October 29, 2014 01:21 PM UTC

S360: Udall 45% Gardner 44%, Hickenlooper 46% Beauprez 43%

•  
• by: Colorado Pols

A new poll from local consultant outfit Strategies 360 offers a different look at the Colorado U.S. Senate and gubernatorial races, with a polling sample they consider to be more representative of the 2014 electorate in Colorado than most public polling we've seen–in particular, a more accurate sample of Latino voters. From their memo today:

Strategies 360 conducted a telephone survey of Colorado voters who are likely to vote in the 2014 General Election. Respondents were randomly chosen from a list of registered voters and interviews were conducted by trained interviewers in both English and Spanish. Interviews were conducted October 20-25, 2014. A combination of landline and mobile phones were called to ensure greater coverage of the population sampled.

A total of 760 interviews were completed. 604 interviews were conducted among a representative sample of likely voters statewide. An additional 156 oversample interviews were conducted among Hispanic likely voters. The sample was weighted to ensure a proportional demographic representation of the likely 2014 electorate. The topline margin of error is ±4.0 and the margin of error for Hispanic voters is ±6.7%.

Currently, Democratic Sen. Mark Udall (45%) and Republican Rep. Cory Gardner (44%) are locked in a statistical tie, with another 8% undecided and 4% supporting a third-party candidate. Several factors contribute to the stalemate:

A massive gender gap. Udall currently holds the same 17-point lead among women that exit polls showed Democratic Sen. Michael Bennet holding in 2010. Meanwhile, Gardner leads among men by a similar 17-point margin. Additionally, Udall leadsamong white women by an 11-point margin and Hispanic women by a 37-point margin. In contrast, Gardner leads among white men by a huge 23-point margin but is currently losing Hispanic men by 29% to 57%. 

Unaffiliated voters. While partisans on both sides are similarly united around their party’s nominee (83% of Democrats back Udall, while 80% of Republicans support Gardner), Udall has more effectively consolidated Unaffiliated voters, which helps negate an expected GOP turnout advantage this year. Today, Unaffiliated Coloradans prefer Udall to Gardner 48% to 37%. 

Hispanics. Much of the media coverage of the U.S. Senate race has centered on Colorado’s Hispanic vote, and for good reason. This race may very well hinge on Hispanic turnout. Currently, Gardner edges Udall among white voters 47% to 43%. In most of the other key U.S. Senate races in 2014, that might be enough for the Republican to secure a win. However, Colorado features the highest proportion of Hispanic voters of any targeted U.S. Senate race this year, and Udall holds a commanding lead among this critical voting bloc: 58% of Hispanic likely voters favor Udall while just 26% favor Gardner. 

Young voters. This race remains close in part because older and middle-aged voters have yet to offer a real edge to either candidate (voters 55 and older split 46% to 46%; voters 35 to 54 lean toward Gardner 42% to 44%). Meanwhile, Udall has built a 10-point lead among voters under 35 (46% to 36%). Furthermore, young voters are disproportionately undecided compared to the older age cohorts. Turnout among this group will be key to any Democratic chances of holding Colorado.

Here's the full memo from Strategies 360. In the gubernatorial race, Gov. John Hickenlooper's somewhat larger lead is attributable to both a large gender gap and a sizable lead among unaffiliated voters–52-35%.

If this poll more accurate than others we're seeing? We do think that Strategies 360's attempt to factor Latino voters gives them a qualitative edge over a lot of the public pollsters–some of whom admit candidly that they have no idea how to account for this pivotal bloc of voters in their surveys. Also, back in 2010, Kevin Ingham, longtime Colorado pollster now with Strategies 360, released a poll on that year's U.S. Senate and gubernatorial races that turned out to be dead-on. So yes, we're inclined to trust these numbers a little more.

At the very least, throw this poll into your averages, and note the conscientious attempt to get it right.