Statistikhengst
Diamond Member
- Banned
- #1
IPSOS/Reuters is now doing doing continuous 5-day rolling polling vis-a-vis both GOP and DEM nominations and is providing its results in double sets:
-the entire crop of candidates
-just the top three
I've already reported it here:
Statistikhengst's ELECTORAL POLITICS - 2015 and beyond: 2016 polling round-up, 22-31.08.2015: DEM Nomination
-and here-
Statistikhengst's ELECTORAL POLITICS - 2015 and beyond: 2016 polling round-up, 22-31.08.2015: GOP Nomination
Here is the direct link to the .pdf:
http://www.ipsos-na.com/download/pr.aspx?id=14798
Overall: 1,152 Adults, MoE = +/-3.3
of them, 415 RRV, MoE = +/-5.5
Margin: Trump +20 (total) Trump +20 (RRV only), Trump +24 (IRV only)
3-Way race: IPSOS
Margin: Trump +23 (total) Trump +24 (RRV only), Trump +25 (IRV only)
Overall: 1,152 Adults, MoE = +/-3.3
of them, 445 DRV, MoE = +/-5.2
Margin: Clinton +22 (all), Clinton +26 (DRV only), Clinton +14 (IDV only)
TOP-3, IPSOS:
Margin: Clinton +23 (all), Clinton +28 (DRV only), Clinton +15 (IDV only)
Within the respective nominations and between them both, there is a lot of consistency right now.
Let's take a look and compare.
GOP nomination, all candidates: Trump +20 (total) Trump +20 (RRV only), Trump +24 (IRV only)
GOP nomination, upper three: Trump +23 (total) Trump +24 (RRV only), Trump +25 (IRV only)
DEM nomination, all candidates: Clinton +22 (all), Clinton +26 (DRV only), Clinton +14 (IDV only)
DEM nomination, upper three: Clinton +23 (all), Clinton +28 (DRV only), Clinton +15 (IDV only)
First, both Trump and Clinton are at +20 and more in overall polling and among RVs of their respective parties only, both in overall races and in the 3-way race. The difference, and this is a key difference, is that Trump's margin actually increases among Independent voters who are R leaning, while Clinton's margin dramatically decreased amond Independent voters who are D leaning. This is good and bad for both, depending on how you want to look at it, but only if the figures stay this way forever, and they probably wont.
Let's put them next to each other:
GOP nomination, all candidates: Trump +20 (total) Trump +20 (RRV only), Trump +24 (IRV only)
DEM nomination, all candidates: Clinton +22 (all), Clinton +26 (DRV only), Clinton +14 (IDV only)
GOP nomination, upper three: Trump +23 (total) Trump +24 (RRV only), Trump +25 (IRV only)
DEM nomination, upper three: Clinton +23 (all), Clinton +28 (DRV only), Clinton +15 (IDV only)
Among all voters in their respective survey groups, Clinton is only 2 points stronger in margin than Trump, but among DRV vs. RRV, she is 6 points stronger in margin than Trump. However, among IRV vs. IDV, Trump is a full 10 points stronger on margin than Clinton.
Same story for a 3-man only race: Among all voters in their respective survey groups, Clinton and Trump have exactly the same margin among their respective groups (+23), but among DRV vs. RRV, she is 4 points stronger. However, among IRV vs. IDV, Trump is a full 10 points stronger on margin than Clinton.
It should also be noted that IPSOS/Reuters is polling two other Democrats who are definitely not in the race and are not going to enter: Gillibrand and Cuomo, and it is more than likely that their votes would go for Clinton, so in reality, her numbers are just ever so slightly understated here.
This means that Clinton's support is being generated more by actual Democrats who are more likely to actually go and vote than by Independents who may or may not go vote. And for Trump, his support is being generated somewhat more by Independent voters who lean R, who may or may not go vote.
The good message for both of these candidates is that this kind of margin looks an awful lot like George H.W. Bush's margins in late 1987 polling over Dole, Robertson, Kemp and DuPont, also a lot like George W. Bush's late 1999 polling margins over McCain, Forbes and Keyes. Those margins lead to massive primary victories when all was said and done.
In a 4 or 5 way race, regardless whether the leading candidate is at 50% or well under 50%, a +20 margin increases the likeliness that a candidate wins pretty much every congressional district within that state, which means that in a primary that divvies up the delegates according to proportional representation, that candidate is still likely to get most all of the delegates (excluding Superdelegates). And in a WTA contest, a +20 absolutely guarantees all of the delegates. This is why a candidate who is under 50% but with such a healthy margin over the others can still win the vast majority of delegates and sail easily toward the nomination.
On the other hand, as in Trump's case, a campaign that is getting more support from Independents who lean toward his party means that he may not do as well in states that do strictly closed primaries. However, at +20 even among R's only, Trump is also guaranteed some massive victories, if these numbers hold through the primary season.
One more thing: some partisans keep saying that Hillary Clinton is struggling. She is not. A +26 to +28 margin is an absolutely CRUSHING margin. And it's kind of hard to say that she is struggling when Trump's margins are very close to hers. The difference is she came from absolutely sky-high margins of +50 or more when no real challengers were on the horizon to the reality of having at least one, maybe two major challengers. That mathematically changes things.
On the other side, Trump came out of nowhere to sweep the GOP field, a surprise that I suspect no one was expecting to see.
One more detail. Look at the % of undecideds under Democrats only or Republicans only: 9% and 7% for Democrats, 10% and 8% for Republicans. That is a very small number for this stage of the game. This means that the field is settling more quickly than I suspected would happen. That is bad news for other prospective GOP candidates, especially Jeb Bush, whose numbers appear to be tanking badly.
And finally, I think it is helpful that IPSOS does full group polling and then does polling for a three-man only race, because as the primaries develop, the chance that the majority of lower tier candidates drop out is very large, anyway.
This is just one poll (a caveat I mention very often), but the data in it is indeed very telling.
IPSOS, as an independent polling organization, started in France in 1975, is now 40 years old and has no dog in the race. Its 2012 presidential polling had a mathematical bias somewhat to the Right:
Statistikhengst's ELECTORAL POLITICS - 2015 and beyond: End Polling Table - USA - alphabetical by pollster name
In 2012, Ipsos/Reuters put out final polls in 4 critical battleground states. It got the call in 3 of those 4 right, was was off to the Right in margin estimate in 3 of 4. Most pollsters missed Florida, so this is not a huge surprise.
It's mathematical bias for 2012, based on its end polling, was: R +1.75
That's a pretty decent figure.
FYI.
-the entire crop of candidates
-just the top three
I've already reported it here:
Statistikhengst's ELECTORAL POLITICS - 2015 and beyond: 2016 polling round-up, 22-31.08.2015: DEM Nomination
-and here-
Statistikhengst's ELECTORAL POLITICS - 2015 and beyond: 2016 polling round-up, 22-31.08.2015: GOP Nomination
Here is the direct link to the .pdf:
http://www.ipsos-na.com/download/pr.aspx?id=14798
GOP nomination:
Overall: 1,152 Adults, MoE = +/-3.3
of them, 415 RRV, MoE = +/-5.5
Margin: Trump +20 (total) Trump +20 (RRV only), Trump +24 (IRV only)
3-Way race: IPSOS
Margin: Trump +23 (total) Trump +24 (RRV only), Trump +25 (IRV only)
DEM nomination:
Overall: 1,152 Adults, MoE = +/-3.3
of them, 445 DRV, MoE = +/-5.2
Margin: Clinton +22 (all), Clinton +26 (DRV only), Clinton +14 (IDV only)
TOP-3, IPSOS:
Margin: Clinton +23 (all), Clinton +28 (DRV only), Clinton +15 (IDV only)
Within the respective nominations and between them both, there is a lot of consistency right now.
Let's take a look and compare.
GOP nomination, all candidates: Trump +20 (total) Trump +20 (RRV only), Trump +24 (IRV only)
GOP nomination, upper three: Trump +23 (total) Trump +24 (RRV only), Trump +25 (IRV only)
DEM nomination, all candidates: Clinton +22 (all), Clinton +26 (DRV only), Clinton +14 (IDV only)
DEM nomination, upper three: Clinton +23 (all), Clinton +28 (DRV only), Clinton +15 (IDV only)
First, both Trump and Clinton are at +20 and more in overall polling and among RVs of their respective parties only, both in overall races and in the 3-way race. The difference, and this is a key difference, is that Trump's margin actually increases among Independent voters who are R leaning, while Clinton's margin dramatically decreased amond Independent voters who are D leaning. This is good and bad for both, depending on how you want to look at it, but only if the figures stay this way forever, and they probably wont.
Let's put them next to each other:
GOP nomination, all candidates: Trump +20 (total) Trump +20 (RRV only), Trump +24 (IRV only)
DEM nomination, all candidates: Clinton +22 (all), Clinton +26 (DRV only), Clinton +14 (IDV only)
GOP nomination, upper three: Trump +23 (total) Trump +24 (RRV only), Trump +25 (IRV only)
DEM nomination, upper three: Clinton +23 (all), Clinton +28 (DRV only), Clinton +15 (IDV only)
Among all voters in their respective survey groups, Clinton is only 2 points stronger in margin than Trump, but among DRV vs. RRV, she is 6 points stronger in margin than Trump. However, among IRV vs. IDV, Trump is a full 10 points stronger on margin than Clinton.
Same story for a 3-man only race: Among all voters in their respective survey groups, Clinton and Trump have exactly the same margin among their respective groups (+23), but among DRV vs. RRV, she is 4 points stronger. However, among IRV vs. IDV, Trump is a full 10 points stronger on margin than Clinton.
It should also be noted that IPSOS/Reuters is polling two other Democrats who are definitely not in the race and are not going to enter: Gillibrand and Cuomo, and it is more than likely that their votes would go for Clinton, so in reality, her numbers are just ever so slightly understated here.
This means that Clinton's support is being generated more by actual Democrats who are more likely to actually go and vote than by Independents who may or may not go vote. And for Trump, his support is being generated somewhat more by Independent voters who lean R, who may or may not go vote.
The good message for both of these candidates is that this kind of margin looks an awful lot like George H.W. Bush's margins in late 1987 polling over Dole, Robertson, Kemp and DuPont, also a lot like George W. Bush's late 1999 polling margins over McCain, Forbes and Keyes. Those margins lead to massive primary victories when all was said and done.
In a 4 or 5 way race, regardless whether the leading candidate is at 50% or well under 50%, a +20 margin increases the likeliness that a candidate wins pretty much every congressional district within that state, which means that in a primary that divvies up the delegates according to proportional representation, that candidate is still likely to get most all of the delegates (excluding Superdelegates). And in a WTA contest, a +20 absolutely guarantees all of the delegates. This is why a candidate who is under 50% but with such a healthy margin over the others can still win the vast majority of delegates and sail easily toward the nomination.
On the other hand, as in Trump's case, a campaign that is getting more support from Independents who lean toward his party means that he may not do as well in states that do strictly closed primaries. However, at +20 even among R's only, Trump is also guaranteed some massive victories, if these numbers hold through the primary season.
One more thing: some partisans keep saying that Hillary Clinton is struggling. She is not. A +26 to +28 margin is an absolutely CRUSHING margin. And it's kind of hard to say that she is struggling when Trump's margins are very close to hers. The difference is she came from absolutely sky-high margins of +50 or more when no real challengers were on the horizon to the reality of having at least one, maybe two major challengers. That mathematically changes things.
On the other side, Trump came out of nowhere to sweep the GOP field, a surprise that I suspect no one was expecting to see.
One more detail. Look at the % of undecideds under Democrats only or Republicans only: 9% and 7% for Democrats, 10% and 8% for Republicans. That is a very small number for this stage of the game. This means that the field is settling more quickly than I suspected would happen. That is bad news for other prospective GOP candidates, especially Jeb Bush, whose numbers appear to be tanking badly.
And finally, I think it is helpful that IPSOS does full group polling and then does polling for a three-man only race, because as the primaries develop, the chance that the majority of lower tier candidates drop out is very large, anyway.
This is just one poll (a caveat I mention very often), but the data in it is indeed very telling.
IPSOS, as an independent polling organization, started in France in 1975, is now 40 years old and has no dog in the race. Its 2012 presidential polling had a mathematical bias somewhat to the Right:
Statistikhengst's ELECTORAL POLITICS - 2015 and beyond: End Polling Table - USA - alphabetical by pollster name
In 2012, Ipsos/Reuters put out final polls in 4 critical battleground states. It got the call in 3 of those 4 right, was was off to the Right in margin estimate in 3 of 4. Most pollsters missed Florida, so this is not a huge surprise.
It's mathematical bias for 2012, based on its end polling, was: R +1.75
That's a pretty decent figure.
FYI.
Last edited:
