Primary Season Polling Wishlist
by Chris Bowers, Tue Feb 20, 2007 at 06:00:09 PM EST
Over the past week, several new national primary polls have been released. Here are the ones for Demcorats:
Latest National Polls
| Candidate |
Marist, 2/15 |
Rasmussen, 2/15 |
Gallup, 2/11 |
Siena, 2/9 |
Rasmussen, 2/8 |
| Clinton |
37 |
28 |
40 |
45 |
28 |
| Obama |
17 |
24 |
21 |
12 |
23 |
| Edwards |
11 |
11 |
13 |
10 |
13 |
| Gore |
11 |
11 |
10 |
10 |
8 |
| Others |
7 |
?? |
8 |
4 |
?? |
| Undecided |
17 |
?? |
3 |
20 |
?? |
These polls are interesting. However, they are also deeply flawed. In the hopes of finding better public polling for the 2008 primary season, I want to make a few small requests to the news organizations commissioning the polls, and to the firms conducting the polls:
- Stop pushing undecideds to make up their minds. I am looking in your direction, Gallup. To produce a poll that shows only 3% of the Democratic electorate as undecided at this point is obviously absurd to the point of shameful. Support for all candidates right now is extremely soft, and as such there should be no attempts whatsoever to force the people who respond to your poll to choose a candidate at this time. If you want to provide an accurate snapshot of current public opinion, you simply can't push undecideds at this point.
- Include all candidates who are running. I am looking at you, Survey USA and Siena. Leaving announced candidates out of your questions is basically an in-kind contribution to the candidates you included in the poll. Why should some candidates, and not others, receive free polling information? This also distorts public opinion, in that voters will see all names on the ballot when they go to vote, and in that it artificially inflates the results for the candidates who are included in the polls. This is really bad stuff.
- Stop including candidates who have not yet declared they are running. Virtually everyone is guilty of this. Another major way to distort public opinion at this time is to lit several candidates who have not yet declared they are running as options in your polls. Right now, everyone is including Clark and Gore, and some polls are including Sharpton. Not only does this distort public opinion because it offers voters choices they do not have, it is not fair to the candidates who are running because it artificially deflates their results. When other candidates declare, then add them. Until that time, stick only to announced candidates.
- Conduct at least as many polls of early states as you conduct nationally. There is no "national primary," so polling it is of questionable value. Further, even though over a fifteen states, including several large states, are lining up for primaries or caucuses on February 5th, we all know that candidate support in states after Iowa and New Hampshire will be heavily influenced by election results in Iowa and New Hampshire. Thus, polls of Iowa and New Hampshire, which are no more expensive than national polls, are actually more useful and interesting at this time. So, please, slow down on the national polls, and give us more insight into the early states.
- Give us favorable ratings and name recognition numbers, too. Trial heats are both fun and interesting. However, it is useful to know candidate potential, not just current candidate standing. Without favorable / unfavorable numbers, as well as name recognition numbers, it is very difficult to know what polls this far out actually mean. Providing these two metrics would give your polls a lot more context, thus making them all that much more meaningful.
- More bells and whistles. I recognize that this request might exceed the budgets of a few polls, but it would also be great to include metrics such as second choices, likely voter methodology, and the strength of candidate support. While expensive, all of this would provide us with a deeper picture of the campaign. Also, personally I would rather have fewer, high quality polls than more, low-quality ones.
There is great potential for informative, quality polling in the 2007-2008 cycle. I think it would be best if we addressed ways it current polling could be improved now, so as to avoid annoying fights over polling methodology as we near election time. Further, producing polls we can all trust would be a tremendous public service, not to mention improve the reputation of the organizations who produce and publish them.
Tags: Polls, primary elections, Democrats, President 2008 (all tags)
You are not logged in.
In order to post a comment, you must be logged in. If you have a member account, please log in to comment.
If not, you can make an account right here. It's quick and free.
