RC Poll: Allen Holds Narrow Lead, but Within Margin of Error; Poll Shows Strong Support for Marriage Amendment

George Allen, the incumbent Republican holds a narrow 3 percent edge over Democrat Jim Webb, but the lead is within the margin of error, according to a poll conducted by The Center for Community Research at Roanoke College.  Independent candidate Gail Parker garnered the support of just two percent of those surveyed.  The Poll also found strong support for the marriage amendment with 53% saying they would vote for its passage. 

The Poll includes interviews conducted with 453 likely voters (registered voters who said they were likely to vote) in the Old Dominion between October 22 and October 29. The Poll has a margin of error of + 4.6 percent.   

The percentage of undecided voters remains relatively high this late in the campaign-12% of likely voters have yet to make a decision in the contest for Senator.  Allen leads by an almost 2-1 margin among those who were undecided but said they were leaning toward voting for a particular candidate (27%-14%), but 59% refused to budge from their indecision.  Supporters of both candidates are confident in their decision as nearly equal numbers say they are very certain that they will vote for Webb (81%) or Allen (79%).  

With regard to the marriage amendment, 53% said they would vote for it, 36% said they would vote against it, and 11% were undecided.  Not surprisingly, the marriage amendment tracks closely with candidate support.  Three-fourths (76%) of Allen supporters also support the amendment, while 59% of Webb supporters oppose it.

Allen enjoys a higher "favorable" rating than Webb (47%-38%), but he also has a higher "unfavorable" rating (34%-27%).  There is still 16% of the likely electorate that has not formed an opinion of Jim Webb.    

Who is Undecided?  

More than three-fourths (78%) of those who are undecided said they are very likely to vote.  Almost two-thirds (64%) of them are women.  Over half (55%) are moderates, but conservatives outnumber liberals (29%-10%).   A plurality of the undecided voters are independent (47%), but there are more undecided Republicans (28%) than Democrats (9%).  They also tend to be older and have lower income levels.  Although the numbers were small, the most important issues for these voters are the War on Terror, gay marriage and education.    

Sources of Support  

Charges that Webb holds negative views toward women may be having an impact in the campaign in so far as there is no gender gap.  Allen holds a two-point lead among men and a three-point edge among women.  Webb leads solidly among both liberals (74%-16%) and moderates (57%-30%), but Allen leads among conservatives (76%-15%), and there are almost three times as many self-described conservatives as liberals.   Both candidates hold serve in their own party, with Webb taking 88% of the Democrats and Allen 88% of the Republicans.  Webb holds a slim two-point advantage among Independents, while Allen holds a 20-point lead among the relatively small number who thinks of themselves as members of some other party.  

Turning conventional wisdom on its head, Allen leads among those with less education, lower incomes and younger voters, while Webb is ahead among more affluent, older and better-educated voters.  The regional breakdown follows tradition, with Allen doing well in Southwest Va., the Shenandoah Valley, and Tidewater, while Webb fares better in Northern Va..  Central Va. and Southside are almost evenly split between the two candidates.  Allen leads among Whites (49%-40%), while Webb leads among African-Americans (71%-11%).

Issues in the Campaign

The War in Iraq (23%), character/trust (11%), taxes (9%), the War on Terror (9%), gay marriage (9%), and the economy and jobs (7%) were identified as the most important issues in the Senate race.  Those who named Iraq as most important support Webb (64%-29%), while those who mentioned the War on Terror favor Allen by 58%-27%.  Taxes (62-29%) and gay marriage (68%-18%) work to the benefit of Allen, while character/honesty (61%-39%) and the economy and jobs (58%-42%) cut in favor of Webb.  Corruption/scandal was most important to only five respondents, all of whom said they would vote for Webb.   

With regard to Iraq, most likely voters do not think the ultimate outcome will be worth the cost in American lives, although the margin is not great.  A plurality of respondents thinks that the U.S. should withdraw troops from Iraq only after the situation is stabilized.  Fewer think there should be a one-year timetable for withdrawal, and less than 20% think troops should come home immediately.  (Slightly different questions were asked of different respondents so exact percentages are not easily expressed.)  To reduce the chances of a terrorist attack in the U.S., a slim plurality prefers working with Middle East nations to improve economic conditions there (38%) over fighting terrorists abroad (37%) or withdrawing from Iraq and minimizing the U.S. presence in the Middle East (12%); 13% answered none of the above.  As would be expected, strong majorities of those who think the War in Iraq will be worth the cost, support fighting terrorists abroad, and think U.S. troops should remain until Iraq is stable support Allen.  Those who favor withdrawal and think the war is not worth the cost tend to support Webb.  It also is interesting that Webb supporters named Iraq as the most important issues while Allen supporters named the War on Terror (the question was open-ended and responses were coded separately).  The public clearly understands the candidates' differing views on the war.     

Kaine and Bush  

Governor Tim Kaine is very popular in the Commonwealth, with an overall approval rating of 67%.  At the same time, President George W. Bush's overall approval rating is 38%, unchanged from a Roanoke College poll one year ago.  Webb does significantly better than Allen (55%-32%) among those who approve of the job Kaine is doing as Governor.  Allen fares even better among those who approve of Bush (81%-5%).     

Analysis

"These results suggest that this seat is still up for grabs," said Dr. Harry Wilson, the director of the Center for Community Research. 

"There is good news for both camps in these results.  For Allen, he does well with younger voters, who may be underrepresented in this survey, while Webb does very well among the most affluent voters, who may be overrepresented.  Also important for Allen is that he is close to the "magical" 50% approval rating for incumbents.  That number is not likely to move much before Election Day, and incumbents with 50% approval ratings rarely lose.  For Webb, African-Americans may be underrepresented here, and they strongly favor him.  In short, this election is likely to be decided by turnout and who can get their supporters to vote.  

"Any poll attempts to get a sample of the electorate that mirrors those who will vote on Election Day.  That is difficult in any election, but it becomes even more important in elections that are this closely contested."        

Methodology  

Interviewing for The Roanoke College Poll was conducted by The Center for Community Research at Roanoke College in Salem, Va. between October 22 and October 29, 2006.  The sample consisted of 453 likely voters in Virginia.  The sample of phone numbers was prepared by Survey Sampling Inc. of Fairfield, Conn. and was created so that all residential telephone numbers, including unlisted numbers, had a known chance of inclusion.   

Questions answered by the entire sample of 453 registered voters are subject to a sampling error of plus or minus approximately 4.6 percent at the 95 percent level of confidence.  This means that in 95 out of 100 samples like the one used here, the results obtained should be no more than 4.6 percentage points above or below the figure that would be obtained by interviewing all registered voters who have a telephone.  Where the results of subgroups are reported, the sampling error is higher.  

A copy of the questionnaire will be provided upon request.

For more about the Institute for Public Opinion Research, click here. 

CONTACT: Dr. Harry L. Wilson, Director, IPOR
(540) 375-2415 (Office)
wilson@roanoke.edu