October 06, 2004
(Updated) Mass Confusion Among Polls In Mayoral Race
'Tribune' Numbers Don't Match Potter And 'Oregonian' Numbers
Note: This post has been updated. Any and all updates appear at the end of the original post.
There appears to be some serious weirdness going on in the world of polls in the Mayoral race between Jim Francesconi and Tom Potter in recent weeks.
Attentive readers of the local blogosphere back on September 26 might have caught some mention of what at the time was a Portland Tribune poll of 400 likely voters that some of us expected to be released shortly at that time. That particular poll -- much to the surprise of many people who heard about it while the Tribune was sitting on it -- showed Francesconi with 41%, Potter with 38%, and 21% undecided.
While people sat around wondering why the Tribune was not publishing the results of this poll, we noted (in the comments to the above-linked post) that the same firm who conducted this at-that-time-unreleased poll had conducted one towards the end of the primary campaign which showed Francesconi with 27%, Potter with 16%, James Posey with 12%, Phil Busse with 6%, "other" at 5% and 24% undecided. Shortly thereafter, the results of the primary election were 42% for Potter and 34% for Francesconi.
Not a good track record for this particular polling firm (Research 2000) when it comes to this particular race.
So boggling were the new results that the Tribune decided to conduct another poll to see what was going on. In last Friday's edition, they published the results of both polls. The newer poll, of 403 likely voters, showed Potter with 44%, Francesconi with 39% percent, and 17% undecided. This article also refers to some of the other polls going on during this period.
For example, at around the same time as all of this was happening, the Potter campaign itself released the results of a poll it had commissioned of 500 likely voters. This poll showed Potter with 49% and Francesconi with 18%. This poll was also reported in the "City Matters" column in last Friday's Oregonian.
Now comes news from KGW of an Oregonian poll supposedly published in today's edition, although for the life of us we don't remember seeing it in the paper. It's also mentioned on the Potter campaign website today.
In this poll, of 316 likely voters, Potter has 53%, Francesconi has 18%, and 29% are undecided -- much closer to the Potter camp's poll than either of the two Tribune polls which show a far tighter race.
We'll leave it to the our various poll-wonk readers to hash out this particular situation. Obviously, however, the Potter campaign and Oregonian polls cannot be correct at the same time the Tribune polls are correct. So: Any takers on pitching a theory as to what's going on here?
Update
For whatever it's worth, there appears to be a general trend in recent polls, as some others in addition to those mentioned in this item reportedly produced very similar results. In other words, the Tribune polls really are the oddity here. Somewhere in all of this, presumably, is some missing or mystery information which would explain the rather dramatic discrepancy.
So we again ask our poll-wonk readers to weigh in. Is such a stark difference, in which one source comes up with numbers starkly different from something like half a dozen other polls, the result of some sort of normal aberration that crops up in polls, faulty methodology, an agenda, or something else?
Comments (1)
Anne Dufay on 08 Oct 2004
I recommend the following for a scholarly, but accessible, answer to your question: http://www.tomdispatch.com/index.mhtml?pid=1881
Or, Arianna, for her own inimitable take on the question:
http://www.ariannaonline.com/columns/column.php?id=737