National Democratic Polls, Low-End Estimate
by Chris Bowers, Fri Apr 27, 2007 at 12:03:09 AM EDT
Yesterday morning, I produced a chart of the rolling averages of national Democratic nomination preference polls from February 23rd through April 24th. I thought about that chart a lot today, and considered different ways that polling averages of this sort could be produced. The result of that thought is a second chart, meant to be a companion of the first chart, that provides a lower estimation of support of all three top-tier candidates.
National Democratic Polls, Low-End Estimate, 2/1-4/24
This chart differs from the first chart in two key ways. First, it incorporates all polls that include the entire, announced Democratic field in their nomination preference question, including those polls that put Al Gore in the question. Second, when a given poll offers both Gore-inclusive and Gore-exclusive results, this chart utilizes the Gore-inclusive data. There are many benefits to this methodology:
National Democratic Polls, Low-End Estimate, 2/1-4/24
| Date | Clinton | Obama | Edwards | Other / Unsure |
|---|---|---|---|---|
| Apr 24 | 34.3% | 25.9% | 16.0% | 23.8% |
| Apr 20 | 34.0% | 25.1% | 15.0% | 25.9% |
| Apr 16 | 33.9% | 24.6% | 15.1% | 26.4% |
| Apr 13 | 34.9% | 24.0% | 15.7% | 25.4% |
| Apr 10 | 35.8% | 22.0% | 15.5% | 26.7% |
| Apr 06 | 35.4% | 22.9% | 15.4% | 26.3% |
| Apr 02 | 35.4% | 22.9% | 15.4% | 26.3% |
| Mar 30 | 34.8% | 23.6% | 15.3% | 26.4% |
| Mar 27 | 34.9% | 25.0% | 14.3% | 25.9% |
| Mar 23 | 35.8% | 25.8% | 12.5% | 26.0% |
| Mar 20 | 36.4% | 25.0% | 11.6% | 27.0% |
| Mar 16 | 36.5% | 25.8% | 12.1% | 25.6% |
| Mar 13 | 36.1% | 25.0% | 12.6% | 26.3% |
| Mar 09 | 36.2% | 25.0% | 12.7% | 26.1% |
| Mar 06 | 35.7% | 25.4% | 12.7% | 26.1% |
| Mar 02 | 35.6% | 24.0% | 11.2% | 29.2% |
| Feb 27 | 35.0% | 22.4% | 10.2% | 32.4% |
| Feb 24 | 35.8% | 21.3% | 10.3% | 32.8% |
| Feb 20 | 36.0% | 20.0% | 10.7% | 33.3% |
| Feb 16 | 35.3% | 19.2% | 11.3% | 34.2% |
| Feb 12 | 38.0% | 17.8% | 11.6% | 32.6% |
| Feb 09 | 35.0% | 18.7% | 11.7% | 34.7% |
| Feb 05 | 36.7% | 17.3% | 10.7% | 34.7% |
| Feb 01 | 37.3% | 17.3% | 12.7% | 32.7% |
This chart differs from the first chart in two key ways. First, it incorporates all polls that include the entire, announced Democratic field in their nomination preference question, including those polls that put Al Gore in the question. Second, when a given poll offers both Gore-inclusive and Gore-exclusive results, this chart utilizes the Gore-inclusive data. There are many benefits to this methodology:
- More data. This chart only excludes polls that ask two-way or three way questions, and as such increases the total data points from 33 to 43.
- Longer trend lines. The increased number of data points allows for a trendline going back to February 1st, rather than February 23rd.
- Less Soft Support. This is the most important difference. Supporters of Gore are placed into the "other / unsure" category. This is justifiable on the grounds that if "Poll Participant X" would choose Gore instead of a "Non-Gore Candidate Y" simply because Gore is included in the question, then "Poll Participant X" clearly is a very, very soft supporter of "Non-Gore Candidate Y." As I have long railed about the excessive pushing of undecideds and leaners in Democratic nomination polls this far from the actual primary / caucus season, this is actually a useful way to up the percentage of "other / unsure" even in polls like Gallup that push undecideds way too hard. At this point, it seems obvious that those who at this point would still support Gore are clearly undecided about the rest of the field (or, as I will discuss tomorrow, not yet engaged in the campaign). By contrast, those who support "Non-Gore Candidate Y" when gore is included in the question are harder supporters of "Non-Gore Candidate Y." thus, this chart produces a lower, but harder, estimate of support for all candidates. Even though I don't like unannounced candidates being included in poll questions, this methodology is a means of making lemonade out of lemon polls.
Tags: Democrats, national Democratic poll average, polls, President 2008, Primary Elections (all tags)
-----------
myDD - skin
-----------
17 Comments