I started by trying to figure out the bias averages of the national polls (OK, I really wanted to try to figure out the bias in state polls, but there just wasn’t a large enough sample to do that). Along the way, though, I found out some other interesting things, and being the gabby sort, I decided to share. Some of it might be useful in examining the polls, as we enter the post-Convention stretch.
I began by noting the 2-way poll results from Polling Report and Real Clear Politics from February 3 to the end of July. I made a spreadsheet with the name of the group or agency which conducted the poll, the starting and ending dates of the polling, and the results for Bush and Kerry. From this information, I drew up another spreadsheet which produced the daily average for the polls. I then compared each poll against the average for the days polled, to see where each poll stood against the average, noting whether Bush or Kerry was getting more or less support than the average in a poll. I then was able to apply that information to determine the average bias, the aggregate variance, and essentially the stability of each poll.
I want to stop at this point, and make an observation about the fairness of these polls. As polls make statements which favor Bush or Kerry, it’s becoming common to see this poll or that dismissed as being unreliable or unbelievable. This is unwise, because in examining the polls presented, I was struck by how well, overall, every poll was conducted. While I have concerns about the demographics base used, and the weighting assumptions inherent in modern polling, the conclusions of each poll were consistent with other polls at the same time. Since thirty-one national agencies or groups have conducted national polls cited in the two sites, such consistency demonstrates either a massive conspiracy, or an impressive level of professional standards. However, within that acknowledgment, there were some clear trends and characteristics, which I think are worth mentioning.
As I said, there were thirty-one separate groups or agencies which conducted polls. Of these, eight conducted only one poll this year (A McKenna, Public Integrity Project, Mother Jones, McLaughlin, University of Connecticut, NBC News, Sacred Heart, and A McHenry) [Of those, U-Conn conducted polls with support from a number of news groups, and NBC News joined the Wall Street Journal for later polling]. Another three groups conducted only two polls (Time/CNN, which split from each other and reformed with other groups for polling, Insider Advantage, and the Battleground Surveys). I have not considered these groups for the purposes of comparative bias, because there just are not enough polls from them to determine a real trend or pattern.
That leaves another twenty groups, however, which have conducted a steady flow of polls. For this report, I will note the five polls which most favor Bush, the five which are least favorable to Bush, the five which are most favorable to Kerry, the five which are least favorable to Kerry, and the five which are, as an aggregate, the least likely to be off the average measure for that date. But before I get to that, I will note here five individual polls which are notable for unusual results.
The Rasmussen polls for April 12-14 and 13-15 were uniquely favorable to President Bush; on those dates the average poll was 40-43 in favor of Kerry, but Rasmussen’s polls for those dates were 46-45 Bush and 46-44 Bush, a bias in favor of Bush.
On the other hand, the Newsweek poll for July 30-31 showed 42-49 favoring Kerry, at a time when the average was 47-48 for Kerry. This was a 5-point bias against Bush.
The CNN/USA Today/Gallup poll for February 16-17 showed 43-55 for Kerry, at a time when the average poll was 45-47 Kerry, which was an 8-point bias in favor of Kerry.
The Battleground poll for March 28-31 showed 43-39 for Bush, at a time when the average was 45-45 tie, a 6-point bias against Kerry.
These five polls do not mean these groups are biased overall, but represent the extremes found when I reviewed the polls taken between February 3 and July 31.
Here are the top five in each of the five categories, then:
Bush Favorable
1. Democracy Corps (+1.8)
2. NPR (+1.7)
3t. NBC/Wall Street Journal (+1.6)
3t. CNN/USA Today/Gallup (+1.6)
5t. Time (+1.0)
5t. Harris (+1.0)
Bush Unfavorable
1. CBS News (no partner) (-3.5)
2. Investor Business Daily (-1.8)
3t. Zogby (-1.6)
3t. Fox News (-1.6)
5. Newsweek (-1.3)
Kerry Favorable
1. LA Times (+2.7)
2t. Time/CNN (+2.5)
2t. CBS News (no partner) (+2.5)
4. CNN/USA Today/Gallup (+1.9)
5t. ARG (+1.7)
5t. Time (+1.7)
Kerry Unfavorable
1. Harris (-2.0)
2. AP/Ipsos (-1.7)
3t. Fox News (-1.4)
3t. Investor Business Daily (-1.4)
5t. Marist (-1.0)
5t. McLaughlin (-1.0)
Aggregate Stability
1. Marist (1.5 avg total variance)
2t. NPR (2.3 avg total variance)
2t. CBS/NY Times (2.3 avg total variance)
4. ARG (2.4 avg total variance)
5. AP/Ipsos (2.5 avg total variance)
I will let the reader draw his/her own conclusions, after making three suggestions:
1. Notice which groups are favorable or unfavorable to both Bush and Kerry
2. Aggregate Stability is a near-unknown quality, but represents a reason to have great confidence in a poll. If a poll does not miss the average by very much, it represents a reliable place for quick vitals.
3. Most of the variance shown is within a reasonable Margin of Error. Without naming the obvious name, however, one group stands out for two reasons. First, this one group is hostile to Bush by more than its Margin of Error, and is favorable to Kerry by almost its Margin of Error, on a regular basis. This group is also unique, because instead of doing its own polls, then joining a partner, this group was teamed up with another media group, but then began doing additional polls on its own, apparently out of displeasure from its partner publishing internals and demographics, which this group conceals.
THE POLLS
Who..............Var Bush..........Var Kerry..........Agg
Marist....................+0.5.....................-1......................1.5
Quinnipiac................-0.5.....................-1.2.....................2.0
NPR.......................+1.7.....................+0......................2.3
CBS/NYT...................-1.1.....................+0.7....................2.3
ARG........................-0.7.....................+1.7....................2.4
AP/Ipsos..................+0.5.....................-1.7....................2.5
ABC/WP....................+1.......................-0.4....................2.6
NBC/WSJ...................+1.6.....................-0.6....................2.6
Pew.......................-1.......................+0.3....................2.6
Time......................+1.......................+1.7....................2.7
Zogby.....................-1.6.....................-0.4....................2.8
Rasmussen.................+0.6.....................+0.1....................3.0
Harris....................+1.......................-2......................3.0
Gallup....................-0.1.....................+1.7....................3.0
LA Times..................+0.......................+2.7....................3.3
Newsweek..................-1.3.....................+0.7....................3.4
IBD/CSM...................-1.8.....................-1.4....................3.7
Democracy Corps...........+1.8......................+2......................3.8
C/U/Gallup................+1.6......................+1.9....................4.1
Fox News..................-1.6......................-1.4....................5.0
CBS News..................-3.5.....................+2.5....................6.0
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment