An internal poll conducted in Illinois’ 13th District,… read more
National Survey Highlights Dem Problems in Key Seats
Last week, I conducted a national survey for American Action Network. The two key political environment questions show the significant opportunity Republicans have this year. While the overall numbers on the President’s approval rating and the generic ballot are good news, one key crosstab highlights the wave that is building.
The generic ballot shows Republicans leading 44%-39%. Besides all of the usual regional crosstabs, we also broke it out by the type of district. We looked at the sample in the 66 Democratic INCUMBENT districts that Charlie Cook lists as either toss-up or leaning Democratic at the time of the survey. In that key crosstab of Swing Democratic Incumbent Seats, the Republican lead grows to 49%-31% on the generic ballot. That is a very powerful crosstab that says the wave is coming.
Among the remaining Democratic districts (Likely/Safe Dem, and open seats), the generic ballot is an unsurprisingly 33% GOP/51% Dem — a sign that the historically safe Dem seat will remain so, while the swing seats will be a bloodbath. By the way, all of in the GOP held seats, the generic is the reverse of the base Dem seats — 52% GOP/32% Dem. Very few, if any, Republican incumbents will be defeated.
Likewise, President Obama’s numbers with likely voters are similar to the national average — 46% approve/51% disapprove. However, in the Swing Democratic Incumbent Seats. he has a much worse 40% approve/57% disapprove. (Keep in mind, many of these Swing Seats are held by Democrats despite the fact that John McCain either won the district in 2008, or, even if losing, outperformed his national result.
Other key findings on the generic ballot:
- Regionally, the Republican wins 47%-39% in the South, 47%-35% in the Midwest, and 46%-36% in the West, while trailing 36%-47% in the Northeast. The Midwest is going to be a killing field for Democrats this year — from western PA through to the Plains, Republicans are going to sweep a LOT of Democrats right out of office.
- The GOPer leads 51%-32% among whites (that’s 61% of the decided white vote). In 1994, GOPers won the white vote 57%-41%. In the nadir year of 2006, Republicans only took 51% of the white vote.
- African Americans break 8% GOP/86% Dem, so nearly all of the core African American support for Dems is built into this survey already. Only 4% of the undecideds are African American, compared to 78% being white (leaving Hispanics as a key group).
- There is a 17 point gender gap — GOPers lead 49%-35% among men, and trail 40%-43% among women.
- The real GOP problem with women comes among women with college degrees — we trail 33%-52%, while leading 45%-36% among women without degrees. Men with college degrees break 53%-32%, so presumably there are going to be a large number of households where votes get cancelled out.
- White women vote GOP 45%-35%, which is not as strong as white men (57%-28%). However, this underscores that Democrats in swing seats with a low percentage of minorities are in HUGE trouble.
- Suburban and rural voters are both going GOP by double digits – suburban by a 48%-36% margin and rural voters break 51%-30%.
- In the states Obama won with less than 54% of the vote, the generic is breaking 48% GOP/34% Dem. Those states are Colorado, Virginia, North Carolina, Iowa, Ohio, Minnesota, Indiana, and New Hampshire — where, not coincidentally, there are a lot of targeted House races (including many that have just bubbled up to target lists in the last few months as voters seek to punish the party in power).
So, what does all these data mean? Democrats are in trouble nationally, but their problems are amplified in the very areas they need to do well enough to survive. While Republicans are hitting home runs nationally (for what that is worth), in the swing seats, the key states, and the key demographics, GOPers are performing like Sammy Sosa on steriods.
The survey was conducted September 14-16, 2010 among 800 likely voters nationwide, and has a margin of error of +/-3.46% in 95 out of 100 cases.