Comments on FiveThirtyEight: Politics Done Right: House Effects In Da House

酒店經紀人, 菲梵酒店經紀, 酒店經紀, 禮服酒店上班, 酒店小姐兼職, 便服酒店經紀, ...

2009-10-08T04:58:53.987-04:00

酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店經紀,
酒店打工經紀,
制服酒店工作,
專業酒店經紀,
合法酒店經紀,
酒店暑假打工,
酒店寒假打工,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店工作,
酒店打工經紀,
制服酒店經紀,
專業酒店經紀,
合法酒店經紀,
酒店暑假打工,
酒店寒假打工,
酒店經紀人,
菲梵酒店經紀,
酒店經紀,
禮服酒店上班,
酒店小姐兼職,
便服酒店工作,
酒店打工經紀,
制服酒店經紀,
酒店經紀,
菲
梵,

艾葳酒店經紀公司提供專業的酒店經紀, 酒店上班小姐,八大行業,酒店兼職,傳播妹,或者想要到打工兼差、...

2009-09-22T02:34:39.787-04:00

艾葳酒店經紀公司提供專業的酒店經紀, 酒店上班小姐,八大行業,酒店兼職,傳播妹,或者想要到打工兼差、打工,兼差,或者八大行業,酒店兼職,想去酒店上班, 日式酒店,制服酒店,ktv酒店,禮服店,整天穿得水水漂漂的,還是想去制服店當上班小姐,水水們如果想要擁有打工工作、晚上兼差工作、兼差打工、假日兼職、兼職工作、酒店兼差、兼差、打工兼差、日領工作、晚上兼差工作、酒店工作、酒店上班、酒店打工、兼職、兼差、兼差工作、酒店上班等,想了解酒店相關工作和特種行業內容,想兼職工作日領、假日兼職、兼差打工、或晚班兼職想擁有快速賺錢又有保障的工作嗎???又可以現領請找專業又有保障的艾葳酒店經紀公司!

艾葳酒店經紀是合法的公司工作環境高雅時尚，無業績壓力，無脫秀無喝酒壓力，高層次會員制客源，工作輕鬆，可日領、現領。
一般的酒店經紀只會在水水們第一次上班和領薪水時出現而已，對水水們的上班安全一點保障都沒有！艾葳酒店經紀公司的水水們上班時全程媽咪作陪，不需擔心！只提供最優質的酒店上班,酒店上班,酒店打工環境、上班條件給水水們。心動嗎!? 趕快來填寫你的酒店上班履歷表

水水們妳有缺現領、有兼職、缺錢卡奴的煩腦嗎?想到日本留學缺錢嗎?妳是傳播妹??想要擁有高時薪又輕鬆的夜間兼職工作,打工機會和,假日打工,假日兼職賺錢的機會嗎??想實現夢想卻又缺錢沒錢嗎!??
艾葳酒店台北酒店經紀招兵買馬!!徵專業的酒店打工,想要去酒店的水水,想要短期日領,酒店日領,禮服酒店,制服店,酒店經紀,ktv酒店,便服店,酒店工作,禮服店,酒店小姐,酒店經紀人,
等相關服務幫您快速的實現您的夢想~!!

^^ very nice 徵信, 徵信社, 徵信, 徵信社, 徵信, 徵信社,徵信, 徵信社, 徵...

2009-03-12T01:25:00.001-04:00

^^ nice blog!! ^@^徵信, 徵信網, 徵信社, 徵信社, 徵信社, 徵...

2009-03-12T01:25:00.000-04:00

徵信社, 感情挽回, 挽回感情, 徵信, 徵信社, 徵信, 捉姦, 徵信公司, ...

2009-02-09T01:22:00.000-05:00

To clarify my previous post:My question is, what i...

2008-08-25T20:51:00.000-04:00

To clarify my previous post:
My question is, what if a pollster leans towards a particular candidate in all its polls, and he's right? I mean, the "lean" is only when you compare it with other pollsters, the "trend". It might still be more accurate to account for the "house effect" of all pollsters. But you've already adjusted your polls and their weights according to a large variety of factors - including their accuracy in previous elections and previous primaries. What makes you sure that adjusting for the house effects won't, at times, overlap with the pollster ratings, etc. you've already given them?

Nate,Isn't it pretty difficult to discern a "bad" ...

2008-08-25T20:44:00.000-04:00

Nate,

Isn't it pretty difficult to discern a "bad" house effect? Put it another way, how do you know if the "house effect" of one poll doesn't actually account for its accuracy? It's a very, very complicated job, and ultimately I don't think it'll help your prediction that much, Nate. It might even adversely affect your accuracy.

Nate said:Right now, our four [sic] most prolific ...

2008-08-25T20:21:00.000-04:00

Nate said:

Right now, our four [sic] most prolific pollsters -- Rasmussen, SurveyUSA and Quinnipiac -- collectively account for about 2/3 of all the data that forms our daily averages.

The three listed pollsters, Rasmussen + SUSA + Quinnipiac only add up to roughly 59%.

I think you mean the four pollsters: Rasmussen + SUSA + Quinnipiac + PPP, for about 66%, yes?

The single smartest thing McCain could do is name ...

2008-08-25T19:46:00.000-04:00

The single smartest thing McCain could do is name Sarah Palin as his VP. She is young, charismatic (for a Republican) and would give the PUMA crowd an excuse to vote for the geezer and the girl...I think if he sees the numbers move towards Obama over the next few days he will, though he probably will pick a non=pick like Pawlenty.

Thanks for finally recognizing the house effect, Nate...it is a politically correct way of getting around the consistent pro-GOP leaning of your partner, Rasmussen, since, as you asserted, house effect does not necessarily mean bias. Whether or not Rasmussen is "biased", the house effect leans the most prolific pollster GOP by 2-3 points...

PeteKent:What did you do with yourself? A coheren...

2008-08-25T16:34:00.000-04:00

PeteKent:

What did you do with yourself? A coherent, reasonably well thought out post. Will wonders never cease?

You wrote, "We must wait for the Conventions to be concluded and the dust to settle before any clear picture can emerge. Obama is likely to get a pretty big bump from that will build over the next four days."

Quite right. Except that it contradicts so many of your earlier posts, in which you cherry-picked minor fluctuations in the data to state so firmly what was happening.

Oh well, wisdom acuired late is better than no wisdom. (Asssuming, that is, that this post is not a temporary aberration.)

McCain has a new ad out featuring PUMA Debra Barto...

2008-08-25T15:13:00.000-04:00

McCain has a new ad out featuring PUMA Debra Bartoshevich:

Hahaha! Sounds like she's selling diet products - I would have taken that as a The Daily Show parody of a McCain ad.

That ad wasn't worth the production costs. ;)

DCM, if the crux of yur argument is that Nate's ac...

2008-08-25T15:03:00.000-04:00

DCM, if the crux of yur argument is that Nate's accuracy ratings aren't really right, that's something to take up with him. Those accuracy ratings are implicit for his methodology to produce other than a GIGO result. Yes, past performance is assumed to be indicative of future performance.

The house effect is implicitly assumed in the accuracy rating. Otherwise, the pollster's results would be made inaccurate by the house lean. Presumably, the lean was also in effect in the prior results upon which the accuracy rating is based.

Putting a house lean on top of the accuracy rating is double counting. It's just plain bad methodology. You're right, it would smooth the results, but I don't think that's to the benefit of accuracy.

CHRISwhat I left out of the previous post is that ...

2008-08-25T14:54:00.000-04:00

CHRIS

what I left out of the previous post is that I do not agree on the term 'house effect'.

They [meaning Franklin & Silver] have both recognized inherent built-in biases in most of the pollsters models.

That is still bias, even if not with intent. Call it by it's given name. In statistics it is 'bias'. That is not a partisan shot, but an analytical fact.

Adam, I understand what you're saying, and I disag...

2008-08-25T14:53:00.000-04:00

Adam, I understand what you're saying, and I disagree that one should "norm" Rasmussen toward CW. That implicitly assumes that the crowd is righter than the pollster, and also assumes linear relationships of the numbers, neither of which I think are true.

Again, the accuracy rating is the main adjuster for so-called "bias," so to pull a pollster's numbers away from the actual results can only reduce accuracy, not increase it.

It is mutually exclusive to say that Rasmussen's about the most accurate pollster, but that he'd be more accurate if his results were skewed more to Dems.

If the idea is to make Nate's numbers look like a bland metapoll, this would accomplish that objective. I don't think that's what he wants.

CHRIS,I would tend to agree with your questioning ...

2008-08-25T14:51:00.000-04:00

CHRIS,

I would tend to agree with your questioning of the spreading around/dampening of polling results.

But as I recall, Nate has indicated in his FAQ posts that his modeling does have some closest neighboring involved & other metrics for projection analysis.

All of those need further adjustments & tweaks.

Still I would more quibble with arbitrary weighting such as the subtracting 166% of the standard error from the house effect coefficient. Why ?

But still at least he is trying to make adjustments for a better predictive model as opposed to moving an agenda or push a narrative only.

I am sure this will not be the final tweaking to his model.

RUDYyou are incorrect in stating that the 'house e...

2008-08-25T14:43:00.000-04:00

RUDY

you are incorrect in stating that the 'house effect' is already reflected in the 'accuracy ratings'.

That is just not true.

But anyways, if any analyst recognizing the data has a built-in proveable 'bias' for whatever reason & in whichever direction then the competent person makes the proper adjustments or you have the classic GIGO syndrome.

That is what I learned in my public affairs graduate school analytical training. It would be incompetent to ignore a known variable.

Anyway, Nate has a new post up now on Today's polling results.

He says he is FIRED UP & READY TO GO...

The race seems to be in a form of suspended animat...

2008-08-25T14:40:00.001-04:00

The race seems to be in a form of suspended animation with nothing seeming to move it. For whatever reason the Biden announcement has yet to have a discernable impact. Now there may be cross currents operating – some liking his regular Joe appeal while others resenting that he is not Hillary and the two cancelling each other out.

We must wait for the Conventions to be concluded and the dust to settle before any clear picture can emerge. Obama is likely to get a pretty big bump from that will build over the next four days. That seems to be the pattern when he dominates the national stage and takes over the news cycle. The prototype being the Rainbow Tour of the Mid-East and Europe that inflated his numbers, if only briefly, as the publicity fueled interest and support that waned quickly in the aftermath of Republican criticisms.

The big wild card for Obama seems to be what happens with the Hillary faction at his convention: What message will the Clintons send and how will it be received. The current polls show that Obama has a problem with the Clinton wing of his party. He is losing something like 25% of these voters to McCain. That it is nonsensical that they would choose to support “McBush” over Obama whose ideological DNA is virtually identical to Mrs. Clinton’s is beside the point. The problem is real and must be dealt with.

Everyone expects Obama to give a wonderful speech at his convention. Still, I cannot see how anything he says can be a game changer and Thursday night represents a significant risk now that the venue is Invesco field before 75,000 supporters. It is almost as if Obama is seeking to turn into a caricature of himself and comparisons with the epic Summer Olympics in Beijing will likely come to mind. The risk of course is that this feeds the arrogance meme.

Then again, who would not want all that wonderful exposure and there are probably tens of millions who have never really heard Obama speak, so he has a big chance to reach out to undecided and loosely moored voters and secure their support. It’s hard to say which way the thing will go.

Net, net, unless the story becomes the turmoil of the PUMAs, I think Obama will have his good four days.

That the movement could be arrested by McCain’s expected Veep pick, and the mini-hoopla going into the Republican convention is open to debate.

Again we will need to wait and see. I expect McCain and has convention to exceed expectations and don’t underestimate how well he might speak when he gives his own acceptance speech.

* * *

Matt JH was at the most positive I have ever seen him today. All blue skies and clear sailing for Obama. He has a point. With all of Obama’s natural disadvantages and the power of the Republican attack machine why is he still narrowly ahead and not behind?

Dunno. But it’s what makes a horserace!

From the article:"""So instead of giving Barack Ob...

2008-08-25T14:40:00.000-04:00

From the article:

"""So instead of giving Barack Obama a large bounce in states like Michigan and Iowa, we instead take some of that and give him a much smaller bounce spread out over a lot of states."""

Doesn't this goes against the very real fact that many states seem to have a local culture? Suppose a poll in CA shows a huge bounce for Obama. Absolutely, the WRONG thing to do would be to spread that bounce into Utah.

I'm sure you have your reasons, but this sounds like insanity, and smells like "yet another metric"---quite similar to the "our processor is faster than yours" debates.

I guess I don't quite see how the implemented "center" house effect is anything but arbitrary.

Finally, "house lean" seems like a bad scientific term. You already have "accuracy", so why not add in "precision". In your usage "house lean" == "precision".

Saddened.

DCM, the reason I said I thought Nate was swayed b...

2008-08-25T14:36:00.000-04:00

DCM, the reason I said I thought Nate was swayed by his lib buddies is that his posts over the last several weeks are laden with defense and angst over his numbers showing more McCain move than CW showed. When parsed, Rasmussen was the reason. That reason was correct.

An accuracy rating already takes house lean into account, so ajusting for it beyond the accuracy rating is double counting.

I'd be arguing exactly the same thing if it were a lib-reputation pollster being affected. In fact, I'm arguing exactly that in my immediately preceding post to this one that Selzer should have more weight (less time decay) because she's less proflific.

I'm interested in advancing the objectivity of the science, not trying to sway the numbers in any particular direction. I agree that swinging Nate's number more toward Obama ultimately doesn't mean squat come election day, as polls don't push votes. But Nate has an opportunity here, and I hate to see him succumb to unwarranted pressure to denature Rasmussen and use more cowbell.

RUDYYou continue to argue the relative merits from...

2008-08-25T14:31:00.000-04:00

RUDY

You continue to argue the relative merits from the exact opposite position to the logical rational stance.

'Accuracy' ratings are a derived indicator of perceived past performance and have litle or no bearing on future interim snapshot results which can never be 'proven' as valid or 'accurate'.

Your argument seems to be solely for arguments sake.

Now if Nate were to fully adjust for these statistical anomalies then perhaps your point would be better received.

But at this time you are clearly on the losing side of any debate.

Time will sort it out & smooths the relevant data.

And just for the record, Shenandoah Valley is rock...

2008-08-25T14:28:00.000-04:00

And just for the record, Shenandoah Valley is rock-ribbed, deep red Republican full stop. Including it with Southwest inflates McCain's "western" numbers here.

Rudy,"Even so, his stuff was among the first to ca...

2008-08-25T14:27:00.000-04:00

Rudy,

"Even so, his stuff was among the first to catch the McCain swing in July post-Berlin, in large part due to the timeliness of the Rasmussen polls. If he had this methodology change in place then, he would have been further behind the curve."

I think you're still a bit off on your analysis here of what the adjustment actually means. Take the scenario listed. If Obama was "actually" (if we could somehow measure it) +5 nationally and +8 in Michigan in early July, and +1 nationally and +4 in Michigan at the end of July, then presume Rasmussen always nails the result except it's +1 Republican. So, it would have had +4 nationally/+7 MI before, +0 nationally/+3 MI after.

You see what I'm saying? Rasmussen, being accurate, would have picked up the O -4 movement perfectly. *However*, it's still off by 1 every time, in a predictable direction.

So, which makes more sense? To count the polls as is and leave that one point in all of them, but just don't weight them as much, or to realize they pick up national movement very well and weight them heavily, but to slightly adjust the results to account for that one point? It seems to me the latter approach gives more accurate results.

Virginia Conservative:It's really amusing to watch...

2008-08-25T14:26:00.000-04:00

Virginia Conservative:

It's really amusing to watch republican hopes rest with Hillary and Bill Clinton.

Adam-Thats kind of correct, but Mark Warner wins b...

2008-08-25T14:24:00.000-04:00

Adam-

Thats kind of correct, but Mark Warner wins big by a coalition, believe it or not, of blue collar whites in Southwest, white collar professionals in NoVA, and African Americans. Its the most surefire way for Democrats to win statewide. In SW they continually elect a moderate Dem, Rick Boucher, to Congress again and again (and hes not some Dixiecrat, hes quite mainstream in the Democrat Party).

The other strategy--winning NoVA, the Richmond suburbs, and Hampton Roads plus African Americans--is much much riskier. Thats why I want to see those crosstabs. And I do think Biden may play well there, in addition to the Eastern Shore (which is just like lower Delaware).

You are welcome Lat. I think a lot of things about...

2008-08-25T14:23:00.000-04:00

You are welcome Lat. I think a lot of things about this set of polling are weird.