An Observation and a Question

October 11, 2012

Comments

A quick thoughts on blogging as well as an observation/comment of some recent commenting trends.

The observation is this: I would note that there have been a lot of positive stories about Romney’s numbers here at OTB over the last several days, just as there were a lot of stories about Obama’s good numbers prior to this week. The reason for this is that we here at OTB (and I think I can speak for James Joyner and Doug Mataconis on this topic) tend to blog in response to the news as it happens. As such, we do not have a specific story-selection strategy (although, clearly, we each have topics that are of special interest to each of us with some overlap).

My is question is: where have all the poll denialists gone? Where has all the talk about improper sampling, skewings, and hackery? I have seen a number of posts about polls this week (as noted above) and yet I do no see all the ranting and raving that had been the norm. It is as if some people react more to these numbers based more on their preferences than they do based on reasons and evidence. Odd, that.

(And yes, that is a snarky and sarcastic paragraph, and it will no doubt annoy some and amuse others. However, my point is sound: we should approach information and address based on reason, not visceral preferences. If we are going to question numbers, have a reason. If we are going to call someone a hack, have evidence. And if we are going to rant and taunt, expect not to be taken seriously).

40 responses to “An Observation and a Question”

Chris Lawrence

October 11, 2012 at 10:20 am

Ah, but you see, all that right-wing pressure about improper sampling and stuff made Gallup et al. stop skewing their polls.

On a (slightly) more serious note, it seems to have moved to the left-wing side of my Twitter feed. Like most contemporary elite politics, rationales seem to temporally follow partisan views, rather than constructing them.
Alex Knapp

October 11, 2012 at 10:21 am

When the facts favor Republicans, it’s the truth we all knew all along.

When the facts don’t favor Republicans, it’s a liberal conspiracy to hide the truth.
Jen

October 11, 2012 at 10:26 am

Well, according to a piece on BuzzFeed, the “unskewed” guy is taking credit for “scaring pollsters straight.” So, according to him, the polls were skewed until he launched his ground-breaking and clearly intimidating site, and now that Romney is tied or leading, the polls are now more accurate.

It’s a really interesting world that these people reside in, isn’t it?
mantis

October 11, 2012 at 10:28 am

My is question is: where have all the poll denialists gone? Where has all the talk about improper sampling, skewings, and hackery?

Not to worry. If Romney’s numbers start tumbling, they will be back in full force.
LaMont

October 11, 2012 at 10:29 am

You don’t say!

I’ll play “Doug’s” advocate – but…both sides do it!
James in LA

October 11, 2012 at 10:33 am

Poll denial is one more regurgitation of the time-honored conservative ploy of discrediting the messenger so we never have to listen to the message. For the last four years, the GOP has spent all its time on this failed strategery [sic], leaving them with no governing achievements 2000 – 2012 on which to run. W, who?

So we get some really crazy crap that is so transparent even Sarah Palin sees through it. The GOP simply telegraphs to the world it is not yet ready — and will not be for a generation — for governing the US in our complex, tightly connected era. To combat this, the GOP nominated a serial liar that is unmasked with ease, and who confirms his disinterest in governing in private.

The author is correct in posing this question, and it really demands an answer. I would prefer billboards across the country, of course…
Gromitt Gunn

October 11, 2012 at 10:37 am

I think that a lot of people are just plain unfamiliar with the psychology of cognitive dissonance and confirmation bias. Even being aware of them, I struggle to maintain objectivity in the face of news stories that contradict how I wish the world to be. And I know that sometimes I fail at maintaining that objectivity or empirically sorting the new facts into my existent world view.

But – at this risk of arrogance, since I clearly know which I think is objectively better – I think that is a primary difference between people educated and accultured into societies that have internalized the Englightenment and the scientific method versus those accultured into societies that have internalized blind faith and “common sense.” We’re in a really weird position right now where there are equally loud contingents that prefer either Society A or Society B. I see Society A as being the way forward and Society B as being a devolution into the Dark Ages (hence, the arrogance).

Those who prefer Society B don’t want to deal with data that undermines faith. They would rather shut the door in the face of uncomfortable facts than revisit assumptions. And, so, yes, this week the polls are unskewed. Who knows if they’ll be skewed or unskewed next week?
Steven L. Taylor

October 11, 2012 at 10:39 am

@Chris Lawrence:

Ah, but you see, all that right-wing pressure about improper sampling and stuff made Gallup et al. stop skewing their polls.

That must be it!

On a (slightly) more serious note, it seems to have moved to the left-wing side of my Twitter feed. Like most contemporary elite politics, rationales seem to temporally follow partisan views, rather than constructing them.

Indeed. I have seen as much denialism, per se, as just utter despair on my Twitter feed (not that such denialism from leftward quarters would surprise me).
Steven L. Taylor

October 11, 2012 at 10:41 am

@Gromitt Gunn:

I think that a lot of people are just plain unfamiliar with the psychology of cognitive dissonance and confirmation bias. Even being aware of them, I struggle to maintain objectivity in the face of news stories that contradict how I wish the world to be.

Indeed. I think it is something that we are struggle with. The difference, I suppose, is often between those who try, at least, to struggle with these issues and those who don’t even think about them.
Alex Knapp

October 11, 2012 at 10:49 am

It hasn’t been absent on the left, but it hasn’t been nearly to the same degree.
Anderson

October 11, 2012 at 10:53 am

There’s been some quibbling, but just as Dems were realistic about how Obama got hammered in the debate, they’ve been realistic about the consequential drop in the polls.

Which seems to be leveling out, btw. Even Rasmussen keeps Obama up in Iowa; Ohio looks steady in the averages; Romney could take CO, VA, and FL and lose, if Obama keeps his firewall states plus OH and Nevada.

I think the giddy “hey, we could have a landslide!” talk is over and done, but assuming Obama gets serious in the next 2 debates and Biden doesn’t masturbate on stage, safer money is on a narrow-to-moderate Electoral College win for Obama, with at least a plurality of the pop vote.
Stonetools

October 11, 2012 at 10:57 am

The people on the left aren’t really disputing the Romney surge. Rather, they are blaming on the surge on the most disastrous debate performance in American history. ( this is not snark- this is how Andrew Sullivan described it). From the left, I’m seeing panic, not denialism.
Gromitt Gunn

October 11, 2012 at 11:19 am

@Steven L. Taylor: That was what I was trying to get at with comments about reason versus faith – it seems that there are large chunks of our population that aren’t even interested in trying. It is part of why I find the assaults on history and science curriculums in public schools – and here in Texas, even on critical thinking skills themselves – so alarming.
Buzz Buzz

October 11, 2012 at 11:45 am

If we are going to question numbers, have a reason.

I believe the most common reason given for people discounting some earlier polls was that the polling sample splits were wildly discordant (e.g. D+13) with the party splits of actual voters in recent elections. I don’t think the more recent polls show the same extreme variance.

My is question is: where have all the poll denialists gone?

I don’t think that “all the poll denialists have gone” anywhere.

If we are going to call someone a hack, have evidence.

This is an excellent idea.

To that end, would you identify some of the specific commenters or comments at OTB that you put in the category of “poll denialists” (i.e. all the ranting and raving that had been the norm) who rejected earlier polls but are embracing recent ones?

Alternately, could you list the partisan sample splits of the polls that OTB commenters “denied” for comparison against the partisan sample splits of the polls that aren’t being “denied”?

(I didn’t see any links whatsoever in your post; I apologize if I overlooked them.)

Because if the party ID splits in the underlying samples haven’t changed in the more recent polls, then you’ve got a solid basis for calling out anyone who criticized the results of the earlier polls on that basis.

But if the underlying samples in the more recent polls now more closely track the recent partisan breakdown of the voters, then the lack of criticism is unremarkable as the complaint has been addressed.

And if we are going to rant and taunt, expect not to be taken seriously

Agreed.
Steven L. Taylor

October 11, 2012 at 12:00 pm

@Buzz Buzz: A quick response: the problem with all the D +x statements is that they assume that partisan ID is a demographic features like age, gender, and race. It rather obviously isn’t. Indeed, if I recall properly, of the major pollsters only Rasmussen weighs the sample specifically for partisan content.

To put it all more simply: one would expect that in a poll that shows Obama ahead that there would be more self-identified Ds, and likewise more Rs in a poll favoring Romney. This is because partisan ID and candidate preference co-varies rather substantially (something that cannot happen with things like race, gender, etc.).

As such: the main thing that shifts is the actual preferences of the persons polled.

Also: all this talk about D+x ignores the fact that more “independents” are actually leaners one way or another to start with.
Steven L. Taylor

October 11, 2012 at 12:04 pm

@Buzz Buzz: See, for example, here: http://www.people-press.org/2012/08/03/party-affiliation-and-election-polls/

Particularly in an election cycle, the balance of party identification in surveys will ebb and flow with candidate fortunes, as it should, since the candidates themselves are the defining figureheads of those partisan labels. Thus there is no timely, independent measure of the partisan balance that polls could use for a baseline adjustment.
Buzz Buzz

October 11, 2012 at 12:15 pm

I already understood that you disagreed with those criticisms. Your crude conflation of statistical criticism with holocaust denial (“poll denialists”) makes your disagreement crystal clear.

But you pointedly avoided the more important portion of my comment, where I asked you to provide some evidence to support your argument:

To that end, would you identify some of the specific commenters or comments at OTB that you put in the category of “poll denialists” (i.e. all the ranting and raving that had been the norm) who rejected earlier polls but are embracing recent ones?

Alternately, could you list the partisan sample splits of the polls that OTB commenters “denied” for comparison against the partisan sample splits of the polls that aren’t being “denied”?
Steven L. Taylor

October 11, 2012 at 12:20 pm

@Buzz Buzz:

Your crude conflation of statistical criticism with holocaust denial (“poll denialists”) makes your disagreement crystal clear.

*sigh*

So much for a attempt at reasonable discourse.
Buzz Buzz

October 11, 2012 at 12:21 pm

So I guess you won’t be providing any evidence to support your argument after all?
Steven L. Taylor

October 11, 2012 at 12:25 pm

@Buzz Buzz:

The problem is, you aren’t arguing in good faith. For example, you ignored what I posted above in response to the sample question.

Feel free to peruse the threads of various posts. The most prominent example was Smooth Jazz. Jan and Florack come to mind as well. I think you qualified. There are numerous others. I am responding to an overall sentiment, which has been pretty clearly observed both in comment sections here and in other parts of the media world.

Alternately, could you list the partisan sample splits of the polls that OTB commenters “denied” for comparison against the partisan sample splits of the polls that aren’t being “denied”?

This strikes me as project that you are free to undertake, if you wish. I would continue to note, however, that you are missing the basic point about what the “partisan ID” variable means and how it relates to candidate preference.
Steven L. Taylor

October 11, 2012 at 12:29 pm

Alternately, could you list the partisan sample splits of the polls that OTB commenters “denied” for comparison against the partisan sample splits of the polls that aren’t being “denied”?

Seriously: by definition polls that favor Romney are going to be more Republican and vice versa. You can’t treat partisan ID the same way we treat race.
Buzz Buzz

October 11, 2012 at 12:31 pm

Feel free to peruse the threads of various posts.

I was asking you to live up to your own standard: If we are going to call someone a hack, have evidence.

And when I ask for such evidence, you accuse me of not arguing in good faith.

It’s sad that you have time to name-call, but not to provide evidence to support your argument.
Buzz Buzz

October 11, 2012 at 12:55 pm

I think you qualified.

Actually, I don’t even remember commenting on the poll methodology or results in the previous poll threads.

But now at least we have a specific accusation from you, one that can be checked against reality.

If we are going to call someone a hack, have evidence.

So the question remains: do you have any evidence?
David M

October 11, 2012 at 12:58 pm

@Buzz Buzz:

Because if the party ID splits in the underlying samples haven’t changed in the more recent polls, then you’ve got a solid basis for calling out anyone who criticized the results of the earlier polls on that basis.

But if the underlying samples in the more recent polls now more closely track the recent partisan breakdown of the voters, then the lack of criticism is unremarkable as the complaint has been addressed.

Partisan ID isn’t necessarily a demographic the pollsters are weighting by like age or gender. It’s more similar to “who are you voting for”, which can change from day to day. Party ID splits will always change and shouldn’t be considered static.
Steven L. Taylor

October 11, 2012 at 1:07 pm

@Buzz Buzz: You are not responding to my response to your questions. This further underscores you are not interested in a conversation.

You have ignored statements about sampling as well as specifc responses to your queries.
Septimius

October 11, 2012 at 1:10 pm

@Steven L. Taylor:

That’s total nonsense. Election polling is designed to obtain an accurate picture of the electorate, not the general population. That’s why election polls that only sample adults are much less accurate than polls that sample registered voters. And, polls that sample registered voters are less accurate than polls that sample likely voters. Voter registration and voter likelihood are fluid factors just like party affiliation. Party affiliation does matter and does make a big difference in the poll results. That’s why some of the polls that had oversampled democrats by double digits were so absurd.

My is question is: where have all the poll denialists gone? Where has all the talk about improper sampling, skewings, and hackery? I have seen a number of posts about polls this week (as noted above) and yet I do no see all the ranting and raving that had been the norm. It is as if some people react more to these numbers based more on their preferences than they do based on reasons and evidence. Odd, that.

I read most of the blog posts on the polls over the past month or so. I didn’t see much ranting and raving. I did see some commenters point out that the partisan breakdown was heavily weighted toward democrats and question the accuracy. That is a totally legitimate point to make. Your “snarky” post criticizing “poll denialists” does nothing to broaden our understanding of the polls or help the dialogue. Honestly, it’s hard to take you seriously.
Anderson

October 11, 2012 at 1:11 pm

Just b/c Buzz Buzz got a “GOP” tattoo on his butt, he thinks everyone else’s party ID is fixed as well.

Obviously, the only sound recourse is voter ID for poll respondents.

Or — stay with me here — we may need to pick a day and actually have everyone show up at various sites around the country and put on the record whom they’re voting for. Then we would have a relatively reliable result. Maybe we can try that next month.
Anderson

October 11, 2012 at 1:14 pm

@Septimius: Since the polls are so inaccurate, they must’ve been wildly off in 2008 and 2004. Right? Oh wait, no they weren’t.

What do you think the pollsters are getting wrong now that they weren’t then?
Steven L. Taylor

October 11, 2012 at 1:16 pm

@Septimius: I didn’t make any arguments about sampling adults v. Likely voters.

Indeed, it is wholly unclear as to what it is that you are calling nonsense.
Buzz Buzz

October 11, 2012 at 1:24 pm

You are not responding to my response to your questions. This further underscores you are not interested in a conversation.

My questions were requests for you to provide evidence to support your argument.

Despite your many comments in this thread, you have still not done so.

You have offered no evidence at all to support the argument you made in your post – that OTB had been overrun with ranting and raving poll denialists who had suddenly stopped commenting because new polls showed Romney with a lead over Obama. ”

You also have offered no evidence to support your new claim that I “qualified” as one of the “poll denialists” you were referring to in your original post. (I don’t think I even commented on the polls at all before this thread.)

Rather than engage in a normal “conversation” where you support your claims by providing evidence, you have instead chosen to retreat into making dishonest ad hominem attacks against me and baseless claims of bad faith.

If your arguments and evidence were strong, I don’t think you’d have to resort to using such tactics.
mantis

October 11, 2012 at 1:29 pm

@Buzz Buzz:

You have offered no evidence at all to support the argument you made in your post – that OTB had been overrun with ranting and raving poll denialists who had suddenly stopped commenting because new polls showed Romney with a lead over Obama. ”

Strawman. He never said anything about the site being “overrun,” and he obviously wasn’t saying they had stopped commenting, but rather stopped having problems with polling methods as soon as their candidates polling improved.

You also have offered no evidence to support your new claim that I “qualified” as one of the “poll denialists” you were referring to in your original post. (I don’t think I even commented on the polls at all before this thread.)

He said he “thought” you qualified. There is uncertainty there.

Rather than engage in a normal “conversation” where you support your claims by providing evidence, you have instead chosen to retreat into making dishonest ad hominem attacks against me and baseless claims of bad faith.

That’s not ad hominem, and the claims aren’t baseless. You are nothing but bad faith, and always have been.

If your arguments and evidence were strong, I don’t think you’d have to resort to using such tactics.

Likewise, if you had strong arguments, you wouldn’t ignore responses to them.
gVOR08

October 11, 2012 at 1:36 pm

Every conservative I ever met considers him/herself a hard headed realist. Self awareness is not a frontline conservative skill.
mantis

October 11, 2012 at 1:38 pm

@Buzz Buzz:

I would also note that it’s difficult to identify if you have been among the poll denialists, as you comment under so many different sock puppet names.
Septimius

October 11, 2012 at 1:38 pm

@Steven L. Taylor:

we should approach information and address based on reason, not visceral preferences. If we are going to question numbers, have a reason.

The commenters around here that questioned the poll numbers had a reason. The polls were heavily weighted toward democrats, sometimes absurdly so. As I wrote previously, that is a totally legitimate reason.

Any campaign would discount a poll that oversampled one party over the other to a degree that would never be reflected in voter turnout. It is an impossibility that 10% more democrats will turn out on election day, yet there were more than a few polls that had 10% more democrats in their sample. Those polls are inaccurate and deserve to be questioned.

Your characterization of ranting and raving and poll denialism among certain commenters is nonsense. I haven’t seen it, and you haven’t provided any evidence of it.
David M

October 11, 2012 at 1:52 pm

@Septimius:

It is an impossibility that 10% more democrats will turn out on election day, yet there were more than a few polls that had 10% more democrats in their sample. Those polls are inaccurate and deserve to be questioned.

The point about the denialists is not they thought those specific polls were wrong, it’s that they used those polls as their evidence that all the polls were wrong. As in, because some poll somewhere may have oversampled Democrats, other unrelated polls were wrong.
Buzz Buzz

October 11, 2012 at 1:53 pm

I would also note that it’s difficult to identify if you have been among the poll denialists, as you comment under so many different sock puppet names.

I don’t comment using other names here.

You may be thinking of jukeboxgrad (see the comment URL under my name, or here for examples).
mantis

October 11, 2012 at 1:59 pm

@Septimius:

The commenters around here that questioned the poll numbers had a reason.

Yeah, they didn’t like the outcomes.
Septimius

October 11, 2012 at 2:13 pm

@David M:

Well, if that was Taylor’s point, he didn’t make it very well. And, I don’t recall the “poll denialists” claiming every poll was wrong because some were skewed. If they were, I’ll stand corrected. But, I don’t think it’s fair to heap scorn or derisively label someone a ‘poll denialist” because they legitimately question a poll sample, especially one with a double digit democrat advantage.
mantis

October 11, 2012 at 2:36 pm

@Septimius:

But, I don’t think it’s fair to heap scorn or derisively label someone a ‘poll denialist” because they legitimately question a poll sample, especially one with a double digit democrat advantage.

Their questions are not legitimate because they are based on a total misunderstanding of polls and various other nonsense.
Barry

October 12, 2012 at 1:31 pm

@Stonetools: “The people on the left aren’t really disputing the Romney surge. Rather, they are blaming on the surge on the most disastrous debate performance in American history. ( this is not snark- this is how Andrew Sullivan described it). From the left, I’m seeing panic, not denialism. ”

I haven’t seen *any* denialism on the left. All I’ve seen is anger at a bad performance, anger at Romney waltzing in an lying like a crack wh*re, taking consolation in the indicators that we’re still leading, and determination to beat them.

I’m sure that ‘some guys’ on Twitter are, well – all a-twitter. However, I haven’t seen anything like the storm of denialism on the right. And that goes 100 times more for publications (as opposed to blogs and Twitter).