Howie laments Bush's numbers in WashPost.com's most recent
Media Notes Extra...
Here we go again: Here we go again:Poll Finds Bush Job Rating at a New Low. Except it's not a new low, it's a new low compared to the last Washington Post survey. Despite two months of "new low" headlines, Bush has actually been stuck in the 36-40 range. An old low, you might say.
Howie's been harping on this topic for months now, but what he's not saying in this particular instance is what this is fueled by...
The first time Howie brought this up (that I'm aware of), he mentioned it in his column titled
"Media Limbo." In it, he said...
Can I just grumble a little about this USA Today /CNN poll?
"President Bush's 'approval rating' has sunk to a new low according to a USA TODAY/CNN/Gallup poll released Monday.
"The latest results show only 36% of those polled saying they 'approve' of the way Bush is handling his job. Bush's previous low was 37%, set last November.
"Sixty percent of those polled said they 'disapprove' of Bush's performance. That matches an all-time worst rating hit last November and again two weeks ago."
Bush is at a new low compared to USA's last poll. CBS has Bush at a new low compared to the last CBS poll. Etc., etc. All true, but they give the collective impression that Bush is sinking week to week. Why do they only compared figures to their own past surveys, when they're fully aware of the others?
At the time, I wrote Kurtz regarding the sheer idiocy of what he's arguing for. What he conveniently overlooks (so that people don't get "the collective impression that Bush is sinking week to week"...even though he is sinking week to week) is that different newspapers (or networks, or what have you) and pollsters use different criteria to arrive at their numbers. Different questions are used, polling pools of different sizes and makeup are used, etc. The only scientifically accurate way of looking at the numbers is to look at a continuum from one's own previous polls to one's most recent poll. Comparing two media outlets' results to each other also raises more questions than it answers, not only requiring that the differences in polling techniques be detailed (to avoid false equivalence), but also requiring that questions of whose polls are examined, whose polls get greater prominence in the piece, etc. be answered. Charges of bias are bound to occur with more frequency in Kurtz's ideal situation, as each media outlet is bound to hype their own numbers over those of other sources. An examination of different outlets' numbers is actually a good idea, but it would require an objective outside source to provide -- not a medium that is doing its own polls.
My letter to Kurtz in which I detailed much of what I wrote above was, of course, unanswered. No surprise, then, that a mere two days later Kurtz was back to complain about this again in his column "The Russ Fuss."
Speaking of that, another "new low" for Bush in the WSJ/NBC poll , even though, as I grumbled the other day, the comparison is not to other recent "new low" surveys but to the last Journal poll.
Again, he ignores the problems in addressing multiple poll results. I responded again, this time in a letter to the editor (which was also cc'ed to Deb Howell -- her response was "I will send this to him. Thanks for your opinion. Deborah." Like him recieving the same points from me directly helped a lot -- thanks, Deb.). Again, no response.
The problem is not that Kurtz wants different polls compared and contrasted. As I mentioned above, that's actually a good idea, but problematic in practice, particularly when the media outlet hosting this comparison is doing its own poll. The problem is why Howie wants this to take place. His grumbling occurs because readers are getting "the collective impression that Bush is sinking week to week," despite this being an accurate impression. He wants to remedy this by comparing the Post's numbers to, say, anther poll in which Bush doesn't appear to be doing quite as badly (even though I'm hard-pressed to think of a poll whose results are showing that Bush's numbers are rising). No, Kurtz wants to give the impression that Bush is holding steady. That even though his poll numbers are collectively sinking, because different polls result in different nubmers and differing rates of decline, we should instead be reporting that Bush's numbers are in a holding pattern of 36-40%. Instead of reporting based on the Post's own evidence, we should instead not be held back by silly things like statistics and quantifiable data. A real reporter doesn't rely on fact, but reports on what's in his or her gut. Because that's where the real truthiness lies.
And this is considered analysis while Froomkin is an opinion-based columnist??? Well, it's the Post. Whadda expect? Reality-based analysis?