Statistikhengst
Diamond Member
- Banned
- #1
I want to pass along some thoughts about Tuesday's election.
First, the GOP victories were far larger in many places than polling, and more importantly, polling aggregates, were pointing to. In the fullness of time, meaning, when the last final canvass has been turned in, I will do an exact analysis, but it is obvious that many, many pollsters were considerably off, including most all Right-Wing leaning pollsters. And most of the time, the mathematical bias was to the LEFT and not to the RIGHT. Now, there are a number of excuses that could be made, but that this point in time, that is meaningless. And believe me, I am just as scathingly hard on pollsters when the polls are off to the Left as when they are off to the Right. Most states have the rule that the final canvass must be in by December 16th: this has to do with the fact that in presidential election years, the so-called "Electoral College" meets between December 16th and 18th of that year, in the respective statehouses, and officially cast their votes for President. In order to keep the laws simple, most states make that the rule for all election years.
So, for people who were wondering why this can take so long, now you know. Some states, like Colorado, have a three-week time limit for the final canvasses. Some states, like Vermont, get it done within 7 days.
That being said, 8 of 10 aggregates for the 10 marquee senatorial races still pointed toward the actual winner, even when the margins were understated. That's 80%.
When ALL final canvasses are in, I will be posting a massive thread with links to each and every canvass, so that you all can see for yourselves how the voting worked out, down to the 100th of a percent.
Therefore, it makes no sense to try to do major and specific comparisons until the final canvasses are in and all is sealed in stone as part of the congressional record.
But there are some general points I can make now. Right now, with 46.2 million votes having been tallied in the Senatorial races, the Republicans currently have a +4.87% margin in the overall popular vote:
Note: the colors are reversed here. They are from Dave Leip's uselectionatlas website, and he uses reverse colors. I highly recommend his site for such stats, because he does a helluva job at keeping stuff updated.
I want to remind again: that figure is only preliminary, it is bound to change. The Senatorial raw vote is currently 20 million votes under where it landed in 2010, it is bound to go upward. Now, whether the overall margin changes upward or downward remains to be seen. But one thing is for sure: the GOP "gets more bang for it's buck" than the Democrats have in the past. A circa +5 margin in the popular vote has been enough to net the GOP +7 seats and when all is said and done, it may well be +9 seats.
Compare that to the Democrats in 2006 (a large blue wave): in order to flip 6 seats in the Senate in that year and just make it to 51, the Ds had a whopping +11.52% margin in the PV. And in 2012, a presidential election year, the D's had a nearly identical +11.82% in the PV for the senatorials, and yet, they picked up only 3 seats.
In 2006, the D's picked up 6 seats, and then they picked up 8 seats in 2008: that makes for 14 seats over two cycles.
In 2010, the R's picked up 6 seats, which, with the MA seat from 2009, makes +7. In 2012, they lost 3 seats, making a net +4. In 2014, they have picked up 7 seats and I suspect it will be +9 when all is said and done, making for a net +13 from 2010 to 2014.
So, we've seen the pendulum swing +13/+14 in both directions now, and that within 8 years, from 2006 to 2014.
That alone should be proof that despite lots of commentary about one party dying, both are alive and well.
I am going to go one step farther and argue that the 2014 election, despite the high probablity of lower turnout and possibly even lower raw vote numbers than 2010 (the last mid-term), is not not only a wave that the polls missed, but also a so-called re-alignment election. In the gubernatorials, the Ds lost a number of very blue states to the R's, albeit with small margins (MA, IL, MD, MI) and in the senatorials, they lost two critical battleground states: CO and IA, which pretty much sealed the deal for the Rs. In the 435 congressional races, I have predicted all year long that it would be about a +11 pickup for the GOP. They have now picked up a net of +13 (+15 minus 2 losses) and currently, the congressional map stands at:
Source: Politico
Of the 22 seats left to decide, forget about the 3 seats in Louisiana for now; they go into a runoff. And in WA-04, that is a R vs. R race, with no D involved. So, the Rs are guaranteed 244 seats, definitely. And in CA-17, it's a D vs D race with no R involved, to the D's are guaranteed to get to at least 181. That leaves 17 seats to look at until the end of this month.
NY-25 and AZ-02 are extremely close, but the Ds are likely to prevail in both, especially in AZ-02, where Ron Barber ended election night 1200 votes behind in 2012 and ended up winning when the absentee ballots were counted. In CA-09, CA-31 and CA-52, the D is about 3 points ahead, will probably win all three. But in CA-26, the D is barely ahead and in CA-16, the R is one point ahead. However, California is notorious for lots and lots of mail-in ballots and the returns usually take the full six weeks to get in before the final canvasses.
You can look at the general map for yourself to see what's up (it's clickable):
House Election Results 2014 Map by State Live Midterm Voting Updates - POLITICO
Geographically speaking, over 3 congressional election cycles, the Rs have literally gutted the Ds in MI, OH, PA, WV, VA and NY. There are now eleven states in the Union that are completely red in terms of HOR delegations, and seven of those states have more than one congressional district.
But geography is not the same as population density. Look at Maine. CD-02, which the GOP picked up, looks considerably larger than CD-01. Only, the vast majority of the population of ME lives in CD-01, not in CD-02. In Ohio, the three very small blue flecks (Cleveland, Youngstown, Columbus) represent more than 25% of the state's population. Likewise, in Illinois, the uppermost eastern counties, though small in size, are about 40% of the state's population (Chicago and environs), if not more. And if you look at the maps of FL, NC, GA, MD, MO, LA, MI. OH and PA, to name some, and look at the blue districts, you will see that some of them have the weirdest possible forms: gerrymandering at work.
Now, all of this in much, much more detail in January/February of 2015. It's my tradition to wait until after the final canvasses and then give myself time to really sift through lots of data before I do an extensive analysis. This analysis was very small and somewhat superficial, but a start.
Lots of pundits are giving their commentary on the 2014 mid-terms and the votemaster has collected some of that punditry here:
Electoral-vote.com
I would say that everyone here would probably find at least one point from each pundit that he or she can agree with.
And in parting, I would like to point to this thread that I created 11 months ago:
Congressional Elections compared to Presidential Terms 1855-present US Message Board - Political Discussion Forum
First, the excel table that is linked at the thread will be updated in the next months.
Second, I would like to re-quote the end of the OP again:
So, once again, electoral history has played out pretty much like it always has.
-Stat
First, the GOP victories were far larger in many places than polling, and more importantly, polling aggregates, were pointing to. In the fullness of time, meaning, when the last final canvass has been turned in, I will do an exact analysis, but it is obvious that many, many pollsters were considerably off, including most all Right-Wing leaning pollsters. And most of the time, the mathematical bias was to the LEFT and not to the RIGHT. Now, there are a number of excuses that could be made, but that this point in time, that is meaningless. And believe me, I am just as scathingly hard on pollsters when the polls are off to the Left as when they are off to the Right. Most states have the rule that the final canvass must be in by December 16th: this has to do with the fact that in presidential election years, the so-called "Electoral College" meets between December 16th and 18th of that year, in the respective statehouses, and officially cast their votes for President. In order to keep the laws simple, most states make that the rule for all election years.
So, for people who were wondering why this can take so long, now you know. Some states, like Colorado, have a three-week time limit for the final canvasses. Some states, like Vermont, get it done within 7 days.
That being said, 8 of 10 aggregates for the 10 marquee senatorial races still pointed toward the actual winner, even when the margins were understated. That's 80%.
When ALL final canvasses are in, I will be posting a massive thread with links to each and every canvass, so that you all can see for yourselves how the voting worked out, down to the 100th of a percent.
Therefore, it makes no sense to try to do major and specific comparisons until the final canvasses are in and all is sealed in stone as part of the congressional record.
But there are some general points I can make now. Right now, with 46.2 million votes having been tallied in the Senatorial races, the Republicans currently have a +4.87% margin in the overall popular vote:
Note: the colors are reversed here. They are from Dave Leip's uselectionatlas website, and he uses reverse colors. I highly recommend his site for such stats, because he does a helluva job at keeping stuff updated.
I want to remind again: that figure is only preliminary, it is bound to change. The Senatorial raw vote is currently 20 million votes under where it landed in 2010, it is bound to go upward. Now, whether the overall margin changes upward or downward remains to be seen. But one thing is for sure: the GOP "gets more bang for it's buck" than the Democrats have in the past. A circa +5 margin in the popular vote has been enough to net the GOP +7 seats and when all is said and done, it may well be +9 seats.
Compare that to the Democrats in 2006 (a large blue wave): in order to flip 6 seats in the Senate in that year and just make it to 51, the Ds had a whopping +11.52% margin in the PV. And in 2012, a presidential election year, the D's had a nearly identical +11.82% in the PV for the senatorials, and yet, they picked up only 3 seats.
In 2006, the D's picked up 6 seats, and then they picked up 8 seats in 2008: that makes for 14 seats over two cycles.
In 2010, the R's picked up 6 seats, which, with the MA seat from 2009, makes +7. In 2012, they lost 3 seats, making a net +4. In 2014, they have picked up 7 seats and I suspect it will be +9 when all is said and done, making for a net +13 from 2010 to 2014.
So, we've seen the pendulum swing +13/+14 in both directions now, and that within 8 years, from 2006 to 2014.
That alone should be proof that despite lots of commentary about one party dying, both are alive and well.
I am going to go one step farther and argue that the 2014 election, despite the high probablity of lower turnout and possibly even lower raw vote numbers than 2010 (the last mid-term), is not not only a wave that the polls missed, but also a so-called re-alignment election. In the gubernatorials, the Ds lost a number of very blue states to the R's, albeit with small margins (MA, IL, MD, MI) and in the senatorials, they lost two critical battleground states: CO and IA, which pretty much sealed the deal for the Rs. In the 435 congressional races, I have predicted all year long that it would be about a +11 pickup for the GOP. They have now picked up a net of +13 (+15 minus 2 losses) and currently, the congressional map stands at:
Source: Politico
Of the 22 seats left to decide, forget about the 3 seats in Louisiana for now; they go into a runoff. And in WA-04, that is a R vs. R race, with no D involved. So, the Rs are guaranteed 244 seats, definitely. And in CA-17, it's a D vs D race with no R involved, to the D's are guaranteed to get to at least 181. That leaves 17 seats to look at until the end of this month.
NY-25 and AZ-02 are extremely close, but the Ds are likely to prevail in both, especially in AZ-02, where Ron Barber ended election night 1200 votes behind in 2012 and ended up winning when the absentee ballots were counted. In CA-09, CA-31 and CA-52, the D is about 3 points ahead, will probably win all three. But in CA-26, the D is barely ahead and in CA-16, the R is one point ahead. However, California is notorious for lots and lots of mail-in ballots and the returns usually take the full six weeks to get in before the final canvasses.
You can look at the general map for yourself to see what's up (it's clickable):
House Election Results 2014 Map by State Live Midterm Voting Updates - POLITICO
Geographically speaking, over 3 congressional election cycles, the Rs have literally gutted the Ds in MI, OH, PA, WV, VA and NY. There are now eleven states in the Union that are completely red in terms of HOR delegations, and seven of those states have more than one congressional district.
But geography is not the same as population density. Look at Maine. CD-02, which the GOP picked up, looks considerably larger than CD-01. Only, the vast majority of the population of ME lives in CD-01, not in CD-02. In Ohio, the three very small blue flecks (Cleveland, Youngstown, Columbus) represent more than 25% of the state's population. Likewise, in Illinois, the uppermost eastern counties, though small in size, are about 40% of the state's population (Chicago and environs), if not more. And if you look at the maps of FL, NC, GA, MD, MO, LA, MI. OH and PA, to name some, and look at the blue districts, you will see that some of them have the weirdest possible forms: gerrymandering at work.
Now, all of this in much, much more detail in January/February of 2015. It's my tradition to wait until after the final canvasses and then give myself time to really sift through lots of data before I do an extensive analysis. This analysis was very small and somewhat superficial, but a start.
Lots of pundits are giving their commentary on the 2014 mid-terms and the votemaster has collected some of that punditry here:
Electoral-vote.com
I would say that everyone here would probably find at least one point from each pundit that he or she can agree with.
And in parting, I would like to point to this thread that I created 11 months ago:
Congressional Elections compared to Presidential Terms 1855-present US Message Board - Political Discussion Forum
First, the excel table that is linked at the thread will be updated in the next months.
Second, I would like to re-quote the end of the OP again:
So, what can we learn from this information?
1.) We have had divided government a lot more than most people realize.
2.) The statistical probability that the President's party suffers major losses in a mid-term election, especially in a 2nd term mid-term, is extremely high.
3.) Divided government is not necessarily a bad thing. See: Eisenhower, Clinton. Likewise, unified government is not necessarily good. See: Hoover, Carter.
So, before the pundits go blabbing their mouths off about the 2014 elections, know that history is on the side of the GOP in this election.
So, once again, electoral history has played out pretty much like it always has.
-Stat
Last edited: