In what should be the last poll released prior to tonight's showdown, the C-SPAN/Reuters/Zogby Tracking Poll suggests that the conventional wisdom on the ground may be true.
Clinton
12/30-1/2: 24
12/29-1/1: 28
12/28-12/31: 30
12/27-12/30: 30
12/26-12/29: 31
Obama
12/30-1/2: 31
12/29-1/1: 28
12/28-12/31: 26
12/27-12/30: 26
12/26-12/29: 27
Edwards
12/30-1/2: 27
12/29-1/1: 26
12/28-12/31: 25
12/27-12/30: 26
12/26-12/29: 24
Keep in mind that tracking polls take into account multiple days of data; this one includes four days. Thus, a candidate's support must increase or decrease significantly in the newest day of polling to show any movement in the overall number. Tracking polls are designed to be stable; a point or two of movement (called statistical noise) shouldn't move the average too much unless a trend is clear (which, of course, is the point).
Here's a possible version of Hillary's numbers over the past few days. In five days, she dropped 7 points, including a 4 point drop from the 1st to the 2nd. Using the released numbers from Zogby, I figured out what set of data was necessary to produce a number that would round to the released data. Please note that 1) these are just an educated guess and 2) Zogby rounds numbers (.5 and .75 rounds up, .25 rounds down).
Again, these are simply an estimate, but they illustrate quite nicely the fall in support necessary to produce the 4 point single day drop. If anyone else wants to experiment with the numbers and post what you come up with, feel free. Edwards steady increase is more steady, while Obama's 3 point jump likely mirrors the large movement present in Hillary's numbers (but obviously, in reverse). Feel free to figure out the Obama/Edwards daily numbers and post them!
On the ground sources as well as various media outlets agree that Obama and Edwards are surging, and Clinton seems to be falling off. The Des Moines Register Poll from a few days ago showed Obama leading by 7 points, but was criticized by many because it seemed to be so far out of line with what other polls suggest, particularly due to the fact it suggested that 2 in every 5 Democratic caucusgoers will be independents (which would double the amount of independent participation from 2004). The DMR stood by the numbers.
If the Zogby tracking poll is correct (and it may be important to note that Zogby isn't considered one of the most reliable polls), it would suggest that the DMR Poll's finding may be correct in pointing to an Obama win this evening. Even if these exact numbers are wrong, they do point toward movement toward Obama and Edwards and away from Clinton, which fits into both the current media narrative as well as the polling data from the last few days. If the caucuses are down to an Obama-Edwards contest, which candidate prevails could depend on how 2nd-tier supporters move after the first round of voting.
The one thing I'd add to the previous posts about 2nd tier voting is that conjectures of how supporters of a certain candidate will move based on similarities between candidates are dangerous. Assuming Biden voters will move to Clinton because of "experience" assumes that Biden voters like him because of his experience AND view Clinton as experienced AND value experience more than all other issues across the entire field of candidates. All of these assumptions lack quantitative evidence. I would suggest a safer way to look at this question is through the polls that have asked about 2nd-choice support. Not surprisingly, Obama and Edwards are the leading 2nd choices (with Edwards slightly more favored than Obama). Clinton trails distantly in 3rd place.
I hope you find this statistical geek-ery interesting as I do. I love breaking down polling data. If you want a class where this stuff is gone into in depth, I'd recommend taking a course with Prof. Charles Franklin (particularly 218) in the Political Science Departmet.
Also, my predictions:
Dem
1) Obama, 34
2) Edwards, 31
3) Clinton, 28
4) Biden, 4
5) Richardson, 3
6) Dodd, 1
(I'm an Obama supporter - that isn't influencing my prediction.)
GOP
1) Romney
2) Huckabee
3) McCain
4) Thompson
Stay posted for more updates from your fellow College Dems. It's going down to the wire!
Here's the tracking data:Clinton, Romney dip as Ron Paul gains
UTICA, New York – Democrat Barack Obama continued his upward momentum through the evening before the Iowa caucuses, capturing the lead ahead of rivals John Edwards and Hillary Clinton...
Obama broke through the 30% barrier for the first time, gaining 31% support after another strong day leading up to the caucuses. But more dramatic was Clinton’s four-point drop in this last day of tracking. Edwards moved into second place by himself after another day where he steadily gained ground.
Clinton
12/30-1/2: 24
12/29-1/1: 28
12/28-12/31: 30
12/27-12/30: 30
12/26-12/29: 31
Obama
12/30-1/2: 31
12/29-1/1: 28
12/28-12/31: 26
12/27-12/30: 26
12/26-12/29: 27
Edwards
12/30-1/2: 27
12/29-1/1: 26
12/28-12/31: 25
12/27-12/30: 26
12/26-12/29: 24
Keep in mind that tracking polls take into account multiple days of data; this one includes four days. Thus, a candidate's support must increase or decrease significantly in the newest day of polling to show any movement in the overall number. Tracking polls are designed to be stable; a point or two of movement (called statistical noise) shouldn't move the average too much unless a trend is clear (which, of course, is the point).
Here's a possible version of Hillary's numbers over the past few days. In five days, she dropped 7 points, including a 4 point drop from the 1st to the 2nd. Using the released numbers from Zogby, I figured out what set of data was necessary to produce a number that would round to the released data. Please note that 1) these are just an educated guess and 2) Zogby rounds numbers (.5 and .75 rounds up, .25 rounds down).
HRC Daily Polling Data (Four Day Rolling Average)
12/26: 31
12/27: 30
12/28: 31
12/29: 32 (31.00 - 31, released 12/30)
12/30: 28 (30.25 - 30, released 12/31)
12/31: 27 (29.50 - 30, released 1/1)
1/1: 23 (27.50 - 28, released 1/2)
1/2: 19 (24.25 - 24, released 1/3)
Again, these are simply an estimate, but they illustrate quite nicely the fall in support necessary to produce the 4 point single day drop. If anyone else wants to experiment with the numbers and post what you come up with, feel free. Edwards steady increase is more steady, while Obama's 3 point jump likely mirrors the large movement present in Hillary's numbers (but obviously, in reverse). Feel free to figure out the Obama/Edwards daily numbers and post them!
On the ground sources as well as various media outlets agree that Obama and Edwards are surging, and Clinton seems to be falling off. The Des Moines Register Poll from a few days ago showed Obama leading by 7 points, but was criticized by many because it seemed to be so far out of line with what other polls suggest, particularly due to the fact it suggested that 2 in every 5 Democratic caucusgoers will be independents (which would double the amount of independent participation from 2004). The DMR stood by the numbers.
If the Zogby tracking poll is correct (and it may be important to note that Zogby isn't considered one of the most reliable polls), it would suggest that the DMR Poll's finding may be correct in pointing to an Obama win this evening. Even if these exact numbers are wrong, they do point toward movement toward Obama and Edwards and away from Clinton, which fits into both the current media narrative as well as the polling data from the last few days. If the caucuses are down to an Obama-Edwards contest, which candidate prevails could depend on how 2nd-tier supporters move after the first round of voting.
The one thing I'd add to the previous posts about 2nd tier voting is that conjectures of how supporters of a certain candidate will move based on similarities between candidates are dangerous. Assuming Biden voters will move to Clinton because of "experience" assumes that Biden voters like him because of his experience AND view Clinton as experienced AND value experience more than all other issues across the entire field of candidates. All of these assumptions lack quantitative evidence. I would suggest a safer way to look at this question is through the polls that have asked about 2nd-choice support. Not surprisingly, Obama and Edwards are the leading 2nd choices (with Edwards slightly more favored than Obama). Clinton trails distantly in 3rd place.
I hope you find this statistical geek-ery interesting as I do. I love breaking down polling data. If you want a class where this stuff is gone into in depth, I'd recommend taking a course with Prof. Charles Franklin (particularly 218) in the Political Science Departmet.
Also, my predictions:
Dem
1) Obama, 34
2) Edwards, 31
3) Clinton, 28
4) Biden, 4
5) Richardson, 3
6) Dodd, 1
(I'm an Obama supporter - that isn't influencing my prediction.)
GOP
1) Romney
2) Huckabee
3) McCain
4) Thompson
Stay posted for more updates from your fellow College Dems. It's going down to the wire!
1 Comments:
haha, cmon
"(I'm an Obama supporter - that isn't influencing my prediction.)"
Nice call, though.
Post a Comment
<< Home