Posts Tagged ‘Poll’

Michigan Primary Prediction and Statistical Analysis

February 28, 2012

It’s been a while since I updated this, but I figured this would be fun.  I made my primary prediction for Michigan late last night and here’s the breakdown I went with:

Santorum – 40%
Romney – 39%
Paul – 11%
Gingrich – 9%
Others – 1%

Then today, I decided to look at the recent polls for Michigan.  The first poll I went with was starting on February 12, since anything before this wouldn’t factor in Santorum’s triple win in Missouri, Colorado, and Minnesota.  I made a scatter plot of Santorum – Romney:

I then added 2 trendlines. The first one (red) was a linear trendline which I knew wouldn’t be reflective of Santorum’s momentum recently. I plotted the second (green) second order polynomial trendline, which did a decent job at predicting his momentum, but I noticed that the relatively low lead he had right after his triple win was skewing the results a bit lower.  Those polls are so old that they don’t correctly account for what’s going on now with just a simple second order polynomial.  So I put in the third (purple) trendline, which was a third order polynomial.  This correctly factored in his slow start to gain momentum, but looks like it drastically overestimates it at the end.

So I thought, how can I better examine Santorum’s momentum.  I took out the polls right after Santorum’s triple-win, which I thought weren’t representative of the momentum he gained, then lost, and regained, and got this plot, with the red representing a linear trendline, and the green representing a second order polynomial:

This plot matched up almost perfectly with my predictions (which I wasn’t expecting it to turn out just like that), but I think accurately portrays the general momentum Santorum has had since his triple win.

Ultimately, gauging the momentum a candidate has is hard to do, but I think the second order polynomial in the second graph accurately portrays Santorum’s momentum.  Ultimately, both of these models could end up being wrong, and there’s no perfect algorithm to predict voters, but the results of my second model, I believe, accurately portray Santorum’s momentum, and cement my confidence that Santorum will have a close victory over Mitt Romney.

Done Predicting,

Ranting Republican

New York 23rd District Election Prediction: Hoffman Wins

November 2, 2009

I already put out my predictions for the New Jersey and Virginia gubernatorial races.  The other major race going on tomorrow is the special election for the New York 23rd Congressional District.  Originally, there were 3 main candidates running: Republican Dede Scozzafava, Democrat Bill Owens, and Conservative Doug Hoffman.  Hoffman entered the race because people had criticized Scozzafava as being too moderate, some saying she was even more liberal than the Democrat.  Top Republicans were split in who they supported, with some Republicans like Newt Gingrich supporting Scozafava, and Sarah Palin supporting Doug Hoffman.

Last week, Scozzafava dropped out of the race and endorsed Owens.  At that point, Scozzafava was trailing in the polls by over 10%, and the race between Owens and Hoffman was close.  Since Scozzafava dropped out, Hoffman has skyrocketed in the polls, and I now expect him to win.

Even though Scozzafava dropped out, it’s too late to change the ballots, so she will remain on the ballot.  Here’s my prediction:

  1. Doug Hoffman (C) – 53%
  2. Bill Owens – 42%
  3. Dede Scozzafava – 5%

I really don’t see Hoffman having any problems now that Scozzafava has dropped out – the district leans Republican and hasn’t gone for a Democrat running for the District since 1992.  I see Hoffman winning pretty easily tomorrow, but we’ll see – it’s been an interesting race so far – there could always be another surprise.

Done Predicting,

Ranting Republican

New Jersey, Virginia Gubernatorial Predictions: Christie, McDonnell Win

November 2, 2009

It’s that time again – election time!  As always, I’ve been tracking the polls, and I’ve done my final predictions.  This year, there are only 2 gubernatorial races: New Jersey and Virginia.  In case you didn’t follow my predictions last year, you may find the color scheme on the maps to be a little bit weird – the colors are switched from what the normal media colors, so sorry about that, but that’s the way the website I use does it.  The maps are courtesy of Dave Leip’s U.S. Election Atlas, and my most current prediction can always be found here.  On to the predictions…

* = Party Pickup (Incumbent Loses Reelection); ^ = Party Pickup (Open Seat)

So here are my predictions, broken down by percentage:

New Jersey

  1. Chris Christie (R) - 46.35%
  2. Jon Corzine (D) – 45.47%
  3. Chris Dagget (I) – 8.18%

Virginia

  1. Bob McDonnell (R) – 57%
  2. Creigh Deeds (D) – 42%
  3. Other candidates – <1%

And here’s my confidence map:

Essentially – I could go either way on New Jersey.  The numbers I posted are the results of an algorithm I’ve developed, and tweaked from last year.  My heart is telling me that Corzine will probably pull out a win, but my mind and the numbers point toward Christie, I’m gonna stick with Christie.  There’s no doubt that McDonnell will win in Virginia.

I don’t see New Jersey called until late into the night tomorrow.  It should be an interesting night.

Done Predicting,

Ranting Republican

1 Day to Go: Final Gubernatorial Elections Prediction

November 3, 2008

Alright, this will finish out my prediction series for today, following my Presidential and Senate election predictions. Again, the colors are switched from what the normal media colors, so sorry about that, but that’s the way the website I use does it. The maps are courtesy of Dave Leip’s U.S. Election Atlas, and my most current prediction can always be found here. On to the predictions…

* = Pickup via defeat of incumbent; ^ = Pickup of an open seat

Democrats: 7 (+1)
Republicans: 4 (-1)
Light gray indicates states with no Gubernatorial races

So, some of the states where you may disagree with my prediction:

  • North Carolina: Pat McCroy (R) had looked like she was going to overtake Bev Perdue (D), but Perdue has begun to come back in the polls, and I think she’s going to be able to hang on and win, but I’m still not confident enough to put it in the “Strong” category.
  • Washington: I’m still fairly confident that Christine Gregoire (D) will keep her seat and beat out Dino Rossi (R), but I’m not confident to put it in the strong category.

Now, on to my map shsowing my confidence in my predictions:

Democrats: 7 (+1)
Republicans: 4 (-1)
Tossup: 0

Light gray indicates states with no Gubernatorial races

And here’s how my prediction changed:

State

Previous

Current

Reasoning

NC

D40T

D40L

I’m confident enough to take this out of the Toss-Up category.

Well, that’s my gubernatorial election prediciton!

Done Predicting,

Ranting Republican
add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::

1 Day to Go: Final Senate Prediction: Democrats Gain 7 Seats

November 3, 2008

Here’s my Senate prediction. I already did my Presidential election prediction as well as my gubernatorial elections prediction.  The colors ARE switched from what the normal media colors, so sorry about that, but that’s the way the website I use does it.  The maps are courtesy of Dave Leip’s U.S. Election Atlas, and my most current prediction can always be found here.  On to the predictions…

* = Pickup via defeat of incumbent; ^ = Pickup of an open seat

Democrats: 19 (+7)
Republicans: 14 (-7)
Light gray indicates states with no Senate races

First, a note about Wyoming and Mississippi:

  • Wyoming has 2 races, Michael Enzi (R) against Chris Rothfuss (D) and John Barasso (R) against Nick Carter (D).  I expect both Republicans to win with results around 63%.
  • Mississippi has 2 races.  The maps are predictions for Thad Cochran (R) against Erik Fleming (D). In the election between Roger Wicker (R) and Ronnie Musgrove (D), I expect Wicker to win with around 55%, a much closer race than the other.

Alright, so let’s look at the states where people may disagree with me:

  • Arkansas: Mark Pryor is unopposed by a Republican. Rebekah Kennedy (Green) is the only opponent, so that’s why I have it so high. It’s not a mistake.
  • Alaska: 2 predictions ago, I had it going to “Uncle Ted” Stevens. Then, the jury found him guilty. I changed my prediction on the U.S. Election Atlas website, but didn’t repost a prediction here (although I did write a blog post saying that Mark Begich would win). Then, I started thinking, and I think Stevens will pull it off. Then, last prediction, I said, “I know the polls disagree (but the Research 2000 poll showing him down 22% is just wrong), but I don’t see Alaskans voting out Uncle Ted.”  Well, I just can’t justify keeping it in Uncle Ted’s column anymore.  I think it’s going to go to Mark Begich, but I’m not confident enough to put it in the “Strong” column.
  • Minnesota: I’m now confident enough in my call for Coleman, and he has been slightly leading in the polls.
  • North Carolina: I think Elizabeth Dole’s attack ad on Kay Hagan gave Hagan the edge she’ll need to win, but still I’m not confident enough to put it in the “Strong” column.

Now, the map indicating the confidence that I have that my prediction is right:

Democrats: 19 (+7)
Republicans: 14 (-7)
Tossup: 3
Light gray indicates states with no Senate races

Alright, so what changes did I make since last time, and why? Here they are:

State

Previous

Current

Reasoning

AK

R50T

D50L

I think the guilty verdict finally did it in for him.  Although I’m not going to mark it as “Strong” because those Alaskan’s sure love their Uncle Ted, and weirder things have happened.

MN

R40T

R40L

I think I’m confident enough to take it out of the Toss-Up category now.

NC

D50T

D50L

Again, I’m confident enough to make this a “Lean” instead of “Toss-Up.”

Well, there you have it – my Senatorial prediction.

Done Ranting,

Ranting Republican
add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::

1 Day to Go: Final Presidential Election Prediction: Obama 291, McCain 247

November 3, 2008

Alright, so here’s my final prediction for the Presidential elections. None of my state-by-state predictions from last time changed, other than a percentages or for 2 states (Indiana and Montana).  I’ve also done predictions for the Senate and Gubernatorial elections.

The maps are provided courtesy of Dave Leip’s U.S. Election Atlas (and you can find the most updated version of my prediction on that website here). The colors on this website are switched from the normal colors (and I’m too lazy to switch them back to the normal colors – but back in the 1980s, these colors were the colors that the media used).

Alright, on to the predictions…

Obama – 291
McCain – 247

So, the states where you might disagree with my position:

  • Ohio: As the current trand of the past week keeps up, it’ll come out right about tied. I think McCain’s last minute blitz campaigning will help him pull out just BARELY on top.  Maybe I’m just too optimistic, but I see McCain winning here.
  • Misssouri, McCain has been leading recently, baring a few polls.
  • North Carolina: I was unsure last week, and I’m still unsure this week. I really haven’t changed my opinion on this.  I had been hoping to be able to take it out of the toss-up category, but I can’t do it for either McCain or Obama.  I really struggled with this one.  It could honestly go eiher way.
  • Florida: Despite Obama’s recent lead in the polls, McCain has come back, and is trending toward taking the lead by tomorrow (and he’s already taken the lead according to some polls).
  • Indiana: I know some polls have shown Obama leading, but I just don’t see it happening tomorrow.

Now, the map indicating the confidence that I have that my prediction is right:

Obama – 291
McCain – 185
Tossup – 62

Now, how has this map changed since my last map? Here’s a chart of the changes:

State

Previous

Current

Reasoning

IN

R40L

R50L

It looks like McCain will reach 50% here.

MT

R50S

R50L

Obama is catching up here, but I don’t think he’ll pull off a win.

So, again, at this point, I think it’s gonna take a small miracle for McCain to win, and I don’t see a small miracle happening tomorrow. All of the states that are toss-ups, I have going to McCain, so, in theory, if the toss-up states go to Obama, Obama would win 353-185.

Done Predicting,

Ranting Republican
add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::

4 Days Away: Gubernatorial Elections Prediction

November 1, 2008

Alright, this will finish out my prediction series for today, following my Presidential and Senate election predictions.  Again, the colors are switched from what the normal media colors, so sorry about that, but that’s the way the website I use does it.  The maps are courtesy of Dave Leip’s U.S. Election Atlas, and my most current prediction can always be found here.  On to the predictions…

* = Pickup via defeat of incumbent; ^ = Pickup of an open seat

Democrats: 6 (+1)
Republicans: 5 (-1)
Light gray indicates states with no Gubernatorial races

So, some of the states where you may disagree with my prediction:

  • North Carolina: Pat McCroy (R) had looked like she was going to overtake Bev Perdue (D), but Perdue has begun to come back in the polls, and I think she’s going to be able to hang on and win.  This could change by my next prediction (since it’s still a toss-up in my prediction), but I don’t think it will.
  • Washington: Last time I had it going to Christine Gregoire (D), the incumbent, but a toss-up.  Today, I’m pretty confident that it’ll stay in her hands.  Dino Rossi hasn’t led / been tied since early October, and I just don’t see Rossi taking the lead.  Interestingly, over the last few days Rossi has closed the gap that had opened up about 7 days ago.  By Monday, I may have to change it back to a toss-up, but as of now, I’m fairly confident in my prediction.

Now, on to my map shsowing my confidence in my predictions:

Democrats: 5 (+1)
Republicans: 5 (-1)
Tossup: 1

Light gray indicates states with no Gubernatorial races

So, there we go.  The Republicans are going to lose 1 Governor position (Missouri), but may have the opportunity to pick one up in North Carolina.  I guess we can’t rule out a small miracle in Washington, but I don’t see the Rossi winning there.  I’ll have a final prediction on Monday.

Done Predicting,

Ranting Republican
add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::

4 Days to Go: Senate Prediction: Democrats Gain 6 Seats

October 31, 2008

Here’s my Senate prediction.  I already did my Presidential election prediction as well as the Gubernatorial Elections prediction.  The colors ARE switched from what the normal media colors, so sorry about that, but that’s the way the website I use does it.  The maps are courtesy of Dave Leip’s U.S. Election Atlas, and my most current prediction can always be found here.  On to the predictions…

* = Pickup via defeat of incumbent; ^ = Pickup of an open seat

Democrats: 16 (+6)
Republicans: 19 (-6)
Light gray indicates states with no Senate races

First, a note about Wyoming and Mississippi:

  • Wyoming has 2 races, Michael Enzi (R) against Chris Rothfuss (D) and John Barasso (R) against Nick Carter (D). I expect both Republicans to win with results around 63%.
  • Mississippi has 2 races. The maps are predictions for Thad Cochran (R) against Erik Fleming (D). In the election between Roger Wicker (R) and Ronnie Musgrove (D), I expect Wicker to win with around 51%, a much closer race than the other.

Alright, so let’s look at the states where people may disagree with me:

  • Arkansas: Mark Pryor is unopposed by a Republican.  Rebekah Kennedy (Green) is the only opponent, so that’s why I have it so high.  It’s not a mistake.
  • Alaska: Last prediction, I had it going to “Uncle Ted” Stevens.  Then, the jury found him guilty.  I changed my prediction on the U.S. Election Atlas website, but didn’t repost a prediction here (although I did write a blog post saying that Mark Begich would win).  Then, I started thinking, and I think Stevens will pull it off.  I know the polls disagree (but the Research 2000 poll showing him down 22% is just wrong), but I don’t see Alaskans voting out Uncle Ted.
  • Minnesota: Again, like last time, for every poll that comes out showing Franken ahead, a poll comes out showing Coleman ahead by the same amount.  Right now, it’s just too close to call, so I’ll keep it where I had it last week.

Now, the map indicating the confidence that I have that my prediction is right:

Democrats: 16 (+5)
Republicans: 19 (-5)
Tossup: 3
Light gray indicates states with no Senate races

Alright, so what changes did I make since last time, and why?  Here they are:

State

Previous

Current

Reasoning

AK

D50L

R50T

I don’t think they’ll vote “Uncle Ted” out.

IA

D50S

D60S

It looks like Tom Harkin will reach 60% here.

KY

R50L

R50S

Mitch McConnell seems to be making a stronger comeback, back from when it was looking like a close race.  I think he’s now safe for sure.

NE

R60S

R50S

I think this an oversight the first time I did the predictions.

NM

D60S

D50S

I think Steve Pearce (R) has gained enough support that he’ll keep Tom Udall (D) from getting above 60%.

NC

R50T

D50T

Elizabeth Dole’s “Godless” attack ad against Kay Hagan was found out to be less than true.  I think there’ll be big backlash against Dole, and I think it’ll go to Hagan, but it is still a little bit too close to call in my opinion.

OR

R40T

D50S

Like I said last week, if Gordon Smith (R) didn’t pull ahead (as he was looking like he might), I was going to slide it over to Jeff Merkley, and that’s what I did.

By Monday, when I do my final update, I should be able to take North Carolina out of the toss-up category, and if more polls come out with Begich leading by a huge margin in Alaska, I’ll switch it back to Begich.  I don’t think I’ll be able to take Minnesota out of the toss-up category, but Al Franken’s latest campaign ad controversy may help Coleman, and I may be able to slide it to the “Lean” category.

Come back here on Monday for my final prediction.

Done Predicting,

Ranting Republican

add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::

4 Days to Go: My Presidential Election Prediction: Obama 291, McCain 247

October 31, 2008

Alright, so last Tuesday, I did a prediction for the Presidential election, predicting an Obama with victory of 286-252, and said I’d update it this Tuesday.  Well, I got a little busy, but here’s my updated prediction.  I’ve also done a prediction for the Senate and Gubernatorial elections.

The maps are provided courtesy of Dave Leip’s U.S. Election Atlas (and you can find the most updated version of my prediction on that website here).  The colors on this website are switched from the normal colors (and I’m too lazy to switch them back to the normal colors – but back in the 1980s, these colors were the colors that the  media used). 

Alright, on to the predictions…

Obama – 291
McCain – 247

So, the states where you might disagree with my position:

  • Ohio: As the current trand of the past week keeps up, it’ll come out right about tied.  I think McCain’s last minute blitz campaigning will help him pull out just BARELY on top.
  • Misssouri, McCain has been leading recently, baring a few polls.
  • North Carolina: I was unsure last week, and I’m still unsure this week.  I really haven’t changed my opinion on this.  I’ll make a hopefully more final decision when I do my final prediction on Monday.  As of now, I think it’ll just barely go to McCain.
  • Florida: Despite Obama’s recent lead in the polls, I think McCain will come back and win here, but if Obama continues to increase his lead, I’ll switch it over to him.

Now, the map indicating the confidence that I have that my prediction is right:

Obama – 291
McCain – 185
Tossup – 62

Now, how has this map changed since my last map? Here’s a chart of the changes:

State

Previous

Current

Reasoning

AK

R60S

R50S

Doesn’t look like McCain will reach 60%.

AZ

R60S

R50S

Doesn’t look like McCain will reach 60%

CA

D50S

D60S

Looks like Obama will manage to reach 60% here.

DE

D50S

D60S

Looks like Obama will manage to reach 60% here.

IN

R50S

D40L

More and more polls are coming out showing that Indiana is going to go to Obama.  I’m not exactly 100% sure yet, but it’s looking like it’ll be decently hard for McCain to come back and be able to win here.

NV

R40T

D50L

Again, like Indiana, the polls just don’t look like McCain will win here.  Originally, I was predicting a declaration of Obama’s victory early in the night and thought that less Obama supporters in Nevada would show up, but Obama has now garnered enough support here in Nevada, that an early victory declaration won’t give McCain a victory in Nevada.

PA

D50S

D50L

Originally, this looked really strong for Obama, but then all 4 candidates (President and VP) headed to Pennsylvania to campaign, which made me suspicious as to Obama’s strength there.  Combined with a blitz of campaigning here on McCain’s part plus Representative John Murtha’s comments about West Pennsylvania, I’m switching this over to the “Lean” category, but I’m still confident that Obama will win.

TN

R60S

R50S

Doesn’t look like McCain will reach 60%.

ME CD1

D50S

D60S

Based on recent polls that broke down demographic areas of the state, it looks like Obama will reach 60% here, but I don’t think he will in District 2 or on a statewide level.

So, at this point, I think it’ll take a small miracle for McCain to win.  All of the states that are toss-ups, I have going to McCain, so, in theory, if the toss-up states go to Obama, Obama would win 353-185.

I’ll do a final prediction on Monday, so check back for my final prediction.

Done Predicting,

Ranting Republican
add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::

2 Weeks Away: Gubernatorial Elections Prediction

October 21, 2008

Alright, this will finish out my prediction series for today, following my Presidential and Senate election predictions.  Again, the colors are switched from what the normal media colors, so sorry about that, but that’s the way the website I use does it.  The maps are courtesy of Dave Leip’s U.S. Election Atlas, and my most current prediction can always be found here.  On to the predictions…

* = Pickup via defeat of incumbent; ^ = Pickup of an open seat

Democrats: 6 (+1) [net gain of 0]
Republicans: 5 (+1) [net gain of 0]
Light gray indicates states with no Gubernatorial races

Now, on to discussion of states where you may disagree with me:

  • North Carolina: It’s all tied up in the polls right now.  Bev Perdue (D) had been leading, but now Pat McCroy (R) has caught up.  It’s looking like he’s right about to statistically overtake her, but I’ll wait a week until I take anything out of the toss-up category.  The next couple of polls that come out will be crucial in gauging North Carolina.
  • Washington: I continue to switch this one back and forth.  At the time of my last prediction, Dino Rossi (R) had been leading for a month, but now Christine Greggoire (D), the incumbent, is back in the lead, and I think she’s going to manage to stay there through election day.  Again, anything could happen here, but I’ll probably be able to slide her over to the Lean Democrat category next week.

Really quick, here’s my confidence map:

Overall, I’m pretty confident that I’ll be able to knock both of these toss-up states into the Lean columns next week without having to change anything, but who knows what could happen.

Thanks for reading, and come back next week for a whole set of new and updated predictions.

Done Predicting,

Ranting Republican
add to del.icio.us :: Add to Blinkslist :: add to furl :: add to ma.gnolia :: Stumble It! ::


Follow

Get every new post delivered to your Inbox.

Join 288 other followers

%d bloggers like this: