Monday, December 5th, 2016

Sometimes you get a bad poll

2

The Marquette Law School poll looks like it has a case of St Norbert’s disease.

A new Marquette Law School Poll finds President Barack Obama leading former Massachusetts Governor Mitt Romney 51 percent to 43 percent among likely voters in Wisconsin. Five percent remain undecided or declined to state a preference, while 1 percent said that they would vote for a third party. Two weeks ago, before the second and third presidential debates, the poll found Obama at 49 percent to Romney’s 48 percent. In the U.S. Senate election, Representative Tammy Baldwin holds 47 percent to former governor Tommy Thompson’s 43 percent, with 10 percent undecided or not offering a preference. In the previous poll, Thompson received 46 percent and Baldwin 45 percent.

That’s certainly contrary to what I was told over the weekend, and it doesn’t look like the presidential campaigns believe those numbers are accurate.

If Wisconsin wasn’t in play, former President Bill Clinton would not be in Wisconsin today, President Barack Obama wouldn’t be visiting three times in the last five days, Vice President Joe Biden wouldn’t be here tomorrow, and Mitt Romney wouldn’t be here tomorrow, either. Congressman Paul Ryan just made three stops in Wisconsin. Obviously the campaigns think the race is closer than the poll would indicate.

Indeed, Politico is reporting an Obama campaign spokesman says that the race is close in Wisconsin because Ryan, “…basically moved into the state.” (No, we can’t make this stuff up.)

I also have to question the Marquette Poll since a poll with the same partisan spread (Democrats plus 5) shows only a three-point difference, and the trend in that poll is movement towards Romney.

Part of the problem is that public polls are having a harder time identifying a likely voter. From the Law Poll’s website:

Turnout could play a big role in the election, according to Poll Director Charles Franklin. “Among all likely voters Obama leads by 8 points, but among those who both are likely to vote and also follow politics most closely, the margin is just two points, 48-46 percent. It works to Obama’s advantage if the less interested voters turn out, while it improves Romney’s chances if they stay home. This shows how get out the vote efforts of both parties can affect the results.” The Senate race similarly tightens among likely voters who follow politics most closely, narrowing to 47 percent for Baldwin to 46 percent for Thompson.

But what’s really frustrating is the inconsistent partisan make-up of the poll. On the website, Franklin said that the partisan voters are consistent in their support for their candidate. So it’s not surprising that when the poll has a net shift of four points towards the Democrats that there is a corresponding shift in the results of the poll:

Last poll 10/11—Likely voters

Republican 46%
Democrat 47%

Tommy Thompson 46%
Tammy Baldwin 45%

Barack Obama 49%
Mitt Romney 48%

This pol 10/25—Likely Voters

Republican 40%
Democrat 45%

Tommy Thompson 43%
Tammy Baldwin 47%

Barack Obama 51%
Mitt Romney 43%

If there was a shift in the basic fundamentals of the race, perhaps such a shift in the polls would be understandable. However, even the MU Law Poll website points out there hasn’t been a shift in attitudes towards the candidates since the first debate. And, as Brian Fraley points out, the poll some pretty odd results:

“This result makes me question the accuracy of the entire poll:”

Who’d do better job on the economy?
Obama 49% Romney 47%

Think that’s weird? Take a look at the results of the foreign policy question.

Who’d do better job on foreign policy?

Obama 54% Romney 40%
The challenger bled among independents in the last two weeks? Why, because of the debates?

Unfortunately we won’t get another poll out of Marquette before the election that better reflects the actual electorate. I don’t normally say question the polls, but question this one.

Be Sociable, Share!

Print this entry