2010 Election Predictions – How we rank individual races

As the Labor Day weekend approaches, primary season is in the “home stretch” now– the eight remaining Congressional primaries will conclude by September 18 (Louisiana’s “everything else” primary will be on Saturday, October 2). For the past several months, we based our calls on individual races on a composite of polls conducted over the last 28 days. Starting with this analysis, we will compress this “look back period” to 14 days, since voters are starting to tune into the fall elections now. Our rankings are based on the following criteria:

(1) Safe Democratic (dark blue on the map)/Safe Republican (dark red on the map) – either a candidate leads by 10 or more points, or a candidate has over 50% in polls;

(2) Lean Democratic (light blue on the map)/Lean Republican (light red on the map) – a candidate leads by 3-9 points;

(3) Tossup (yellow) – a candidate leads by less than 3 points;

(4) No data available/primary has not been held (gray) – In general, we don’t analyze statewide polls until its primary has been conducted. However, since pollsters have been releasing multiple polls with different Democratic vs Republican matchups for states that have not held their primaries, we will modify this rule as follows: if at least one of the major parties has a clear front runner in a contested primary, we will use the various poll matchups to get a sense of how a Senate/Governor race is leaning and, based on that composite of polls in the last 14 days, will assess which party is ahead.
(5) No Senate/gubernatorial race in 2010 for this state (black);

For the past three weeks, the following has changed with regards to poll numbers released: (1) Democratic polling firm Public Policy Polling (PPP) is now using a “likely voter” model for determining its poll sample – this theoretically will produce a more Republican poll sample, and (2) Rasmussen is now using follow up question(s) to get initially undecided voters to indicate a preference for one of the candidates (this action will cause the undecided number to plummet). We will report both numbers Rasmussen uses for purposes of our poll analysis.

2010 Election Predictions – Senate Races 

Senate Projections

(9/1 projection: 54 Democrats, 46 Republicans, or Republican gain of +5)
 (8/19 and 8/26 projections: 55 Democrats, 45 Republicans, or Republican gain of +4)

Five things on the map are worth noting: (1) We do not yet have poll data for the Hawaii and Vermont Senate races. Vermont recently conducted its primary, while Hawaii will not have its primary until September 18; (2) The Alaska race has potentially gotten more complicated – most of the absentees were counted on Tuesday, and incumbent Senator Lisa Murkowski was unable to narrow her 1,668 vote deficit. There have been some polls assuming that the Republican challenger Joe Miller would be the nominee, and those currently show him with a respectable lead in the polls. Now that Senator Murkowski is out of the picture, we can get a clearer picture on the race; (3) Though Florida is currently classified as a “Tossup”, for the past three polls, Republican Marco Rubio has outpolled Republican turned Independent Charlie Crist. We expect this trend to continue, because with the Democratic nominee chosen, we can expect further desertions of Democrats from the Crist camp; (4) Now that party nominees have been chosen for the special Senate election in West Virginia, the first post primary poll released shows Democratic Governor Joe Manchin with only a 48-42% lead. This was the first poll showing him with less than 50%, and we will watch to see whether this can become a race to watch  (like the Connecticut Senate race); (5) in a recent Senate poll showing Vitter with a substantial lead, Senator Vitter is clearly helped by the fact that 58% of voters favor repeal of healthcare reform, and that only 38% approve of President Obama’s performance.

2010 Election Predictions – Governor’s Races  

Governor Projections

(9/1 projection: 35 Republicans, 15 Democrats, or Republican gain of +11)
(8/26 projection: 31 Republicans, 19 Democrats, or Republican gain of +7)
(8/19  projection: 32 Republicans, 18 Democrats, or Republican gain of +8) 

Since our last analysis, there have been four changes. For the first time in the California and New Mexico Governor’s races, the Republican leads by a significant enough margin for us to reclassify these races from “tossup” to “leans Republican.” In Ohio, the Republican challenger’s poll leads have grown large enough for us to reclassify this race from “leans Republican” to “solid Republican.” And finally, a recent poll by Minnesota Public Radio puts the Minnesota Governor’s race  at a tie – the poll even notes that the third party candidate is drawing support more from Democrats.

A couple of things on the map are also worth noting: (1) while we have no poll data for Vermont, they just held their primary, so we expect to see poll data very soon; (2) In three  states, Democratic leads in the polls are due to third party candidacies. In Colorado, the Republican and Constitution Party candidates are getting a combined 43% (as opposed to the Democrat ‘s 41%). In Massachusetts, the embattled Democratic incumbent leads in the polls with an anemic 30%, while the Republican + Independent total is 41%. Finally, in Rhode Island, the Democrat “leads” with 39%, while the Republican + Independent total is 52%; (3) In the Florida Governor’s race, a third party candidate has dropped out; (4) In Colorado, while the Democrat leads due to a third party candidacy, his leads have begun to soften. We will keep an eye out to see whether we need to reclass this race in a subsequent analysis.

2010 Election Predictions – Generic Congressional Vote

(9/1 poll composite: 46.6% Republican, 41.8% Democrat – +4.9% Republican)
(8/26 poll composite: 46.7% Republican, 41.0% Democrat – +5.7% Republican)
(8/19 poll composite: 46.6% Republican, 41.2% Democrat – +5.4% Republican)

These strengthening Republican poll leads have ignited a wave of panic among Democrats. The chairman of the Democratic Congressional Campaign committee has now communicated  that if a Democrat is not doing well in the polls and/or is not raising enough money, they will not get any help from the campaign committee – essentially, a financial “death penalty” for underperformers.

To put the “Generic Congressional Vote” question in perspective, when Republicans retook the House (and Senate) in 1994, the popular vote in the 435 House districts was 52-45% Republican. If we were to split the undecideds equally between both parties, you would have a 52.5% Republican vote (we’re including the Newsweek number in this analysis), which equates to a 228 seat Republican majority – a net gain of 49 Republicans, or 10 more than needed to regain control of the House. We also believe at this time that the 228 seat figure is a conservative estimate, since it doesn’t consider voter intensity, which clearly favors the Republicans right now. There’s another factor at play too: voters are not happy with Congressional incumbents, and in a recent Rasmussen poll, only 37% wanted to re-elect their own Congressman.  Since 60% of incumbent House members are Democrats, this is obviously not a sentiment the Democrats want voters to have.

2010 Election Predictions – Individual House Races

(9/1: 250 Republicans/185 Democrats (+ 71 Republican, “Watch list” of 28 Democrats/2 Republicans)

(8/26: 253 Republicans/182 Democrats (+ 74 Republican, “Watch list” of 25 Democrats/2 Republicans)

(8/19: 254 Republicans/181 Democrats (+ 75 Republican, “Watch list” of 24 Democrats/1 Republican)

One of the things we have repeatedly noted in our discussions of the “Obama plunge” and its effect on the fall elections is that actual poll results (and in the case of Washington state, “open primary” election results) take precedence over the theoretical criteria used to determine whether a seat is vulnerable.  This past week, we have seen polls released for 19 House races. Though the numbers above would seem to signify a continued Democratic renaissance, the truth is actually the opposite. What we saw was the following:

(1) In 12 of the 19 House races, polling data confirmed our belief as to who was safe, on the “watch list”, or vulnerable;

(2) Two formerly safe Democrats (4 year incumbent Joe Donnelly of Indiana and 10 year incumbent Mike Ross of Arkansas) are now on the “watch list.” While Mike Ross’ district voted 60% for McCain, Obama actually carried Donnelly’s district with 54% With the exception of “cap and trade”, Rep. Donnelly has supported the Democratic leadership, while Mike Ross has been more independent in his voting record;

(3) Three freshman Democrats have been moved from “vulnerable” to “watch list”: Alan Grayson of Florida, and New Mexico Democrats Harry Teague and Martin Heinrich. While McCain narrowly carried Teague’s district, Obama got 52% in Grayson’s district, and 60% in Heinrich’s district;

(4) Two Democrats (Russ Carnahan of Missouri and Heath Shuler of North Carolina) have seen their numbers rise above 50% in recent polls and thus have been moved off the “watch list” for now.

A final note on the polls so far: we have cumulatively seen polls done on 34 Democratic held seats where the Democrat is behind in the polls. On average, these districts voted 51% for Obama, which means that these seats are “Ground Zero” for the “Obama plunge.” Similarly, we have seen poll results for an additional 23 Democratic held seats where Democrats lead but with less than 50% – Obama on average got 52% in these districts, and these seats are similarly vulnerable against the “Obama plunge.” These 57 (last week, it was 51) House seats alone can flip the House to Republican control. Additionally, there are 42 more vulnerable/”watch list” Democratic held seats where no polling has been done.