Thursday, May 24, 2012

Taking A Deep Dive In Virginia

February 10, 2012 | 6:00 a.m.

Behind the paywall, we're taking an in-depth look at the races that will decide who controls the Senate next year, based on exit poll data from 2006, 2008 and 2010 -- including unreported breakdowns specially provided to Hotline by Edison Research.

The analysis found, as I wrote in National Journal last week, that two divergent forces are driving the race for the upper chamber: On one hand, Democrats will benefit because of higher minority turnout driven by the presidential race (senators running for re-election this year didn't have the presidential race at the top of the ticket in 2006). On the other hand, Republicans will benefit because white voters -- particularly those with college educations -- broke hard away from Democrats in 2010, and they still have a largely unfavorable view of President Obama.

Infographic

Virginia, where former Democratic Gov. Tim Kaine and former GOP Sen. George Allen are locked in an expensive and nail-biting battle for a Senate seat, provides an example of the former more than the latter. And that could mean the conventional wisdom -- that the race will go down to the wire -- is wrong: Virginia's changing demographics mean Kaine's odds look a lot better than they appear.

That's thanks almost completely to the surge of minority voters who turned out to elect Obama. In 2006, minorities made up 21 percent of the electorate, according to Edison Research data. By 2008, when Obama's campaign poured millions into the state to turn out new or irregular voters, minorities made up 30 percent of the electorate -- a 9-point increase. Only Texas saw a larger increase in the influence of minority voters between 2006 and 2008.

Of course, 2012 is not 2008, and neither Democrats nor Republicans believe minority turnout will be as high as 30 percent of the electorate this year. But it could be close, and Allen's dismal performance among minority voters in 2006 should worry the GOP.

Virginia also showed a big education gap, a divide among white voters who graduated from college and those who didn't. In 2006, Allen beat Democrat Jim Webb by a small 53 percent to 47 percent margin among college-educated white voters. Allen stomped Webb among white voters without a college education by a much wider 66 percent to 34 percent margin.

A Quinnipiac University poll released this week shows largely the same story: Allen and Kaine are running close among college-educated white voters (Kaine leads by a statistically insignificant 47 percent to 45 percent), Allen is winning by 30 points (59 percent to 29 percent) among non-college whites, and Kaine is leading among non-white voters by an overwhelming margin (68 percent to 18 percent).

Exit Polls

-      ---- Whites ----
Year   College  Non-Col Minorities
'06      46%      33%      21%
'08      40       30       30
+/-      -6       -3       +9

'06 Election Results

        ---- Whites ----
       College  Non-Col Minorities
Webb     47%      34%      77%
Allen    53       66       23

The Quinnipiac University poll, conducted Feb. 1-6 among 1,544 registered voters for a margin of error of +/- 2.5 percent (margins of error for subgroups is higher).

-                                   -White-  Non 
-           All Dem GOP Ind Men Wom Col Non White
T. Kaine    45% 90%  5% 44% 42% 47% 47% 29%  68%
G. Allen    44   4  89  40  49  39  45  59   18
Other        1  --  --   1   1  --  --   1    1
Not vote     2   1   1   4   3   1   1   2    3
Undec        9   5   5  12   6  11   7   9   10

But play with the poll's findings a bit and things look a lot better for Kaine than the statistical tie Quinnipiac pollsters reported: Reweight the poll to assume an electorate that looks more like the non-presidential 2006 (46 percent college whites, 33 percent non-college whites and 21 percent minorities) and Kaine would lead Allen 45 percent to 44 percent, a statistically insignificant margin.

If the electorate looks like it did in 2008 (40 percent college whites, 30 percent non-college whites, 30 percent minorities), Kaine's lead grows to 48 percent to 41 percent, a higher margin than any public survey has shown.

Take these numbers with a grain of salt: Republicans point out that Quinnipiac pollsters surveyed registered voters, resulting in a sample they say skews more toward Democrats than they anticipate the actual electorate will be.

But it underscores a key point about the new way to win a statewide election in Virginia: Because of the growing role minorities play in the state, and their predisposition toward voting for Democrats, Republicans are forced to win larger percentages of the white vote. This time around, Republicans who study the state believe they'll need somewhere between 63 percent and 67 percent of the white vote to win the Senate seat.

Allen's margin among non-college whites will almost certainly rise from 59 percent. His big challenge going forward will be to build a lead among the college-educated portion of the white vote. His failure to do so in 2006 cost him the race against Webb; his 2012 bid hinges on changing that.

Subscribers, stay tuned to Hotline over the course of the next few weeks as we dive deep into each Senate race's electoral breakdowns. The results will shed light on which side really has the advantage this fall, and long into the future.

Join the Discussion

The National Journal Group has the right (but not the obligation) to monitor the comments and to remove any materials it deems inappropriate.

Comments powered by Disqus

 

Search This Blog


Archives

Monthly Archives

Recent Posts

Recent Comments


About

Contact On Call:


Staff

Reid Wilson, Editor-in-Chief
Sean Sullivan, Editor
Julie Sobel, Deputy Editor

Contributing Editors:
Josh Kraushaar and Quinn McCord
Contributing Writers:
Steven Shepard, Dan Roem, Tim Alberta, Stephanie Palla, Sarah Mimms, Kevin Brennan, Chris Peleo-Lazar and Scott Bland



Disclaimer

On Call editors reserve the right to delete inappropriate comments. The Hotline, National Journal Group, Inc. and Atlantic Media Company are not responsible for the content of the comments that remain.