Last Post

I am going to call it a night. I think that Republicans are clearly going to have the upper hand--procedurally. They will have an ample majority in the House and the Democrats will not have control in the Senate.

Senate
Outcome So-far: +6 Republican
Projected Outcome: +7 Republican, 52 Democrats, 48 Republicans

House
Outcome So-far: +59 Republican
Projected Outcome: +64 Republican, 243 Republicans, 192 Democrats

Republicans Expected to Gain 60 House Seats

Senate
Outcome So-far: +3 Republican
Projected Outcome: +7 Republican, 52 Democrats, 48 Republicans

House
Outcome So-far: +21 Republican
Projected Outcome: +60 Republican, 239 Republicans, 196 Democrats

A Note to Readers

On election night, I will be providing live updates of the results. I will update an election board, which will show the status of the most vulnerable House and Senate seats, along with a running probability of a Republican takeover in each chamber. Please visit my site so that you can all receive this information in realtime.

-Jacob

FINAL List of Vulnerable House Seats

In the final analysis of individual U.S. House seats, 89 Democrats and six Republicans are now listed as vulnerable. In the past week, Chellie Pingree (ME-1) was added to the list of vulnerable Democratic Representatives.

As the reader may notice, the Republicans are likely to retake control of the U.S. House of Representatives. The Democrats will probably win four or five Republican seats. However, the Republicans have at least a 90 percent probability of winning each of 41 Democratic seats. They have at least a 50 percent probability of winning each of the next 22 most vulnerable seats. They even have over a 10 percent probability of winning each of the 64th through 82nd most vulnerable seats. This analysis implies that the Republicans are expected gain approximately 41 + 1/2*41 = 62 seats and lose about four seats. Thus, the Republicans are poised to net about 58 seats, while they only need to gain 39 seats to take control of the House.

Correction: Note that WV-3 is not open.

Democrats by State (Open seats in bold)


Democrats by Vulnerability (Open seats in bold)


Republicans (Open seats in bold)

FINAL House Prediction: Republicans Gain 52

Based on polling data released over the last nine months, Republicans are expected to gain between 41 and 62 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 52 seat gain by Republicans, one more seat than the previous estimate. The Republicans seem to be poised to retake control of the U.S. House of Representatives with the largest electoral landslide since 1948. In fact, based on a Monte Carlo simulation, the Republicans have a 97 percent probability of retaking the House.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 3.9 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is 0.6 percent greater than the mid-October estimate. The final estimate implies that, if the election were held today, Republicans would achieve approximately a 27 seat majority over the Democrats--a historically modest, but commanding advantage.

The charts below provide the change in voter preferences over time, along with the probability distribution of the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

FINAL Senate Prediction: Republicans Gain Seven

Safe Democratic, 8 Seats (15 percent+) :
CT (Dodd)
DE (Kaufman)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Wyden (OR)
Leahy (VT)

Strong Democratic, 2 Seats (5.0 -15.0 percent):
Boxer (CA)
Murray (WA)

Weak Democratic, 2 Seats (0.0 - 5.0 percent):
IL (Burris)
WV (Byrd)

Weak Republican, 4 Seats (0.0 - 5.0 percent):
Bennet (CO)
NH (Gregg)
Reid (NV)
PA (Spector)

Strong Republican 5 Seats (5.0 - 15 percent):
Grassley (IA)
KY (Bunning)
OH (Voinovich)
MO (Bond)
Feingold (WI)

Safe Republican 16 Seats (15.0 percent+):
AK (Murkowski)
Shelby (AL)
McCain (AZ)
Lincoln (AR)
FL (LeMieux)
Crapo (ID)
IN (Bayh)
Isakson (GA)
KS (Brownback)
Vitter (LA)
ND (Dorgan)
Burr (NC)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

Most Vulnerable House Seats: Situation Continues to Worsen for Democrats

In the latest analysis of individual House Seats, seven additional Democrats are now listed as vulnerable. The seven Democrats are House veterans Sanford Bishop Jr. (GA), Leonard Boswell (IA), Bruce Braley (IA), Bob Etheridge (NC), James Oberstar (MN), Nick Rahall (WV), and Ike Skelton (MO). Hence, 88 Democrats are listed as vulnerable, while only six Republicans are at risk of losing.

As the reader may notice, the Republicans are likely to retake control of the U.S. House of Representatives. The Democrats will probably win four or five Republican seats. However, the Republicans have at least a 90 percent probability of winning each of 37 Democratic seats. They have at least a 50 percent probability of winning each of the next 26 most vulnerable seats. They even have over a 10 percent probability of winning each of the 64th through 82nd most vulnerable seats. This analysis implies that the Republicans are expected gain approximately 37 + 1/2*44 = 59 seats and lose about four seats. Thus, the Republicans are poised to net about 55 seats, while they only need to gain 39 seats to take control of the House.

Democrats by State (Bold indicates open seat)


Democrats by Rank (Bold indicates open seat)


Republicans (Bold indicates open seat)

HOUSE UPDATE: Republicans Remain Poised to Win 50+ Seats

Based on polling data released over the last nine months, Republicans are expected to gain between 39 and 63 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 51 seat gain by Republicans, the same as the previous estimate. If the polling holds steady, Republicans seem to be poised to retake control of the U.S. House of Representatives with the largest electoral landslide since 1994. In fact, based on a Monte Carlo simulation, the Republicans have a 95 percent probability of retaking the House.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 3.3 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is identical to the September estimate. The mid-October estimate implies that, if the election were held today, Republicans would achieve approximately a 25 seat majority over the Democrats--a modest, but decisive advantage.

Next week, I will publish a final estimate for the mid-term election. Per usual, the charts below provide the change in voter preferences over time, along with the probability distribution of the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

SENATE UPDATE: Republicans Predicted to Gain Six Seats

Safe Democratic, 7 Seats (15 percent+) :
CT (Dodd)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Wyden (OR)
Leahy (VT)

Strong Democratic, 4 Seats (5.0 -15.0 percent):
Boxer (CA)
DE (Kaufman)
IL (Burris)
Murray (WA)

Weak Democratic, 2 Seats (0.0 - 5.0 percent):
Reid (NV)
WV (Byrd)

Weak Republican, 3 Seats (0.0 - 5.0 percent):
Bennet (CO)
NH (Gregg)
PA (Spector)

Strong Republican 4 Seats (5.0 - 15 percent):
Grassley (IA)
OH (Voinovich)
MO (Bond)
Feingold (WI)

Safe Republican 17 Seats (15.0 percent+):
AK (Murkowski)
Shelby (AL)
McCain (AZ)
Lincoln (AR)
FL (LeMieux)
Crapo (ID)
IN (Bayh)
Isakson (GA)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Burr (NC)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

Most Vulnerable House Seats, Part IV

In the tables below, I assign probabilities to Republican gains in the 81 most vulnerable Democratic seats. I also list probabilities to Democratic gains in the six most vulnerable Republican seats.

As the reader may notice, the Republicans are likely to retake control of the U.S. House of Representatives. The Democrats will probably win four or five Republican seats. However, the Republicans have at least a 90 percent probability of winning each of 33 Democratic seats. They have at least a 50 percent probability of winning each of the next 23 most vulnerable seats. They even have over a 10 percent probability of winning each of the 57th through 75th most vulnerable seats. This analysis implies that the Republicans are expected gain approximately 33 + 1/2*42 = 54 seats and lose about four seats. Thus, the Republicans are poised to net about 50 seats, while they only need to gain 39 seats to take control of the House.

Democrats by State (Bold indicates open seat)


Democrats by Rank (Bold indicates open seat)


Republicans (Bold indicates open seat)

Most Vulnerable House Seats, Part III

In the tables below, I assign probabilities to Republican gains in the 81 most vulnerable Democratic seats. I also list probabilities to Democratic gains in the six most vulnerable Republican seats.

As the reader may notice, the Republicans are likely to retake control of the U.S. House of Representatives. The Democrats will probably win four or five Republican seats. However, the Republicans have at least a 90 percent probability of winning each of 30 Democratic seats. They have at least a 50 percent probability of winning each of the next 26 most vulnerable seats. They even have over a 10 percent probability of winning each of the 57th through 75th most vulnerable seats. This analysis implies that the Republicans are expected gain approximately 30 + 1/2*45 = 52 seats and lose about five seats. Thus, the Republicans are poised to net about 47 seats, while they only need to gain 39 seats to take control of the House.


Democrats by State (Bold indicates open seat)



Democrats by Rank (Bold indicates open seat)



Republicans (Bold indicates open seat)

HOUSE UPDATE: Bottom Falls Out of Democratic Party

Based on polling data released over the last nine months, Republicans are expected to gain between 39 and 62 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 51 seat gain by Republicans, up from the previous estimate of 48 seats. If the polling holds steady, Republicans seem to be poised to retake control of the U.S. House of Representatives with the largest electoral landslide since 1994. In fact, based on a Monte Carlo simulation, the Republicans have a 95 percent probability of retaking the House.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 3.3 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is substantially greater than the August estimate of 2.0 percent. The September estimate implies that, if the election were held today, Republicans would achieve approximately a 25 seat majority over the Democrats--a modest, but decisive advantage.

Over the next two months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the probability distribution of the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

Most Vulnerable House Seats, Revisited

Based on a recent critique from Nate Silver, I have decided to conduct a seat-by-seat analysis of the U.S. House. In the tables below, I assign probabilities of Republican gains in the 88 most vulnerable Democratic seats. I also list probabilities of Democratic gains in the 10 most vulnerable Republican seats.

As the reader may notice, the Republicans have a much greater than 65 percent probability--the estimate that Nate Silver asserts--of winning the House. The Democrats are likely to win four or five Republican seats. However, the Republicans have at least a 92 percent probability of winning each of 31 Democratic seats. They have at least a 50 percent probability of winning each of the next 24 most vulnerable seats. They even have over a 10 percent probability of winning each of the 56th through 75th most vulnerable seats. This analysis implies that the Republicans are expected gain approximately 31 + 1/2*44 = 53 seats and lose about five seats. Thus, the Republicans are poised to net about 48 seats, while they only need to gain 39 seats to take control of the House.

What's the probability the the Republicans could only gain 38 or fewer seats? Using a back-of-envelope calculation, I would say that the Democrats only have about a two percent chance of retaining the House. Now, I am assuming independence between the individual seats and zero uncertainty in my seat-specific probability estimates. So, in reality, the Democrats may have as much as a 10 to 15 percent chance of retaining the House (coincidentally, my analysis on the generic ballot question yields about this probability). However, I do not think that the Democrats have anywhere near one-in-three odds of maintaining the majority.


Competitive Democratic Seats


Competitive Republican Seats

SENATE UPDATE: Despite DE, Republicans Still Predicted to Gain Six Seats

Using a similar methodology as my model for the U.S. House, I predict that the Republicans will gain six seats in the U.S. Senate. Here is a seat-by-seat summary (margin of victory in parentheses; party turnover in italics):

Safe Democratic, 7 Seats (15 percent+) :
CT (Dodd)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Wyden (OR)
Leahy (VT)

Strong Democratic, 4 Seats (5.0 -15.0 percent):
Boxer (CA)
DE (Kaufman)
IL (Burris)
Murray (WA)

Weak Democratic, 2 Seats (0.0 - 5.0 percent):
Reid (NV)
WV (Byrd)

Weak Republican, 4 Seats (0.0 - 5.0 percent):
Bennet (CO)
NH (Gregg)
PA (Spector)
Feingold (WI)

Strong Republican 4 Seats (5.0 - 15 percent):
FL (LeMieux)
Grassley (IA)
OH (Voinovich)
MO (Bond)

Safe Republican 16 Seats (15.0 percent+):
AK (Murkowski)
Shelby (AL)
McCain (AZ)
Lincoln (AR)
Crapo (ID)
IN (Bayh)
Isakson (GA)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Burr (NC)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

Most Vulnerable House Seats

Now that Nate Silver has come out with his own seat-by-seat house projection, I have decided to publish my list of most vulnerable Democratic and Republican seats. My methodology is not very sophisticated and I do not assign probabilities to whether a given seat switches party. However, this list should provide the reader an idea of which seats will be the first to switch from Democratic to Republican.

My ranking of the most vulnerable seats is essentially based on the 2008 presidential vote, demographic composition, and ideological orientation of each district. A Partisan Index score of under 30.0 should be interpreted as a solidly Republican district, while a score of over 70.0 indicates that the district is solidly Democratic. Districts between 30.0 and 45.0 are Republican leaning and districts between 55.0 and 70.0 are Democratic leaning. Districts in the 45.0 to 55.0 range indicate marginal status.

I list the one hundred most vulnerable Democratic seats and the twenty-five most vulnerable Republican seats. Open seats are in bold.

HOUSE UPDATE: Republicans Remain in Position to Take Majority

Based on polling data released over the last eight months, Republicans are expected to gain between 35 and 60 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 48 seat gain by Republicans, unchanged from the previous estimate. If the polling holds steady, Republicans seem to be poised to retake control of the U.S. House of Representatives with the largest electoral landslide since 1994. In fact, based on a Monte Carlo simulation, the Republicans have an 85 percent probability of retaking the House.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 2.0 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is slightly greater than the July estimate of 1.9 percent. The August estimate implies that, if the election were held today, Republicans would achieve approximately a 19 seat majority over the Democrats--a modest, but decisive advantage.

Over the next three months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the probability distribution of the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

SENATE UPDATE: Republicans to Gain Six Seats

Using a similar methodology as my model for the U.S. House, I predict that the Republicans will gain six seats in the U.S. Senate. Here is a seat-by-seat summary (margin of victory in parentheses; party turnover in italics):

Safe Democratic, 6 Seats (15 percent+) :
CT (Dodd)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Leahy (VT)

Strong Democratic, 5 Seats (5.0 -15.0 percent):
Boxer (CA)
IL (Burris)
Reid (NV)
Wyden (OR)
WV (Byrd)

Weak Democratic, 2 Seats (0.0 - 5.0 percent):
Murray (WA)
Feingold (WI)

Weak Republican, 6 Seats (0.0 - 5.0 percent):
Bennet (CO)
DE (Kaufman)
FL (LeMieux)
NH (Gregg)
OH (Voinovich)
PA (Spector)

Strong Republican 3 Seats (5.0 - 15 percent):
Grassley (IA)
MO (Bond)
Burr (NC)

Safe Republican 15 Seats (15.0 percent+):
AK (Murkowski)
Shelby (AL)
McCain (AZ)
Lincoln (AR)
Crapo (ID)
IN (Bayh)
Isakson (GA)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

HOUSE UPDATE: Republicans Edge Up to a 48 Seat Gain

Based on polling data released over the last seven months, Republicans are expected to gain between 30 and 65 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 48 seat gain by Republicans, up five seats from the previous estimate. If the polling holds steady, Republicans seem to be poised to retake control of the U.S. House of Representatives with the largest electoral landslide since 1994. In fact, based on a Monte Carlo simulation, the Republicans have an 80 percent probability of retaking the House.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 1.9 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is somewhat greater than the June estimate of 1.4 percent. The July estimate implies that, if the election were held today, Republicans would achieve approximately an 18 seat majority over the Democrats--a modest, but decisive advantage.

Over the next five months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the probability distribution for the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

NOTE: I tweaked my charts for the expected margin of majority and the probability distribution for the number of seats gained by Republicans.

Senate Prediction: Republicans to Gain Five Seats

Using a similar methodology as my model for the U.S. House, I predict that the Republicans will gain five seats in the U.S. Senate. Here is a seat-by-seat summary (margin of victory in parentheses; party turnover in italics):

Safe Democratic, 5 Seats (15 percent+) :
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Leahy (VT)

Strong Democratic, 5 Seats (5.0 -15.0 percent):
Boxer (CA)
CT (Dodd)
IL (Burris)
Reid (NV)
Wyden (OR)

Weak Democratic, 4 Seats (0.0 - 5.0 percent):
FL (LeMieux)
Murray (WA)
WV (Byrd)
Feingold (WI)

Weak Republican, 4 Seats (0.0 - 5.0 percent):
DE (Kaufman)
NH (Gregg)
OH (Voinovich)
PA (Spector)

Strong Republican 4 Seats (5.0 - 15 percent):
Bennet (CO)
Grassley (IA)
MO (Bond)
Burr (NC)

Safe Republican 15 Seats (15.0 percent+):
Shelby (AL)
McCain (AZ)
Lincoln (AR)
Crapo (ID)
IN (Bayh)
Isakson (GA)
Murkowski (AK)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

Polling Remains Steady, Republicans Expected to Gain 43 Seats

Based on polling data released over the six five months, Republicans are expected to gain between 25 and 60 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 43 seat gain by Republicans, down slightly from the previous estimate. This result is almost identical to the previous month's estimate. However, one should note that polling has been sparse for the last several weeks. If the polling holds steady, Republicans seem to be poised to retake control of the U.S. House of Representatives.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 1.4 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is slightly greater than the May estimate of 1.3 percent. The June estimate implies that, if the election were held today, Republicans would achieve a bout a 11 seat majority over Democrats--a very modest advantage.

Over the next five months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the probability distribution for the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

Senate Prediction: Republicans to Gain Four Seats

Using a similar methodology as my model for the U.S. House, I predict that the Republicans will gain four seats in the U.S. Senate. Here is a seat-by-seat summary (margin of victory in parentheses; party turnover in italics):

Safe Democratic, 6 Seats (15 percent+) :
Boxer (CA)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Leahy (VT)

Strong Democratic, 6 Seats (5.0 -15.0 percent):
CT (Dodd)
IL (Burris)
Reid (NV)
Wyden (OR)
Murray (WA)
Feingold (WI)

Weak Democratic, 2 Seats (0.0 - 5.0 percent):
PA (Spector)
FL (LeMieux)

Weak Republican, 3 Seats (0.0 - 5.0 percent):
Bennet (CO)
DE (Kaufman)
NH (Gregg)

Strong Republican 6 Seats (5.0 - 15 percent):
McCain (AZ)
Grassley (IA)
IN (Bayh)
MO (Bond)
Burr (NC)
OH (Voinovich)


Safe Republican 13 Seats (15.0 percent+):
Shelby (AL)
Lincoln (AR)
Crapo (ID)
Isakson (GA)
Murkowski (AK)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

Republicans Are Expected to Gain 42 Seats

Based on polling data released over the last five months, Republicans are expected to gain between 25 and 60 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 42 seat gain by Republicans, down slightly from the previous estimate. Such an electoral accomplishment would be sufficient to unseat the Democratic Party from majority status in the lower chamber. The Republican gains would also represent the largest electoral landslide since 1994.

A statistical model that analyzes the results of 15 independent polling firms indicates that Republicans lead Democrats by a margin of 1.3 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is slightly smaller compared to the April estimate of 1.8 percent. The May estimate implies that, if the election were held today, Republicans would achieve a 11 seat majority over Democrats--a very modest advantage.

Over the next six months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the probability distribution for the number of seats gained by Republicans on election day. See my April 30, 2010 post for an explanation of my methodology.

Senate Prediction: Republicans Gain Five Seats

Using a similar methodology as my model for the U.S. House, I predict that the Republicans will gain five seats in the U.S. Senate. Here is a seat-by-seat summary (margin of victory in parentheses; party turnover in italics):

Safe Democratic, 6 Seats (15 percent+) :
Boxer (CA)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Leahy (VT)

Strong Democratic, 4 Seats (5.0 -15.0 percent):
IL (Burris)
Wyden (OR)
Murray (WA)
Feingold (WI)

Weak Democratic, 3 Seats (0.0 - 5.0 percent):
CT (Dodd)
Reid (NV)
PA (Spector)

Weak Republican, 4 Seats (0.0 - 5.0 percent):
Bennet (CO)
DE (Kaufman)
FL (LeMieux)
NH (Gregg)

Strong Republican 6 Seats (5.0 - 15 percent):
McCain (AZ)
Grassley (IA)
IN (Bayh)
MO (Bond)
Burr (NC)
OH (Voinovich)


Safe Republican 13 Seats (15.0 percent+):
Shelby (AL)
Lincoln (AR)
Crapo (ID)
Isakson (GA)
Murkowski (AK)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Coburn (OK)
DeMint (SC)
Thune (SD)
UT (Bennett)

A Projection for the Senate: Republicans Gain Five Seats

Using a similar methodology as my model for the U.S. House, I have constructed predictions for the Senate. Here is a seat-by-seat summary (margin of victory in parentheses; party turnover in italics):

Safe Democratic, 6 Seats (15 percent+) :
Boxer (CA)
Inouye (HI)
Mikulski (MD)
Schumer (NY-A)
Gillibrand (NY-B)
Leahy (VT)

Strong Democratic, 5 Seats (5.0 -15.0 percent):
CT (Dodd)
IL (Burris)
Wyden (OR)
Murray (WA)
Feingold (WI)

Weak Democratic, 2 Seats (0.0 - 5.0 percent):
Reid (NV)
Specter (PA)

Weak Republican, 3 Seats (0.0 - 5.0 percent):
Bennet (CO)
DE (Kaufman)
NH (Gregg)

Strong Republican 8 Seats (5.0 - 15 percent):
McCain (AZ)
Lincoln (AR)
FL (LeMieux)
Grassley (IA)
IN (Bayh)
MO (Bond)
Burr (NC)
OH (Voinovich)


Safe Republican 12 Seats (15.0 percent+):
Shelby (AL)
Crapo (ID)
Isakson (GA)
Murkowski (AK)
KS (Brownback)
KY (Bunning)
Vitter (LA)
ND (Dorgan)
Coburn (OK)
DeMint (SC)
Thune (SD)
Bennett (UT)

Revised Model Predicts 45 Seat Gain for Republicans

Based on polling data released over the last four months, Republicans are expected to gain between 25 and 60 seats in the U.S. House of Representatives (with a 90 percent level of credibility). According to the model, the most likely outcome is about a 45 seat gain by Republicans. Such an electoral accomplishment would be sufficient to unseat the Democratic Party from majority status in the lower chamber. The Republican gains would also represent the largest electoral landslide since 1994.

A statistical model that analyzes the results of 15 independent polling firms indicates Republicans lead Democrats by a margin of 1.8 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. This figure is little changed from the March estimate of 2.0 percent. The March figure, however, has been revised downward from its preliminary estimate of 6.8 percent as a result of a technical correction in the data, along with the incorporation of post-health care polling. The April estimate implies that, if the election were held today, Republicans would achieve a 13 seat majority over Democrats--a very modest advantage.

Over the next seven months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the probability distribution for the number of seats gained by Republicans on election day. I will also post an explanation of my methodology below.

Methodology:

I use a mixture of beta-binomial models. Let Xi by the percentage of the national electorate supporting the Republicans according to pollster i. Let theta be my parameter of interest (i.e. the percentage of the votes that prefer the Republican Party). Then {theta | X1, X2, ... Xm} ~ 1/m*{X1 + X2 + ... + Xm}. In other words, I calculate the probability distribution of theta by simply taking the unweighted mean of all of the pollsters' distributions. One advantage of this method is that each pollster's belief is given equal weight, regardless of the number of polls that they have conducted. This characteristic prevents frequent pollsters (who often use internet or automated polling) from being disproportionately influential on the results.

How do I find the distribution of each Xi? Ah, now we have arrived at the interesting part. For this step, I use the beta-binomial model from Bayesian statistics. That is, Xi | Yj ~ beta(a[j], b[j])*binomial(y[j] | p, n[j]) = beta(a + y[j], b + n[j]- y[j]), where a[j]=a[j-1] +y[j-1], b[j]=b[j-1] +n[j-1] - y[j-1]. Y[j] is the number of people in the sample of poll j by pollster i that support the Republicans, while n[j] is the sample size of poll j. The polls are ordered chronologically so that the prior (beta(a[j], b[j])) is the posterior produced by the preceding poll. By using this recursive formula, we will have obtained the posterior distribution of Xi when we have iterated through to the most recent poll.

The initial prior for all posteriors of {X1...Xm} is beta(7.2, 6.6), which corresponds roughly to the results of the 1994 election (though note that the weight on this prior is very low). Furthermore, I normalize the y[j]s and n[j]s by a vector of constants in order to account for the staleness of each poll, i.e. the more recent polls receive the most weight. I apply this Bayesian inference to the following 15 pollsters:

Rasmussen (14 polls since January 1, 2010)

CNN (5)

Quinnipiac (1)

Democracy Corps (5)

YouGov (12)

Gallop (8)

PPP (4)

National Review (2)

Ipsos (1)

Newsweek (1)

Public Opinion Strategies (2)

Pew (2)

ABC/Washington Post (3)

Franklin & Marshall (1)

NPR (1)

I estimate the number of seats held by Republicans after the election by using the following formula: # of seats = theta*435 + 5. Thus, the number of seats is roughly proportional to the level of national support for the Republican caucus. However, I do add five seats to the estimate in order to compensate for structural advantages that Republicans have in the Congressional elections.
______________________________________________
NB: I have adjusted one element of my model since March. I have altered the weight that is applied to each poll. Previously, I normalized the sample size of each poll according to the date at which the data was retrieved--such that the most recent polls received the most influence. I still give the most recent polls more weight. However, the precise algorithm for determining the weightings has changed. My previous strategy was to give the most recent polls a weight of one (and older polls a weigh in [0, 1]). Now, I give all polls a weight of (1/2)^(M), where M is the number of months between the collection date of the poll and the day of the election.

Why the change? Well, my ultimate objective is the produce a probability distribution for the number of seats gained by Republicans on election day. However, my former approach really created a probability distribution for the number of seats gained by Republicans if the election were held today. The mode (or expectation) of the distribution for the number of Republican acquisitions was correct, but the distributional precision was too high because I gave the most recent polls a full weight--even though the most recent polls are still months removed from the election. By weighting the sample sizes on the basis of the amount of time remaining before the election, I can provide a distribution that has a much larger and more plausible variance. In order words, the discrepancy between the spreads of the March distribution and April distribution is entirely due to this modification in my methodology.

I should finally mention that this alteration, nevertheless, has little to no effect on the central tendency of my predictions because it does not change the relative magnitudes of the weights. (The new algorithm also has the nice property that weight for a given poll does not change from month-to-month since the magnitude is based on the date of the election, rather than the post date of my projections.)

Republicans Are Expected to Gain Over 50 Seats

Based on polling data released over the last three months, Republicans are expected to gain between 50 and 60 seats in the U.S. House of Representatives. Such an electoral accomplishment would be sufficient to unseat the Democratic Party from majority status in the lower chamber of Congress. The Republican gains would also represent the largest electoral landslide since 1994.

A statistical model that analyzes the results of 15 independent polling firms indicates Republicans lead Democrats by a margin of 6.8 percent on the so-called generic ballot--a measure of national voter preference between the two major congressional caucuses. The model suggests that, if the election were held today, Republicans would achieve a 35 seat majority over Democrats--a slim, but decisive majority.

Over the next eight months, I will update this model and estimate Republican electoral gains. The charts below will provide the change in voter preferences over time, along with the current probability distribution for Republican seats gained. I will also post an explanation of my methodology below.

Methodology:

I use a mixture of beta-binomial models. Let Xi by the percentage of the national electorate supporting the Republicans according to pollster i. Let theta be my parameter of interest (i.e. the percentage of the votes that prefer the Republican Party). Then {theta | X1, X2, ... Xm} ~ 1/m*{X1 + X2 + ... + Xm}. In other words, I calculate the probability distribution of theta by simply taking the unweighted mean of all of the pollsters' distributions. One advantage of this method is that each pollster's belief is given equal weight, regardless of the number of polls that they have conducted. This characteristic prevents frequent pollsters (who often use internet or automated polling) from being disproportionately influential on the results.

How do I find the distribution of each Xi? Ah, now we have arrived at the interesting part. For this step, I use the beta-binomial model from Bayesian statistics. That is, Xi | Yj ~ beta(a[j], b[j])*binomial(y[j] | p, n[j]) = beta(a + y[j], b + n[j]- y[j]), where a[j]=a[j-1] +y[j-1], b[j]=b[j-1] +n[j-1] - y[j-1]. Y[j] is the number of people in the sample of poll j by pollster i that support the Republicans, while n[j] is the sample size of poll j. The polls are ordered chronologically so that the prior (beta(a[j], b[j])) is the posterior produced by the preceding poll. By using this recursive formula, we will have obtained the posterior distribution of Xi when we have iterated through to the most recent poll.

The initial prior for all posteriors of {X1...Xm} is beta(7.2, 6.6), which corresponds roughly to the results of the 1994 election (though note that the weight on this prior is very low). Furthermore, I normalize the y[j]s and n[j]s by a vector of constants in order to account for the staleness of each poll, i.e. the more recent polls receive the most weight. I apply this Bayesian inference to the following 15 pollsters:

Rasmussen (11 polls since January 1, 2010)

CNN (3)

Quinnipiac (1)

Democracy Corps (4)

YouGov (6)

Gallop (3)

PPP (3)

National Review (2)

Ipsos (1)

Newsweek (1)

Public Opinion Strategies (1)

Pew (1)

ABC/Washington Post (1)

Franklin & Marshall (1)

NPR (1)

I estimate the number of seats held by Republicans after the election by using the following formula: # of seats = theta*435 + 5. Thus, the number of seats is roughly proportional to the level of national support for the Republican caucus. However, I do add five seats to the estimate in order to compensate for structural advantages that Republicans have in the Congressional elections.