With Labor Day over and the campaign season in full swing, a numerical review of one of the most competitive gubernatorial elections in the nation is warranted. This race, featuring incumbent Scott Walker against challenger Marry Burke, has been considered Walker's to lose ever since he handily beat back former US Representative and Milwaukee Mayor Tom Barrett during the 2012 recall election (54.1% to 46.3%). This victory was a gain over the 2010 election between the same two contenders, in which Walker won 52.3% to 46.5%.
Come below the fold for more analysis.
With the election still nearly two months away, anything can still happen. However, there is enough recent polling data and historical information to allow us to form some educated opinions. Specifically, we can look at the polling data from the most recent two months (August & September). This data is late enough such that both candidates should have name recognition amongst the electorate, while also reflecting the incumbent's electoral viability due to judgments made throughout the expiring term.
Likely voter polling data for Walker V. Burke over this time period is inclusive of :
We Ask America*
9/3-9/3
Walker - 44
Burke - 48
Marquette University
8/21-8/24
Walker - 46.5
Burke - 48.6
Rasmussen Reports*
8/13-8/14
Walker - 48
Burke 47
*Please note - FiveThirtyEight gives Rasmussen a +1.3 GOP house effect and that We Ask America showed Romney winning three states 30 days prior to losing them in the general election. Please take with a grain of salt.
Although wonton averaging is frowned upon by professionals, for the sake of argument, we will average here.
August - September Average
Walker - 46.2
Burke - 47.9
(Burke +1.7)
What does this mean? Not a whole lot. This small dataset shows an exceptionally close race with Burke in a better position than most thought she would be. However, perhaps the bigger item to take away from these three points, is that Walker is averaged at 46.2%.
We can glean a bit more info from this dataset if we look at other gubernatorial elections in Wisconsin over the last few cycles, with their respective polling data from August and September.
2012 Recall (6/5)
Final
Walker - 53.1
Barrett - 46.3
Marquette University
4/26 - 4/29
Walker - 47
Barrett - 46
PPP
4/13-4/15
Walker - 50
Barrett - 45
Marquette University
3/22-3/25
Walker - 47
Barrett - 45
Average
Walker - 48.3
Barrett - 45.7
(Walker +2.7)
During the recall, we see that at two months out, Walker was averaged to win 2.7 points more than Barrett, with pollsters likely voters giving him an averaged 48.3% share (2.7% higher than Barrett). Walker would go on to win, as predicted.
Granted, recall elections are a bit odd. Thus, let's take a look at the 2010, 2006, and 2002 elections. Incumbents (which Walker currently is) are given an I.
2010 Gov. Election
Final
Walker - 52.3
Barrett - 46.6
Rasmussen
9/15-9/15
Walker - 51
Barrett - 43
Rasmussen
8/24-8/24
Walker - 47
Barrett - 44
Rasmussen
8/10-8/10
Walker - 49
Barrett - 41
Average
Walker - 49
Barrett - 42.7
Here again, we see the relatively early polling reflecting the actual outcome, even when both candidates were non-incumbents, with the average difference of 6.3% being very close to the final difference of 5.7%. Please note, that 2010 is considered a GOP wave year.
Going back to 2006, what is considered to be a DNC wave year, we get the following data:
2006 Gov. Election
Final
Green - 45.3
Doyle (I) - 52.7
Zogby
9/11
Green - 45.7
Doyle (I) - 48.2
Zogby
8/28
Green - 43.1
Doyle (I) - 48.7
Rasmussen
8/20
Green - 41
Doyle - 49
Strategic Vision
8/18
Green - 44
Doyle (I) - 45
Research 2000
8/17
Green - 38
Doyle (I) - 48
Average
Green - 42.4
Doyle (I) - 48.2
The differences here are more pronounced, with the final difference being 7.4% and the polling average from this dataset being 5.9%. However, note that Doyle, the incumbent, was polling at 48.22% at this point in the game. Doyle was then in the same situation Walker is in now, facing the first election after one term as gov. Doyle was at 48.22% vs. a relatively unknown challenger, whereas Walker is currently at 46.2%.
So, we just talked about two wave years: 2006 and 2010. What if we go way back to a less turbulent time? How about 2002? Sadly, there isn't much data here, but let's see what it says.
2002 Gov. Election
Final
McCallum (I) - 41.4%
Doyle - 45.1%
Badger Poll
9/26
McCallum (I) - 35%
Doyle - 43%
Again, we see the August/September polling reflecting the ultimate winner, yet understating the amount of support for each candidate.
From this information, we see that, not only when the incumbent is down this early on, does the incumbent go on to lose, but that the August/September predictions have consistently reflected the final outcome. Although, it is exceedingly important to note that the final result has shown a smaller margin between the two than predicted. Furthermore, more voters come home, on average, to the GOP candidate v. the DNC candidate (4.4% to 2.9%) - although it varies by year. In 2002, McCallum's final vote was 41.4%, with the poll showing only 35% (McCallum increased 6.4% v. Doyle's 2.1%). However, in 2006, Doyle's final vote was 52.7%, with the polls showing 48.2% (Doyle increased 4.5% v. Green's 2.9%). Perhaps this shows that more voters come home to the incumbent? This would hold true during the recall where Walker's polling shows 48.3% , with the final value being 53.1% (or a 4.8% increase). This is compared to Barrett's 0.6% increase (45.7% to 46.3%).
Please understand that 2002 and 2012 were both odd years, with 11% of the vote going to the Libertarian (former Gov. Thompson's brother Ed) in 2002 and 2012 being the recall. So, salt, salt, salt.
Now, let's jump back to the present. We know that polling can be an art, as much as a science. So, let's expand the window, but only look at Marquette University's post primary polling, as they are considered the gold standard and the DNC candidate would have been chosen.
Marquette University
8/21-8/24
Walker - 46.5
Burke - 48.6
Marquette University
7/17-7/20
Walker - 46
Burke - 47
Marquette University
5/15-5/18
Walker - 46
Burke - 46
Just looking at Marquette's polling, we see that Burke's share of the poll has increased as she has become more well know in the state. From May through August, she has increase her share by 2.6%. Walker, on the other hand, has increased his share by only 0.5%.
From all of the information above, we can see that Burke is in a stronger position than many would have guessed prior to the primary, where Marquette had Walker at 48% and Burke at 41% in March. She should also take some comfort in the fact that every candidate ahead at this point went on to win. Perhaps a greater takeaway is that Marquette's gold standard doesn't see Walker breaking 46.5%. Every candidate to win since 2006 has polled at least 47% by this point (with 2002 having the GOP incumbent at only 35%). She may also appreciate the Marquette trend line moving in her directing.
Now, all things being equal, Burke may be able to argue for the edge based upon historic analysis. However, things to note:
1) Debates are coming. Walker debates like Dick Cheney. Hard to rattle and could convince just about anybody that orange is purple and the sun will rise in the west. Burke has never debated at this level, if ever at any level.
2) John Doe documents will likely continue to be forthcoming. The 7th Circuit Court of Appeals will hear more arguments in the case on the 9th. Although this has had little impact on this race (or the recall, for that matter), it doesn't help Walker.
3) Walker has more money flowing in than any Wisconsin governor has ever had. The undecided are being pushed harder than ever with the GOP camp having the noise advantage (by an order of magnitude).
4) Burke is still making a name for herself. She is less known than any gubernatorial candidate in the last 12 years (Barrett was mayor of the largest city and a former US Rep, Walker was county executive of the largest county, Green was a US Rep, Doyle was AG, while Burke was an unelected commerce secretary seven years ago).
5) Walker's job record is 107k jobs vs. 250k jobs promised. Burke is hammering this and will continue to do so. The odds of those numbers getting significantly better in the next two months are not good.
So, that about sums up where things are.
PS - It is surprising that the DNC Chair has been so quick to retreat from her negative comments on Walker regarding women's rights under his leadership. Although her language was unprofessional, it would seem that bringing up the following points instead of retreating would have been more prudent (from her perspective)
- Walker fired a female MD while County Executive due to her being a model while she was a college student. This, while Milwaukee County had a shortage of doctors to begin with.
- Walker refused to investigate rape at the Milwaukee County Mental Health Complex, despite overwhelming evidence to do so.
- His campaign referred to his own Lt. Gov. as "a one women circus", "radioactive", "not worth [his campaign's] time", "the bane of [his] existence", and hoped she'd "topple over in her high heels". Not to mention, his team was "really beginning to dislike her". Odd language to endorse regarding a partner...
- Signed legislation forcing women to be vaginal probing prior to receiving abortions (overturned by a federal court - with his office still appealing).
- His campaign was also quoted as thinking that single mothers who received state assistance were "hilarious".
If anything, it would look like Walker and his campaign thought little of women's rights, women candidates, or women down on their luck. It would seem like that would be the statement to make. Also, most of the above points are from the Rindfleisch files. Look it up.