Polls vary widely in race for VA governor

THE ROANOKE COLLEGE POLL

Conducted by The Institute for Policy and Opinion Research

McAuliffe widens lead over Cuccinelli; Democrats also lead down-ticket races

Democrat Terry McAuliffe has opened a 15-point lead over Republican Ken Cuccinelli (46%-31%), while 14 percent of likely voters in Virginia remain undecided in the 2013 Gubernatorial election, according to The Roanoke College Poll. Libertarian candidate Robert Sarvis claimed 9 percent of respondents.*

The Roanoke College Poll interviewed 838 likely voters in Virginia between October 21 and October 27 and has a margin of error of +3.4 percent.

In the down-ticket races, Democrat Ralph Northam leads Republican E. W. Jackson for lieutenant governor (Northam—48%, Jackson—32%, Undecided—20%), and, for the first time, Democrat Mark Herring has opened a significant lead over Republican Mark Obenshain for attorney general (Herring—46%, Obenshain—35%, Undecided—19%).

The Virginia Race for Governor—2013

Cuccinelli remains very unpopular with voters, while McAuliffe is barely in favorable territory (Cuccinelli—28% favorable, 52% unfavorable; McAuliffe—36% favorable, 33% unfavorable). Robert Sarvis is still largely unknown to most voters (16% favorable, 13% unfavorable, 55% don’t know enough to have an opinion).

The election is shaping up as a referendum on Cuccinelli. More than half (58%) of his supporters said they are voting for Cuccinelli while a plurality (44%) of McAuliffe voters said they are voting against Cuccinelli.

When asked which candidate would do a better job on a variety of issues, most responses track with voter preference fairly closely. McAuliffe was judged to do a better job on the economy and jobs (44%-33%), health care (44%-31%), education (42%-29%), transportation (40%-26%), abortion (44%-27%), energy (42%-27%) and the environment (46%-25%). He held a narrower edge over Cuccinelli on gun control (40%-33%) and taxes (37%-35%). Cuccinelli’s only advantage was on government spending (36%-33%).

Regarding other characteristics, McAuliffe led on being the most trustworthy (36%-25% with Sarvis registering his highest number at 16%), caring the most about “people like you” (38%-25%), and understanding the problems and challenges facing the Commonwealth (39%-30%). As for who is the most intelligent candidate, McAuliffe was named by 27 percent, while Cuccinelli was cited by 24 percent, and 10 percent thought Sarvis was the most intelligent.

As expected, the most important issues in the campaign were the economy (28%), jobs (18%), and health care (10%). Other issue named by 5 percent or more of respondents were government spending/budget, education, and abortion/women’s issues (5% each).

The largest change since the last Roanoke College Poll was a surge for McAuliffe among male likely voters. In fact, the gender gap all but disappeared in this poll (females—46%-29% McAuliffe; males 46%-32% McAuliffe). McAuliffe continued to do extremely well among blacks, and of, course, among Democrats and liberals. Cuccinelli fares better with Republicans, conservatives, and those who support the Tea Party. His lead among white likely voters has evaporated, and he now trails by 2 percent in that group. In what are typically considered to be crucial constituencies, McAuliffe has significantly increased his leads among moderates (54%-19%) and Independents (45%-29%). The results for specific subgroups may be found here.

Sources of campaign information

This may be an age of technological innovation in political campaigning, but this election is more old-school. Less than 10 percent of likely voters have been following the candidates on any social media (4% each for Cuccinelli and McAuliffe, 2% for Sarvis). And barely 20 percent have visited one or more candidates’ websites (15% for Cuccinelli, 14% for McAuliffe, and 5% for Sarvis). Regarding sources of political news, 81 percent of respondents identified television news as important for them, followed by the Internet (64%), newspapers (62%), friends and family (55%), and radio news (53%). Just over one-fourth (27%) of respondents said they had watched a gubernatorial debate. A plurality (40%) of that group said they thought that McAuliffe had won the debate.

Elected officials’ approval ratings and favorable/unfavorable views

President Barack Obama’s favorable rating is 50 percent (42% unfavorable), which is somewhat higher than earlier in October. Gov. Bob McDonnell’s rating is slightly lower at 40 percent, while his unfavorable rating has risen to 33 percent.

Obama’s job approval rating is slightly improved from earlier this month, with 43 percent of respondents approving and 45 percent disapproving of the job he is doing. McDonnell’s approval rating is marginally lower at 41 percent, while Congressional approval must soon hit the floor if this is not the floor (4%, rounded up from the actual 3.5%).

A preference for divided government and compromise

A majority of Virginia’s likely voters (52%) expressed a preference for divided government. At the same time, a larger majority (78%) thought that elected officials should compromise in order to get things done rather than sticking to their principles if it means the government would not function smoothly. (Of course, this poll was conducted in the shadow of the federal government shutdown.) A majority (58%) also said they would vote for an elected official who compromised on something they believed was a core value or principle. Yet, when asked if elected officials should act as trustees who make decisions they believe are correct and necessary, a majority (61%) said that elected officials should represent the views of their constituents even if they disagree with those views.

Analysis

“Terry McAuliffe continues to widen his lead and we may now be seeing some coattails as well, particularly in the attorney general’s race,” said Dr. Harry Wilson, director of the Institute for Policy and Opinion Research. “Ken Cuccinelli is still the driving force in this election, both for his supporters and his detractors.”

“At this point, Cuccinelli has to hope that voting for him is a much stronger motivator than voting against him,” Wilson said. “McAuliffe’s comparatively lower ratings on honesty and intelligence do not suggest that voters are enamored of him. As many voters anecdotally told our interviewers, he is the ‘lesser of the evils.’”

*Likely voters were defined as registered voters who said they were very likely or somewhat likely to vote in November. All references to likely voters in this release, other than this paragraph, employ that definition. Using a more strict voter model—including only those registered voters who said they were very likely to vote and voted in 2012–the percentages barely change (McAuliffe—48%, Cuccinelli—33%, Sarvis—8%; N=670). Including likely voters and adding in those who said they are leaning toward a candidate maintains McAuliffe’s lead over Cuccinelli (48%-33%), while Sarvis rises to 10 percent, and only 8 percent remain undecided.

Methodology

Interviewing for The Roanoke College Poll was conducted by The Institute for Policy and Opinion Research at Roanoke College in Salem, Va. between October 21 and October 27, 2013. A total of 838 likely voters in Virginia were interviewed. The sample included both land lines and cell phones and was created so that all cell phone and residential telephone numbers, including unlisted numbers, had a known chance of inclusion. Cell phones constituted 32 percent of the completed interviews.

Questions answered by the entire sample of 838 likely voters are subject to a sampling error of plus or minus approximately 3.4 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 3.4 percentage points above or below the figure that would be obtained by interviewing all likely voters in Virginia who have a home telephone or a cell phone. Where the results of subgroups are reported, the sampling error is higher.

Quotas were used to ensure that different regions of the Commonwealth were proportionately represented. The data were statistically weighted for gender, race, and age.