Statistikhengst
Diamond Member
- Banned
- #1
On the heels of the first part of an ABC/WAPO poll showing Clinton at +6 over Jeb Bush and +16 over him in a theoretical three-way race with Donald Trump as an Independent, PPP (D), undisputably the best pollster from 2012, has released it's latest national poll, and it is most interesting:
http://www.publicpolicypolling.com/pdf/2015/PPP_Release_National_72215.pdf
This .pdf is 108 pages long. I recommend that you download the .pdf and view it with a program like Adobe in order to find the page numbers more quickly.
pp 1-2: General Summary
pp.3-61: GOP nomination
pp. 62-73: DEM nomination
pp. 74-end: General election and assorted additional internals
1,087 RV for the GE matchups, MoE = +/-3.0
Republican nomination: 524 RRV, MoE = +/-4.3
Democratic nomination: 496 DRV, MoE = +/-4.4
There is a lot of information from this poll to glean.
Two caveats, just to save time:
1.) It's just one poll, but it is the 85th national poll since the 1st poll was released on December 5, 2012, more than 2.5 years ago. There have now been, including this PPP (D) poll, a sum-total of 365 head-to-head matchups from the 85 polls, and Hillary has won 356 of them (97.53% of all matchups), the GOP has won 6 matchups (1.64% of all matchups) and there were 3 absolute ties (0.83%). Virtually all of the Clinton wins have been outside the MoE of any given poll, and the vast majority of them have been +5 or well above. So, yes, it's just one poll, but it is well in-line with 2.5 of electoral polling history. No candidate in our entire history has been polled in the early phase for a presidential cycle as much as Hillary Clinton. That is a historical fact.
2.) Any statement I make, based on the data from the poll, automatically means "according to this PPP (D) poll". I'm not going to write it every time, that just wastes time and energy. Everyone knows that I am referring to this poll.
Why so much information? Well, because no other pollster is willing to release so many internals as PPP.
In a nutshell: Hillary's favorability is bad (-10), but it is so far better than favorability for the majority of GOP candidates so that it is as if it plays no role. In DEM nomination polling, as expected, Hillary easily sweeps the field of 5 declared candidates. In GOP nomination polling, Donald Trump is on top of a very large field. But in GE election polling, where Hillary wins over every single candidate with whom she is matched up, from +3 to +13, Donald Trump does absolutely the worst, at -13 behind Hillary. And in a theoretical 3-way matchup: Clinton / Bush / Trump, she beats Bush by +18 and Trump is almost tied with Bush, ala 1992 and 1912, but with a much, much, much larger margin.
That was the nutshell. Now, here are the specifics.
Clinton:
among Democrats: Fav 69 / unfav 22, +47 (p. 62 of the .pdf)
by gender among Democrats: women - fav 72 / unfav 19, +53, men - fav 65 / unfav 27, +38
However, among all voters, Hillary is deeply underwater: Fav 41 / unfav 51, -10 (p.74, pdf)
That is a bad fav rating to have and normally would indicate that a candidate is losing, and badly at that. This is mostly because while we expect her to triumph in the Liberal vote and get crushed in the conservative vote (which is the case), she is deeply underwater in the moderate vote:
Fav 39 / unfav 48, -9 (p.83, .pdf)
Only, most of the Republican candidates are doing far, far worse than Hillary in favorability ratings among the same general election RVs:
Carson: Fav 31 / unfav 34, -3
Walker: Fav 31 / unfav 38, -7
Rubio: Fav 31 / unfav 40, -9
Fiorina: Fav 21 / unfav 33, -12
Cruz: Fav 27 / unfav 45, -18
Paul: Fav 27 / unfav 45, -18
Huckabee: Fav 27 / unfav 49, -22
Bush: Fav 26 / unfav 52, -26
Trump: Fav 29 / unfav 60, -31
Christie: Fav 19 / unfav 61, -42
So, all of the GOP candidates are also underwater, but 7 of those 10 are doing worse than Hillary. If you think that Trump's fav deficit is the largest, you are wrong. At -31, that is a terrible statistic for any candidate, but Chris Christie is doing far worse, at -42. And the level of deficit does not correspond to any hard data in the general election matchups (see: below). For instance, Carson is only at -3, but still loses to Clinton by 8 points in the GE. Both Paul and Cruz are at -18 in fav/unfav and yet, Paul does very well against Clinton, is usually in low single digits behind her, where Cruz does considerably worse. This tells me that fav/unfav is only one consideration in the voters' mind and that fav/unfav does not translate 1:1 to electability/unelectability.
Back to Hillary's national fav/unfav: 41 / 51, -10. Among women, nationally, all GE voters, she is also underwater, 44 / 48, -4, but remember, she was way up among Democratic women, which means she is getting crushed among Conservative women and probably also underwater among moderate women in order to get to this statistic. Among men, she is at -15 (39 / 54), and yet, she wins every single matchup against all GOP comers and often is also ahead in the men's vote as well. That is an important data point, will be detailed later.
So, fav/unfav ratings only play a subordinate role in all of this. The data shows it very clearly.
You can see the gender internals for all candidates in this rubrik starting on p.86 of the .pdf.
Clinton 57
Sanders 22
undecided 12
Webb 5
Chafee 3
O'Malley 2
Margin: Clinton +35
The polling shows a slight gain for Sanders, but a +35 margin is still a crushing blowout margin and Clinton's topline remains right up there close to the 60 mark, where it has been pretty much all the time. This race is non-competitive.
Trump 19
Walker 17
Bush, Jeb 12
Rubio 10
Carson 10
Huckabee 8
Paul 4
Cruz 4
Fiorina 4
Christie 3
Kasich 3
Undecided 2
Perry 1
Santorum 1
Jindal 1
Gilmore, Graham and Pataki 0
Margin: Trump +2
With a MoE of +/-4.3, this is actually a statistical tie between Trump and Walker. Whether or not this is a post-announcement bounce for Walker remains to be seen. But only five GOP candidates of 17 are in double digits. Paul has lost steam, but Fiorina has actually ticked-up a bit. This is one of the first times that Kasich has been polled and he clocks in at 3 for a start, let's see where he goes.
values in parentheses are from the previous PPP poll, from June 16, 2015, where applicable.
Clinton 45 (47) / Paul 42 (40), margin = Clinton +3 (+7)
Clinton 46 (46) / Rubio 41 (43), margin = Clinton +5 (+3)
Clinton 46 (45) / Bush 41 (41), margin = Clinton +5 (+4)
Clinton 46 (46) / Walker 41 (42), margin = Clinton +5 (+4)
Clinton 46 (47) / Huckabee 40 (42), margin = Clinton +6 (+5)
Clinton 47 (46) / Carson 39 (41), margin = Clinton +8 (+5)
Clinton 48 (48) / Cruz 40 (42), margin = Clinton +8 (+2)
Clinton 46 (45) / Christie 38 (41), margin = Clinton +8 (+4)
Clinton 47 (46) / Fiorina 37 (40), margin = Clinton +10 (+6)
Clinton 50 / Trump 37, margin = Clinton +13
And a three-way matchup, with Trump running as a possible Independent:
Clinton 45 / Bush 25 / Trump 23, margin = Clinton +18
That is the second three-way matchup released in as many days. If we compared that to the ABC/WAPO matchup from July 20th, 2015:
Clinton 46 / Bush 30 / Trump 20, margin = Clinton +16
First, in all ten one-to-one matchups, Hillary wins all of them, and only one win, against Rand Paul, is right at the cusp of the MoE (+/-3.0). She wins with margins from +3 to +13. Three of those margins are +5, 3 of them are +8. 9of those ten margins are better than President Obama's 2012 win, and five of them are also better than Obama's 2008 win. Clinton's winning margin compared to the last PPP (D) has improved in 9 of 10 matchups, the exception being the Clinton/Paul matchup, where Paul comes the closest to her, but has really slipped in GOP nomination polling, the first piece of evidence for the "Trumpian Divide"
The second piece of evidence are Trump's numbers themselves: Trump is at the top of the poll for the Republican nomination, or better put, at a statistical tie with Scott Walker, but is absolutely at the bottom in presidential matchups and this is not the first poll to show this, either. This is now the sixth poll to show Trump far behind Hillary in GE-matchups, and indeed, by a large double digit margin. That being said, this losing margin is smaller than the margins recorded in the Suffolk Poll (-17), the CNN poll (-24), the The FOX NEWS poll (-17), but larger than the yougov/Economist poll (-11).m There was also a Quinnipiac taken 3 weeks before Trump's entry into the race (-18).
And just for comparison, although I am sure that Bernie Sanders (D) is not going to get the Democratic nomination, in a matchup against Trump, Sanders also wins by +10, 47/37. That says something.
In all of these matchups, there is a story to be told in womens' vote vs. the men's vote, and it is a story not quite what people expect to hear. The numbers show that Hillary still has room to grow in the women's vote, where she has often been at +20 above her GOP challengers, but way behind in the men's vote. But here, often, the men's vote margins are smaller than usual, and she is often AHEAD in the men's vote, which is the bad message for the GOP.
Hillary vs. GOP, women's vote vs. mens vote:
vs. Bush: women +9, men -1
vs. Carson: women +11, men +5
vs. Christie: women +12, men +4
vs. Cruz: women +10, men +6
vs. Fiorina: women +16, men +2
vs. Huckabee: women +7, men +5
vs. Paul: women +10, men -4
vs. Rubio: women +7, men +3
vs. Trump: women +21, men +4
vs. Walker: women +11, men -1
Those internals can be found starting on p. 88 of the .pdf.
In only 3 of 10 matchups does Clinton lose in the men's vote. This is critical, because it means that the healthy margins we are seeing for the former Secretary of State are not just being generated by large margins in the women's vote. In fact, 5 of those margins (vs. Bush, Huckabee, Rubio, Cruz and Paul) are SMALLER than Obama's +11 in the womens' vote from 2012. Obama lost in the male vote in 2012. If this statistic maintains over the next year, then the GOP has a real problem on it's hand, because the assumption is that in order to at least make the race competitive, the GOP needs to win the men's vote handily.
What is missing from the matchups, what I would like to have seen, would have been a Clinton / Kasich matchup and still, a Clinton / Perry matchup.
AFTER all matchup questions, PPP asked the following:
33. Do you agree or disagree with the comments Donald Trump made about John McCain over the weekend?
Agree 16 / disagree 60 / don't know 23, margin = disagree +44
So, the long and short of it is that Hillary is still easily cruising to the DEM nomination, the GOP field is still unsettled, but 5 candidates appear to have nestled themselves into the top tier, Clinton easily beats the GOP field in GE matchups and the double paradox exists within the GOP field: the guy likely to come the closest to Hillary is tanking in GOP nomination polling, and the guy who is at the top of GOP polling does the worst against Hillary, aka, the Trumpian Divide.
More internals at the poll. Again, it is just one poll, but it says alot.
http://www.publicpolicypolling.com/pdf/2015/PPP_Release_National_72215.pdf
This .pdf is 108 pages long. I recommend that you download the .pdf and view it with a program like Adobe in order to find the page numbers more quickly.
pp 1-2: General Summary
pp.3-61: GOP nomination
pp. 62-73: DEM nomination
pp. 74-end: General election and assorted additional internals
1,087 RV for the GE matchups, MoE = +/-3.0
Republican nomination: 524 RRV, MoE = +/-4.3
Democratic nomination: 496 DRV, MoE = +/-4.4
There is a lot of information from this poll to glean.
Two caveats, just to save time:
1.) It's just one poll, but it is the 85th national poll since the 1st poll was released on December 5, 2012, more than 2.5 years ago. There have now been, including this PPP (D) poll, a sum-total of 365 head-to-head matchups from the 85 polls, and Hillary has won 356 of them (97.53% of all matchups), the GOP has won 6 matchups (1.64% of all matchups) and there were 3 absolute ties (0.83%). Virtually all of the Clinton wins have been outside the MoE of any given poll, and the vast majority of them have been +5 or well above. So, yes, it's just one poll, but it is well in-line with 2.5 of electoral polling history. No candidate in our entire history has been polled in the early phase for a presidential cycle as much as Hillary Clinton. That is a historical fact.
2.) Any statement I make, based on the data from the poll, automatically means "according to this PPP (D) poll". I'm not going to write it every time, that just wastes time and energy. Everyone knows that I am referring to this poll.
Why so much information? Well, because no other pollster is willing to release so many internals as PPP.
In a nutshell: Hillary's favorability is bad (-10), but it is so far better than favorability for the majority of GOP candidates so that it is as if it plays no role. In DEM nomination polling, as expected, Hillary easily sweeps the field of 5 declared candidates. In GOP nomination polling, Donald Trump is on top of a very large field. But in GE election polling, where Hillary wins over every single candidate with whom she is matched up, from +3 to +13, Donald Trump does absolutely the worst, at -13 behind Hillary. And in a theoretical 3-way matchup: Clinton / Bush / Trump, she beats Bush by +18 and Trump is almost tied with Bush, ala 1992 and 1912, but with a much, much, much larger margin.
That was the nutshell. Now, here are the specifics.
Favorability:
Clinton:
among Democrats: Fav 69 / unfav 22, +47 (p. 62 of the .pdf)
by gender among Democrats: women - fav 72 / unfav 19, +53, men - fav 65 / unfav 27, +38
However, among all voters, Hillary is deeply underwater: Fav 41 / unfav 51, -10 (p.74, pdf)
That is a bad fav rating to have and normally would indicate that a candidate is losing, and badly at that. This is mostly because while we expect her to triumph in the Liberal vote and get crushed in the conservative vote (which is the case), she is deeply underwater in the moderate vote:
Fav 39 / unfav 48, -9 (p.83, .pdf)
Only, most of the Republican candidates are doing far, far worse than Hillary in favorability ratings among the same general election RVs:
Carson: Fav 31 / unfav 34, -3
Walker: Fav 31 / unfav 38, -7
Rubio: Fav 31 / unfav 40, -9
Fiorina: Fav 21 / unfav 33, -12
Cruz: Fav 27 / unfav 45, -18
Paul: Fav 27 / unfav 45, -18
Huckabee: Fav 27 / unfav 49, -22
Bush: Fav 26 / unfav 52, -26
Trump: Fav 29 / unfav 60, -31
Christie: Fav 19 / unfav 61, -42
So, all of the GOP candidates are also underwater, but 7 of those 10 are doing worse than Hillary. If you think that Trump's fav deficit is the largest, you are wrong. At -31, that is a terrible statistic for any candidate, but Chris Christie is doing far worse, at -42. And the level of deficit does not correspond to any hard data in the general election matchups (see: below). For instance, Carson is only at -3, but still loses to Clinton by 8 points in the GE. Both Paul and Cruz are at -18 in fav/unfav and yet, Paul does very well against Clinton, is usually in low single digits behind her, where Cruz does considerably worse. This tells me that fav/unfav is only one consideration in the voters' mind and that fav/unfav does not translate 1:1 to electability/unelectability.
Back to Hillary's national fav/unfav: 41 / 51, -10. Among women, nationally, all GE voters, she is also underwater, 44 / 48, -4, but remember, she was way up among Democratic women, which means she is getting crushed among Conservative women and probably also underwater among moderate women in order to get to this statistic. Among men, she is at -15 (39 / 54), and yet, she wins every single matchup against all GOP comers and often is also ahead in the men's vote as well. That is an important data point, will be detailed later.
So, fav/unfav ratings only play a subordinate role in all of this. The data shows it very clearly.
You can see the gender internals for all candidates in this rubrik starting on p.86 of the .pdf.
DEM nomination:
Clinton 57
Sanders 22
undecided 12
Webb 5
Chafee 3
O'Malley 2
Margin: Clinton +35
The polling shows a slight gain for Sanders, but a +35 margin is still a crushing blowout margin and Clinton's topline remains right up there close to the 60 mark, where it has been pretty much all the time. This race is non-competitive.
GOP nomination:
Trump 19
Walker 17
Bush, Jeb 12
Rubio 10
Carson 10
Huckabee 8
Paul 4
Cruz 4
Fiorina 4
Christie 3
Kasich 3
Undecided 2
Perry 1
Santorum 1
Jindal 1
Gilmore, Graham and Pataki 0
Margin: Trump +2
With a MoE of +/-4.3, this is actually a statistical tie between Trump and Walker. Whether or not this is a post-announcement bounce for Walker remains to be seen. But only five GOP candidates of 17 are in double digits. Paul has lost steam, but Fiorina has actually ticked-up a bit. This is one of the first times that Kasich has been polled and he clocks in at 3 for a start, let's see where he goes.
General Election matchups:
values in parentheses are from the previous PPP poll, from June 16, 2015, where applicable.
Clinton 45 (47) / Paul 42 (40), margin = Clinton +3 (+7)
Clinton 46 (46) / Rubio 41 (43), margin = Clinton +5 (+3)
Clinton 46 (45) / Bush 41 (41), margin = Clinton +5 (+4)
Clinton 46 (46) / Walker 41 (42), margin = Clinton +5 (+4)
Clinton 46 (47) / Huckabee 40 (42), margin = Clinton +6 (+5)
Clinton 47 (46) / Carson 39 (41), margin = Clinton +8 (+5)
Clinton 48 (48) / Cruz 40 (42), margin = Clinton +8 (+2)
Clinton 46 (45) / Christie 38 (41), margin = Clinton +8 (+4)
Clinton 47 (46) / Fiorina 37 (40), margin = Clinton +10 (+6)
Clinton 50 / Trump 37, margin = Clinton +13
And a three-way matchup, with Trump running as a possible Independent:
Clinton 45 / Bush 25 / Trump 23, margin = Clinton +18
That is the second three-way matchup released in as many days. If we compared that to the ABC/WAPO matchup from July 20th, 2015:
Clinton 46 / Bush 30 / Trump 20, margin = Clinton +16
First, in all ten one-to-one matchups, Hillary wins all of them, and only one win, against Rand Paul, is right at the cusp of the MoE (+/-3.0). She wins with margins from +3 to +13. Three of those margins are +5, 3 of them are +8. 9of those ten margins are better than President Obama's 2012 win, and five of them are also better than Obama's 2008 win. Clinton's winning margin compared to the last PPP (D) has improved in 9 of 10 matchups, the exception being the Clinton/Paul matchup, where Paul comes the closest to her, but has really slipped in GOP nomination polling, the first piece of evidence for the "Trumpian Divide"
The second piece of evidence are Trump's numbers themselves: Trump is at the top of the poll for the Republican nomination, or better put, at a statistical tie with Scott Walker, but is absolutely at the bottom in presidential matchups and this is not the first poll to show this, either. This is now the sixth poll to show Trump far behind Hillary in GE-matchups, and indeed, by a large double digit margin. That being said, this losing margin is smaller than the margins recorded in the Suffolk Poll (-17), the CNN poll (-24), the The FOX NEWS poll (-17), but larger than the yougov/Economist poll (-11).m There was also a Quinnipiac taken 3 weeks before Trump's entry into the race (-18).
And just for comparison, although I am sure that Bernie Sanders (D) is not going to get the Democratic nomination, in a matchup against Trump, Sanders also wins by +10, 47/37. That says something.
In all of these matchups, there is a story to be told in womens' vote vs. the men's vote, and it is a story not quite what people expect to hear. The numbers show that Hillary still has room to grow in the women's vote, where she has often been at +20 above her GOP challengers, but way behind in the men's vote. But here, often, the men's vote margins are smaller than usual, and she is often AHEAD in the men's vote, which is the bad message for the GOP.
Hillary vs. GOP, women's vote vs. mens vote:
vs. Bush: women +9, men -1
vs. Carson: women +11, men +5
vs. Christie: women +12, men +4
vs. Cruz: women +10, men +6
vs. Fiorina: women +16, men +2
vs. Huckabee: women +7, men +5
vs. Paul: women +10, men -4
vs. Rubio: women +7, men +3
vs. Trump: women +21, men +4
vs. Walker: women +11, men -1
Those internals can be found starting on p. 88 of the .pdf.
In only 3 of 10 matchups does Clinton lose in the men's vote. This is critical, because it means that the healthy margins we are seeing for the former Secretary of State are not just being generated by large margins in the women's vote. In fact, 5 of those margins (vs. Bush, Huckabee, Rubio, Cruz and Paul) are SMALLER than Obama's +11 in the womens' vote from 2012. Obama lost in the male vote in 2012. If this statistic maintains over the next year, then the GOP has a real problem on it's hand, because the assumption is that in order to at least make the race competitive, the GOP needs to win the men's vote handily.
What is missing from the matchups, what I would like to have seen, would have been a Clinton / Kasich matchup and still, a Clinton / Perry matchup.
AFTER all matchup questions, PPP asked the following:
33. Do you agree or disagree with the comments Donald Trump made about John McCain over the weekend?
Agree 16 / disagree 60 / don't know 23, margin = disagree +44
So, the long and short of it is that Hillary is still easily cruising to the DEM nomination, the GOP field is still unsettled, but 5 candidates appear to have nestled themselves into the top tier, Clinton easily beats the GOP field in GE matchups and the double paradox exists within the GOP field: the guy likely to come the closest to Hillary is tanking in GOP nomination polling, and the guy who is at the top of GOP polling does the worst against Hillary, aka, the Trumpian Divide.
More internals at the poll. Again, it is just one poll, but it says alot.
Last edited:
