliable election polls by outfits like Butch & Seabass Swamp Tours and Polls has Santorum leading Romney by 5 percent in Alabama or, maybe, falling behind by 6.7 percent. It could be almost anything depending on the fishing that day.

Wonder where these polls come from? Me neither, but here's my best guess:

1. Find a target group of voters who potentially, based on last census, are alive. In the polling business this is know as N, found in popular politically words like "nincompoop." Acquiring the correct N is rather tricky since it's dang near impossible to question everyone unless there's free beer and pizza. So during the dark ages (1032 to 1345 and pretty much all of 2008) pollsters picked a representative group by employing a stratified random sample technique. This has the advantage of sounding really cool and being random, as in choosing a prom date or toilet paper.

2. Devise questions so soul-searchingly simple that only an honest answer is possible; as demonstrated by this real-world dramatization:

Norm: Harriett, would you get me another beer?

Harriett: Get it yourself!

3. Collect, conjugate, massage and morbulate (from the Greek "morbull" or "to cook in olive oil") the data. Data are best quickly injected into statistical formulas to reduce spoilage. These are excitingly confusing as in, π(2X + 3Y)/cosine Z=% where Z = the birthrate of a mythical woodland creature with large pointed ears. The result is a percentage. Duh, just look at the formula which allows a pollster to use insightful statements like, "The concept of degrees of freedom reveals sample variance failing because … hey, what slime bucket ate the last blueberry cake donut?!."

So, when the Queen-of-Hearts Research Group reports Ron Paul leading all GOP candidates in Guam by 20 percent, first, question the sample used. Did they survey Ron's immediate family or did they include second cousins, live-in servants, and illegal immigrants who clean the pool? Then consider if Norm and Harriett have or would ever vote. Finally, decide if there really is a way to trap cosines in paper bags.

This all proves polls are often hogwash, derived from putrid data much as an alchemist creates gold from a moldy pan of lasagna. Dig in but, like watching polls of this presidential election year, don't be surprised if you get indigestion … actually, you'll probably just want to vomit.

(Joe Barnhart is a free-lance writer who lives in Dillon.)
 

Reader Comments(0)