6.18.2006
Newspaper Survey: Error Margin +/- %1000?
The Plattsburgh (news)Paper is still trying to stay hep with the Internet Age. For example, through the PeePee Web site, it’s acting like a cool cat, getting interactive, man, with its readers. Far out.
For example, the online surveys it conducts from time to time. Recently one survey asked for reader feedback in regards to the job performance of a controversial politician. With real surveys, the margin for error is usually low, maybe +/- %2.5, meaning the figure given could be higher or lower by that rate.
But what is the value of a survey with a wildly high response rate that makes any margin of error impossible to calculate? According to a PeePee article published 6/10/06, most surveys generate around 100 responses; the highest response on one hot topic was 196 (gas prices). But the survey regarding the controversial pol generated over 4700 responses; a majority, 75 per cent, were in favor of the man staying on the job.
As the article notes, at this time the paper has no way to stop people from manipulating the vote in any way they choose. So the question must be asked: why bother having a survey at all? An editor responded that the surveys are not intended to be “scientific,” but only an indicator of the public’s views.
But how can a survey be even an indicator of popular feeling when there’s no way to prevent extreme manipulation? The only thing indicated is someone can fudge the figures and there’s no way the paper can prevent it. The concept of interactivity is invalidated.
But the Plattsburgh Paper lumbers along, a squaresville dinosaur facing extinction because it can’t adapt to the new scene. It poses as a mammal, covering itself with some cool-looking fur, yet in the end it just keeps laying eggs. It’s beat –- but not in a good way.
Maybe it should try playing bongo drums, man.
# posted by Stan Spire @ 5:31 AM 
