An EPIC/MRA poll conducted in early February and released this week reveals that 71 percent of Republican voters support a second DeVos run for governor.  Not surprisingly, 78 percent of Democrats oppose a second bid by DeVos to be our state’s chief executive.  What do these numbers mean?  First, that Republicans liked DeVos and think he was the victim of unfair attacks, so they are willing to give him a second chance at seeking the job.  And second, that Democrats know the odds on fooling Michigan voters a second time with deceptive, false attacks about “jobs to China” are pretty slim.  Democrats must realize that if DeVos runs again, he will have a better policy agenda and be a better debater backed up by just as much money as the last time.  They also know the Democratic candidate in 2010 won’t be able to blame John Engler anymore (at least I hope not) and will be forced to defend the sorry and sad legacy of Governor Granholm.   And frankly, I think DeVos deserves another shot and by running, he could potentially head off a divisive and costly Republican primary.

The Freep/WDIV is out with a poll today. On first glance, it doesn’t seem like good news for Team Romney. But, like I said, on first glance.

Let’s look again.

Rudy Giuliani and John McCain are 1 and 2 in in the Freep/WDIV poll of whom MI Republicans favor for the GOP nomination. Newt Gingrich is next, followed by Mitt Romney. But there’s a big problem with this ranking and the tell came in a sidebar to the main story:

Plus-or-minus 8 percentage points???? Most people in the biz will tell you that if you get above 4 percent points on the margin of error (or MOE in polling parlance) then you have a bad sample. Seriously people, I think I could give you results within 8 percentage points just by guessing, that’s how unreliable the numbers are. But let’s look at how off these numbers could be.

Let’s assume the numbers for Giuliani and McCain were too high and Romney’s were too low (my blog, my hypothetical), and let us further assume that these numbers are off by margin of error, which is 8 percentage points. Doing some simple math, McCain and Giuliani’s numbers drop to around 22 and Romney’s number would be around 16 percentage points. That’s a six point spread.

Now, I am not saying that my hypothetical is the case, but given such a wide MOE, you just can’t trust the numbers as presented.

Another problem with the Freep/WDIV poll…it sampled Democrats, Independents and Republicans, and then tried to extrapolate who would support which Republican or Democrat in primaries, respectively, or their party’s nomination. By the Freep’s own admission, the poll sampled too few Republicans.

The other issue I have with the poll is, because it doesn’t offer its methodology or the questions asked, we don’t know who was asked the questions. Did they just ask anyone who answered the phone…likely voters…likely primary voters….likely voters who voted on previous primaries? It doesn’t say and we don’t know.

My suspicion is that pollsters asked whether a potential respondent planned on voting in the next presidential primary or participating in a caucus, and if they said yes, they were asked questions. That’s a very weak qualifier. A more accurate one would be to ask whether one planned to participate in the upcoming primary/caucus and have done so in the past. Previous vote history is a better indicator of future vote pattern.

Of course, the Michigan presidential primary is not until Feb. 2008, a full year away, making all of this piffle just mental gymnastics.

Technorati Tags: Politics, Michigan, Romney, McCain, Giuliani, Polls