(or) I Wouldn't Touch Results With A Ten-Foot Poll

While I know exceptionally little of the art of statistics analysis, I have worked for four different telephone research firms, including Gallup itself, and there are a few other variables that are more or less invisible.

First, no matter how much research firms attempt to get their sample indicative of the population, polls will always be skewed toward people who answer polls. Duh, you say, but that is an important factor. There are any number of groups who will never answer a survey and we'll never know what they have in common because they wouldn't answer a poll about it.

Second: interviewer technique. As a fairly successful telephone interviewer myself, my own approach was to sell the survey as patriotic duty and with that I was luckiest with older conservatives. Most of the younger, hipper crowd I would pitch the survey to would usually refuse or hang up, seemingly thinking their time is too important and that they are clever to have beaten me. So, that's why you liberals really should answer these polls. If you don't, you make the country look more conservative.

Third, the demographic of interviewers. Even if we don't talk about it, much of the time we can tell the race people are by their voice alone. I have completed surveys with people of all races, but most frequently they were with my fellow pale folk. People respond more readily to someone they think is like them. And for those of you who don't know, Gallup is headquartered in my native Nebraska, a state not well known for its racial diversity.

I don't really have a conclusion to this as I'm not a statistician, but perhaps those who are can use this information to better analyze the data available. Or maybe it's all just a mug's game.

 

Music For Your Ears To Taste: "Language City" by Wolf Parade

~S-B