UPDATE: After most recent poll, Republican spinmeisters going crazy

I noted yesterday that Minnesota Democrats Exposed was furiously spinning the good news for the Franken campaign from the latest poll. I didn’t think the situation was so bad for them that they needed to start breaking out the lies already. Apparently, though, they’re getting that worried about the state of this race.

Norm “W” Coleman was never going to be able to win this race on his record. The fact is, he hasn’t voted in the interests of most Minnesotans, although over the past two years, he’s tried to lay claim to being a moderate. The only way his campaign could win was through personal attacks, and now that Franken has managed to climb out of the muck, Coleman’s in serious trouble. How much trouble? Enough that conservative bloggers are distorting his poll numbers to shore up the base.

Yesterday, MDE made the following claim:

In a three-way race, Coleman still holds a statistically significant lead over Franken, while Franken still can’t break 40% of the vote. [emphasis added]

The fact is, because of the poll’s margin of error, the Senate race is now a statistical tie! Coleman’s lead is statistically insignificant, despite Brodkorb’s claims to the contrary. I know that scares the heck out of the conservatives, but the numbers don’t lie.

A more detailed explanation of why this is a statistical tie follows after the break [WARNING: boring technical stuff ahead].

The technical details

I should emphasize here that there’s a high probability Coleman is still in the lead. However, his lead is much smaller than last time; so much so that we can’t say with certainty that he really is winning. Here’s why. The American Association of Public Opinion Research (AAPOR) explains that “We often talk about a MOSE as if it applies to the whole survey, but note that in reality it applies to each answer in the survey”. Taking that into consideration, here are the possible ranges of Franken and Coleman’s numbers, at a 95 percent confidence level:

Coleman: 42.2 to 49.8
Franken: 35.2 to 42.8

The pollster is 95 percent confident that Coleman and Franken’s numbers fall within the ranges above. Why such a large range? It’s because of the error inherent in any survey, since we’re only taking a sample, and not asking everyone. Because of that error, we can’t be sure that Coleman’s really ahead. There’s a possibility that Coleman really has 42.2 percent of the vote, and Franken 42.8 percent.

Let me emphasize again: Coleman is probably still in the lead, if not by much and not for long. But I find it amazing that GOP spinmeisters are claiming his lead is statistically significant when it clearly is not. Is this just ignorance of how polls work, or is the GOP lie-machine revving up to try to lie Coleman back into office?

Related posts: Want more Daily Liberal? Subsrcibe or bookmark this site to keep up to date!




Comments

No comments yet.

RSS feed for comments on this post. TrackBack URI

Leave a comment