With half of the voters surveyed just over a week before Election Day saying they planned to vote, or already have voted, for Udall, 43 percent said they favored Weh.
Eight percent said they were still undecided in the U.S. Senate race. Absentee voting for the Nov. 4 election began Oct. 7. The total for the poll’s head-to-head question amounts to 101 percent because of rounding error.
An earlier Journal Poll taken in mid-September showed Udall with a 13-point advantage in the Sen-
ate race – 51 percent backing Udall and 38 percent favoring Weh. Eleven percent of voters were undecided in the earlier survey.
Since September, Weh appears to have made gains by connecting with conservative voters, particularly in rural areas of the state, said Journal pollster Brian Sanderoff, president of Research & Polling Inc. But that has not resulted in deep cuts into Udall’s base of supporters, he said.
“Tom Udall has had a comfortable lead,” Sanderoff said. “The race narrowed as people began hearing Allen Weh’s conservative message, but getting to the point that the race could turn around, that would be a real uphill battle for Allen Weh.”
Sanderoff said the Senate race in New Mexico is probably affected by a “national mood” of dissatisfaction with a gridlocked Congress and Democratic President Barack Obama, but that national current doesn’t seem to have taken a major toll on Udall’s popularity in New Mexico.
As a result, national political groups looking to influence control of the Senate have not been active in the Udall-Weh campaign so far, Sanderoff noted.
Weh also showed slight improvement in his Republican base, with 81 percent of GOP members saying they backed the Albuquerque businessman and retired Marine colonel.
Udall’s support among Republicans lagged at 12 percent. But Udall showed strength among independents, or voters who decline to state a party affiliation.
The survey showed Udall with support from 53 percent of independents, compared with Weh’s 37 percent. Ten percent of independents were undecided.
Udall’s strength among independents is a shift after the September Journal Poll showed independents almost evenly split between Udall, at 41 percent, and Weh, at 38 percent. In September, 22 percent of independents remained undecided.
Ethnicity and gender
Udall held a strong advantage over Weh among Hispanic voters, with 62 percent favoring Udall and 27 percent supporting Weh.
Weh has kept his advantage over Udall among Anglo voters. Fifty-two percent of Anglo voters said they supported Weh,; 43 percent said they backed Udall.
Male voters for the first time in the Senate race leaned toward Weh. Forty-nine percent of men said they preferred the Republican, compared with 46 percent backing Udall.
A month earlier, Udall led among men by 10 points, with 50 percent to Weh’s 40 percent.
Female voters said they favored Udall 53 percent to Weh’s 37 percent.
“The gender gap is big here,” Sanderoff said.
In the Albuquerque area, Udall held a steady and significant lead over Weh. In the state’s Democratic-leaning north-central region, Udall grew his lead to a nearly 3-to-1 advantage.
But Weh has a strong new lead in the state’s southern and southwestern region. Weh also built on his advantages in the state’s Republican-leaning eastside and northwestern region.
The Journal Poll is based on a scientific, statewide sample of 614 proven voters who cast ballots in the 2010 and 2012 general elections and said they were likely to vote again this year.
The poll was conducted Oct. 21-23. The full voter sample has a margin of error of plus or minus 4 percentage points. The margin of error grows for subsamples.
Research & Polling Inc. generated a random sample in which each New Mexico county received a representative proportion primarily based on turnout patterns in the 2010 general election for governor. When necessary, Research & Polling weights the surveys to reflect the known distribution of age, gender and party affiliation, based on the 2010 election. Historically, voter turnout is much lower in non-presidential election years, such as 2010 and 2014.
Racial and ethnic proportions are based on Research & Polling estimates of turnout patterns.
All interviews were conducted by live, professional interviewers, with multiple callbacks to households that did not initially answer the phone.
Both landlines (67 percent) and cellphone numbers (33 percent) of proven general election voters were used.