Closed Thread
 
Thread Tools Display Modes
  #101  
Old 09-06-2019, 10:17 PM
Bone's Avatar
Bone is online now
Extrajudicial
Moderator
 
Join Date: Jul 2003
Posts: 10,889

Moderating


Quote:
Originally Posted by DSeid View Post
Carnal K you are welcome. It really is not worth engaging with the silliness that Lance Turbo is posting on this. We can use our time more productively ... say by engaging with the guy muttering on the street corner.
Knock it off. Clearly people are interested in engaging but if you're not, then rather than snide backhanded remarks, simply don't.

[/moderating]
  #102  
Old 09-07-2019, 04:36 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Quote:
Originally Posted by Bone View Post
Knock it off. Clearly people are interested in engaging but if you're not, then rather than snide backhanded remarks, simply don't.

[/moderating]
I take offense at this moderation. I wasn't backhanded in my snideness at all!

Moderation deserved and noted. I'll contain my snideness to my own thoughts going forward.

Quote:
Originally Posted by Lance Turbo View Post
But that's not how Nate Silver does it.

Here is Nate's Iowa forecast for Nov 2016.

The highest weighted poll is indeed A+ Selzer & Company with a huge 3.84 which was in the field Nov 1-4. However, the second highest weighted poll is C+ RABA Research at 2.43 in the field Nov 1-2 followed closely by C- Survey Monkey at 2.41 in the field Nov 1-7.

They are well ahead of:
A- Quinnipiac University 2.01 Oct 20-26
A- Ipsos 0.68 Oct 17-Nov 6
A+ Selzer & Company 0.57 Oct 3-6
A+ Monmouth University 0.25 Sep 12-14

In particular note the decay of the Iowa gold standard Selzer poll that is a month old. Fresh C+ data has four times the weight of month old A+ data.
I appreciate your bringing in how 538 actually operates in the discussion. Yes, as election eve approaches 538 overweights recency. (Details here.)
Quote:
Recency, which is less important early in the campaign and becomes more important later on.
It certainly makes sense to do that. That last week is where many voters finally decide and information older than that ages in value quickly.

The current approval rating is not the circumstance that overweights recency over house reliability. Quite the opposite circumstance.

So now that you are wanting to engage with how "Silver does it", let's. I've pointed out to you a few posts up that C+ rated Rasmussen is one of five houses that report at least weekly, three that are rated higher than it it is.

And there are two new less frequently reporting highly rated house polls that pop in each week on average.

IF, for simplicity and as the edge case, we assumed that highly rated houses all suddenly aged to zero weight after a week (which of course they do not, we know he has them lose weight slower the better rated they are) and all other polls stayed static but C+ rated Rasmussen, how much do you think would Rasmussen need to move to move the tracker by 2, or even 1, given it being one of just those seven, five of which are more highly rated and therefore significantly more heavily weighted?

The answer of course depends on how much a C+ house's weight is discounted relative to a B or A house. Counting them the absurd all the same weight Rasmussen would need to move 7 to move the needle one and 14 to move it two, and the number goes up dramatically as we apply whatever decreased weight it is actually given. Given that Rasmussen actually moved about 3 over that time period it is certainly the case that it did not meaningfully or significantly impact the tracker's output in this particular case, and given that its weight is not the same, that older quality house results do have input more than a week out, and it is highly unlikely that Rasmussen has moved double digits over time while the quality houses have all stayed flat, it is almost certainly such as a general rule.

A single C+ house moving 5 or so points differently than all the quality house results move will have no perceptible impact on the curve, and aggregated with a multiple other C+ houses would start to look like a mostly flat line, as their individual noise would cancel each other out. (Which is what happens with your Harris overweighted and no house correction aggregated line, even as individual quality house results as a group moved by 4.5. in that time period.)
  #103  
Old 09-07-2019, 05:15 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Cantankerousness aside we can learn a lot about how the 538 weights polls based on pollster rating and freshness by looking at that Iowa 2016 forecast. There's nothing special about Iowa here. That just happens to be the one I had open. Also, there are certainly some differences between this model and the Trump approval model, but many of the underlying principles should hold.

Here are some insights that should clear up some misconceptions held by other posters in this thread.

ē Month old data from the same pollster, even an A+ pollster, is weighted significantly less than fresh data (e.g. Selzer poll from beginning of October is about 15% the weight of Selzer poll from beginning of November although sample size plays a role here as well).

ē A fresh C- poll with a larger sample size can have greater weight than a two week old A- poll (e.g. Survey Monkey versus Quinnipiac).

ē A C+ poll in the field at the same time as a B poll can have a higher weight due to a larger sample size (e.g. RABA (2.43) > Emerson (2.27)).

ē Even a C- pollster can have higher weight than the average weight of other pollsters (e.g. Survey Monkey).

There's probably some other things we can draw from that, but these are the ones that leapt to mind. Many of those bullet points, in some form or other, were called absurd earlier in this thread.

I think it's safe to say that Nate Silver does not consider C- to C+ pollster data to be 'utter shit'. Two of three highest weight polls fall into this category and a rational analyst does not feed 'utter shit' into his model.
  #104  
Old 09-07-2019, 05:50 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by DSeid View Post
The current approval rating is not the circumstance that overweights recency over house reliability. Quite the opposite circumstance.
Unsupported and likely incorrect.
  #105  
Old 09-08-2019, 09:13 AM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Quote:
Originally Posted by Lance Turbo View Post
Unsupported and likely incorrect.
I am curious - why do you think that it is "likely incorrect" that current approval ratings are more like polling results several months out, when recency is relatively less valued, than they are like polling results as Election Day approaches, the case in which Silver overweights recency?

Do you really not understand why election eve is "special"?


And do you have any response to the math demonstrating the limit to how much a single tracker, let alone one of the lowest weighted ones, can move the output given its being one of seven pollsters reporting at least weekly and on average at least two high quality houses dropping new data each week, even ignoring the rapidity that C+ trackers age compared to the data from the higher rated houses (thus, for simplicity sake, ignoring the impact of higher weighted houses results more than a week old)?

I mean other than to just state "likely incorrect" with no supportive argument or rationale.


The math really isn't very hard to follow and it really is head-shaking to have someone respond to "2 + 2 = 4" with "likely incorrect."


Last item is a request - I don't program but can you run your program also showing a line of the aggregate (as you define it) that includes Rasmussen please?

My suspicion is that even within your universe, with your method's lack of weighting by rating, and counting each instance of a daily tracker as much as the results by houses that report out every three weeks or less, you will see that your new aggregate line is barely different than your current one, and will not conform to 538's output any better.

Thank you in advance.
  #106  
Old 09-08-2019, 09:26 AM
CarnalK's Avatar
CarnalK is offline
Guest
 
Join Date: Jul 2000
Posts: 18,455
It's certainly reasonable that 538 might not treat the election as the ultimate end of the approval rating and gauge recency differently than in primary and general election polling.
  #107  
Old 09-08-2019, 10:36 AM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Not sure I understand that statement Carnal K.

But we know that 538 considers recency as less important early on and as gaining in importance as Election Day approaches. They state that. The week before the election is the extreme case of recency weighting.

And we know that counting only recent (one week old or less) polls (an extremely high value on recency over quality) a single C+ tracker moving 3 to 5 points would have an insignificant input on 538's outputs.
  #108  
Old 09-08-2019, 10:40 AM
CarnalK's Avatar
CarnalK is offline
Guest
 
Join Date: Jul 2000
Posts: 18,455
I'm saying that "as election day approaches " might not be considered for approval ratings because it's a constantly running number. Nobody is taking general election polls in December 2020 but they will be doing approval polls.
  #109  
Old 09-08-2019, 10:55 AM
CarnalK's Avatar
CarnalK is offline
Guest
 
Join Date: Jul 2000
Posts: 18,455
Meant to add, that's not unreasonable because job approval isn't hard locked to voting preference. For instance, GOP presidents have often had approval ratings amongst minority groups that they've never come close to getting in vote tallies.
  #110  
Old 09-08-2019, 10:58 AM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by CarnalK View Post
I'm saying that "as election day approaches " might not be considered for approval ratings because it's a constantly running number. Nobody is taking general election polls in December 2020 but they will be doing approval polls.
Pretty much this.

Polls six month out from an election and polls seven months out from an election have pretty much equal predictive value so it doesn't really make a ton of sense to drastically devalue the seven month poll in favor of the six month poll just because it's a month old.

Furthermore, polling far out out from an election is typically sparser than polling with an election imminent. For this reason you may not want to be too aggressive with time decay because then your model would tend to reflect only the most recent poll and be less useful as a forecast for and election six months in the future.

538's Trump approval model is trying to keep it's finger on the pulse of the American electorate. For this reason it probably handles data decay much more like a situation where an election is imminent than a forecast of an election six months from now.

Finally, I didn't respond 'likely incorrect' to the statement '2 + 2 = 4'. I quoted the exact statement that I thought was likely incorrect.
  #111  
Old 09-08-2019, 11:07 AM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Carnal K,

AH.

I'd suspect you are right and that recency is not increased in weight for approval rating even as election day approaches, it would not the same sense to ... but by then no one is really caring about approval rating for its own sake and it is not factored into any of the elections forecasts.

Obviously it mattered that Obama had crawled back into net positive approval territory by election day (at this point he had, briefly, fallen almost as low as Trump consistently sits, underwater by about 8) but it was the which candidate would you be vote for polls that were the actual polling inputs, as should be.
  #112  
Old 09-08-2019, 11:41 AM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Quote:
Originally Posted by Lance Turbo View Post

... 538's Trump approval model is trying to keep it's finger on the pulse of the American electorate. ...
Actually, not quite. Again, we go to the source. Their goal is less a single moment pulse than having predictive power.
Quote:
... In the case of approval ratings, there’s no election to predict — so we instead choose the settings based on how well they would have predicted a president’s future approval ratings. It asks, for instance, what settings would best have predicted Bill Clinton’s approval ratings in March 1998 based on data through February 1998. ...
To be precise that is the actual goal. In service of that goal, again, as clearly stated by 538, highly rated houses are weighted more heavily and lose weight more slowly, and poorly rated houses and weighted lightly and lose weight more quickly. The curve is then calculated using fairly aggressive parameters as that turns out to do a good predictive job:
Quote:
The model uses three second-degree (quadratic) smoothers, with bandwidths of 10, 20 and 30 days, and averages these three estimates together. The bandwidth reflects the number of days used to calculate the polynomial: For instance, a 20-day bandwith applied on Feb. 27 would mean that polls from Feb. 7 to Feb. 27 are used in the calculation. The model also requires that a minimum of five polls be used in the calculation, so it will lengthen the time intervals under consideration in the event of sparse data.
Yeah this is getting into the weeds. But I would still love direct answers to the questions and request I made.

Last edited by DSeid; 09-08-2019 at 11:42 AM.
  #113  
Old 09-08-2019, 12:06 PM
Sunny Daze's Avatar
Sunny Daze is online now
Member
 
Join Date: Feb 2014
Location: Bay Area Urban Sprawl
Posts: 12,866
Quote:
Originally Posted by DSeid View Post
Actually, not quite. Again, we go to the source. Their goal is less a single moment pulse than having predictive power.
I think you are saying the same thing Lance did about this.

Lance, FWIW, I found your models interesting. I can see the value in peeking at Rasmussen because they are so important to the Republican understanding of the world.

I particularly approve of your NL West model.
  #114  
Old 09-08-2019, 12:37 PM
elucidator is offline
Charter Member
 
Join Date: Mar 2000
Location: Further
Posts: 60,072
Ravens vs Vikings, Ravens just tried a quantum froth stochastic variation. Didn't work.
__________________
Law above fear, justice above law, mercy above justice, love above all.
  #115  
Old 09-08-2019, 02:01 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
I'm not sure how you are reading this and still coming to the conclusion that month old data could still be weighted as heavily as fresh data.

Quote:
The model uses three second-degree (quadratic) smoothers, with bandwidths of 10, 20 and 30 days, and averages these three estimates together. The bandwidth reflects the number of days used to calculate the polynomial: For instance, a 20-day bandwith applied on Feb. 27 would mean that polls from Feb. 7 to Feb. 27 are used in the calculation. The model also requires that a minimum of five polls be used in the calculation, so it will lengthen the time intervals under consideration in the event of sparse data.
The month old data only falls within the 30 day bandwidth while the fresh data falls into each of the 10, 20, and 30 day bandwidths.
  #116  
Old 09-08-2019, 03:51 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by Sunny Daze View Post
I particularly approve of your NL West model.
Thanks.

I think I have figured out your Twitter handle. Thanks for following.
  #117  
Old 09-08-2019, 05:56 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Lance Turbot, you remain consistent in ignoring and failing to respond to the basic math that illustrated the limits that single poll of 7 published weekly or more frequently can have, especially when it is a low rated one and five others are highly rated, and most fixated on my answer to a hypothetical universe in which there were only two pollsters reporting, one A+ reporting monthly, and and one C+.

(And I remain curious as to whether or not you'll run your script adding a line for your "all else and Rasmussen" to show how much in your model of "all else" adding Rasmussen weighted as strongly as all others and counting every day's report as a full weight moves that curve's needle.)

But yes in the case of only two pollsters reporting, one high quality and one poor, I would suspect that 538 would consider the daily tracker as the equal of one report per period of time such as 3 or 4 weeks, aging its weight faster than it ages the highly rated tracker, and then applying the smoothing curve as above for the sparse data circumstance - lengthening the time intervals under consideration in the event of sparse data such that five fully counting polls are included. IOW barring the circumstance being election eve the older highly rated house's result would still dominate the result, and should.


Meanwhile I had used Iowa as an illustration of my thought process. At the time the most recent poll available was Change Research, a C+ rated pollster, and had Warren way on top above both Biden and Sanders by 11. I stated clearly that its recency did not offset it being only C+ and that I would not believe the race had significantly shifted from the previous A+ rated result showing Biden up 9 over Warren and the even older B YouGov Biden +5 (with Sanders in second and warren in third, down by 12) or Selzer older yet showing Biden +8 until a highly rated firm said the same thing.

Would you have assessed differently? Would you have concluded that the C+ rated pollster was proof of a significant change in Iowa with Warren now on top, worth more because it was more recent?


Of course today a new YouGov is out - Biden +3 over Sanders in second and +12 over Warren, farther below Biden and Sanders than she was in the last YouGov poll.



Looking back it seems that placing the meaning of a C+ firm as of low value by itself, mostly but not completely ignoring it, was more predictive of future polling, was a more rational course, than taking it very seriously would have been.


(Note for Warren fans - she leads as second choice and as one of the ones being considered. She is still positioned well.)
  #118  
Old 09-08-2019, 07:18 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by DSeid View Post
Lance Turbot, you remain consistent in ignoring and failing to respond to the basic math that illustrated the limits that single poll of 7 published weekly or more frequently can have, especially when it is a low rated one and five others are highly rated, and most fixated on my answer to a hypothetical universe in which there were only two pollsters reporting, one A+ reporting monthly, and and one C+.
All evidence indicates that one poll could have a noticeable effect with a large swing. Even a C+ polster.

Quote:
Originally Posted by DSeid View Post
(And I remain curious as to whether or not you'll run your script adding a line for your "all else and Rasmussen" to show how much in your model of "all else" adding Rasmussen weighted as strongly as all others and counting every day's report as a full weight moves that curve's needle.)
You have a lot nerve assigning me tasks after your (admitted) snideness toward me throughout this thread. Do it yourself. The 538 database can be downloaded as a csv. Fire up Excel and get it done.

Alternatively, you could try asking nicely.

Quote:
Originally Posted by DSeid View Post
But yes in the case of only two pollsters reporting, one high quality and one poor, I would suspect that 538 would consider the daily tracker as the equal of one report per period of time such as 3 or 4 weeks, aging its weight faster than it ages the highly rated tracker, and then applying the smoothing curve as above for the sparse data circumstance - lengthening the time intervals under consideration in the event of sparse data such that five fully counting polls are included. IOW barring the circumstance being election eve the older highly rated house's result would still dominate the result, and should.
This is wrong in few ways. Even on day 1 the A+ would only have somewhere in the neighborhood of twice the weight of the C+ poll, and you can read right off the page that they count Rasmussen as a new data point every three days. They do this because Rasmussen posts a three day rolling average. Counting it as a new poll every three days doesn't count any interviews twice, but also doesn't throw any out. The idea that it is treated as one report every 3 or 4 weeks is completely unsupported.

You really can't seem to get over the idea that a fresh C+ poll could be weighted more highly than a stale A+. I don't know what to tell you at this point. It is obvious that this can happen and it is obvious that this should happen. You yourself have posted a couple different quotes from 538 that support the idea that this does indeed happen.

Quote:
Originally Posted by DSeid View Post
Meanwhile I had used Iowa as an illustration of my thought process. At the time the most recent poll available was Change Research, a C+ rated pollster, and had Warren way on top above both Biden and Sanders by 11. I stated clearly that its recency did not offset it being only C+ and that I would not believe the race had significantly shifted from the previous A+ rated result showing Biden up 9 over Warren and the even older B YouGov Biden +5 (with Sanders in second and warren in third, down by 12) or Selzer older yet showing Biden +8 until a highly rated firm said the same thing.

Would you have assessed differently? Would you have concluded that the C+ rated pollster was proof of a significant change in Iowa with Warren now on top, worth more because it was more recent?
Your thought process is consistent with 538's and mine here. 150 days from an election or 200 days from an election polls have roughly equal predictive power so it doesn't improve a forecast to significantly decay polling this far out, particularly when polling is sparse. This was explained in a previous post.

Furthermore, a poll that is a large outlier from all other polling should be regarded with suspicion. 538 handles this by adjusting the weight of such a poll downward algorithmically. I agree with the way they handle this.

Finally, there's a reason 538 doesn't actually do a forecast this far out for something like a the Iowa Caucuses. The uncertainty band would be huge were they to try to put something out there now. Maybe even big enough to include that Change Research poll.

Last edited by Lance Turbo; 09-08-2019 at 07:20 PM.
  #119  
Old 09-08-2019, 07:54 PM
Bone's Avatar
Bone is online now
Extrajudicial
Moderator
 
Join Date: Jul 2003
Posts: 10,889

Moderating


Quote:
Originally Posted by DSeid View Post
Lance Turbot, you ...
Please be circumspect in typing other's usernames. This seems like a typo to me, but it's always a good idea to use the appropriate handles when responding.

[/moderating]
  #120  
Old 09-08-2019, 08:11 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Lance Turbo, I did request very politely in post 105 with a thanks in advance explaining that I donít program (or do Excel). Again please and thank you. Or someone else who knows how.


Bone. An honest typo (and I suspect Freudian typo not autocorrect) but seeing it pointed out did make me laugh. I had caught myself and corrected that typo a few times before as I just have that image, like a fish slapping scene from Monty Python, in my head. Missed it this time. Sorry. No ill intent I promise.
  #121  
Old 09-08-2019, 08:52 PM
CarnalK's Avatar
CarnalK is offline
Guest
 
Join Date: Jul 2000
Posts: 18,455
Quote:
Originally Posted by Lance Turbo View Post
They do this because Rasmussen posts a three day rolling average. Counting it as a new poll every three days doesn't count any interviews twice, but also doesn't throw any out. The idea that it is treated as one report every 3 or 4 weeks is completely unsupported.
538 quite explicitly lowers weight of frequent polls, so those three day running polls most certainly have less impact than another poll. DSeid quoted this already:
Quote:
The weights also account for how often a pollster measures Trump’s approval ratings. If it does so more often than about once per 20 days, each instance of the poll is discounted so that the pollster doesn’t dominate the average just because it’s so prolific. Daily tracking polls also receive special handling from the formula so that interviews are not double-counted.
https://fivethirtyeight.com/features...roval-ratings/

Last edited by CarnalK; 09-08-2019 at 08:53 PM.
  #122  
Old 09-08-2019, 08:56 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Post 105, as a whole, was not polite.

Whatever. Here's the 3 week period at the start of the dip.

Code:
	Rasmussen	non-Rasmussen	Combined
7/22/19	-1.636364	-10.086567	-8.894872
8/11/19	-6.4		-10.168142	-9.601504
Net net	-4.763636	-0.081575	-0.706632
  #123  
Old 09-08-2019, 08:59 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by CarnalK View Post
538 quite explicitly lowers weight of frequent polls, so those three day running polls most certainly have less impact than another poll. DSeid quoted this already:

https://fivethirtyeight.com/features...roval-ratings/
Correct. We discussed it earlier.

That doesn't really change anything I posted.
  #124  
Old 09-08-2019, 09:03 PM
CarnalK's Avatar
CarnalK is offline
Guest
 
Join Date: Jul 2000
Posts: 18,455
You didn't discuss it. You pretended it didn't matter. You said basically "sure, he discounts them but not completely" and then continued on.
  #125  
Old 09-08-2019, 09:07 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
BTW Lance Turbo,

Do you believe that one poll (of seven that report each week, and others also impacting the output) "could have a noticeable effect with a large swing. Even a C+ polster."

Or do you believe that 538 would deal with such "a poll that is a large outlier from all other polling ... by adjusting the weight of such a poll downward algorithmically"?

Or are we back at quantum states in which you hold both those opposing belief states at the same time until you see what the measurement is?

Me?

I believe that 538 has explained how they deal with outliers well in a recent article. It goes into the pot, weighted by the houses rating and size factors. But there's more ...
Quote:
... the Polling 101 answer. When you see a poll that looks like an outlier, just throw it into the average. If you want, you can give some consideration to the sample size and the quality of the pollster.

But if youíve read FiveThirtyEight for a while, youíve probably heard that Polling 101 answer before. So Iím also going to give you the Polling 201 answer. But I want you to promise that youíll abide by it fairly strictly, rather than interpret it too liberally. Pinky swear? OK, great. Then here goes:

If a poll shows a significant change in the race, you should tend to presume itís an outlier unless itís precipitated by a major news or campaign event.

Corollary: You should be much more open to the possibility that a poll reflects a real change if itís among the first polls following a major news or campaign event.

What do I mean by a ďmajorĒ news or campaign event? Some fairly specific types of things. When I made you pinky swear earlier, I was asking you to stick precisely to this list:

1. Debates.

2. Candidates entering or exiting the race, or clinching their nominations.

3. Primary and caucus results (e.g., the Iowa caucuses occur and that has knockoff effects on the next set of states).

4. The conventions.

5. The announcement of vice presidential candidates.

6. The final week of the campaign.

7. Spectacular, blockbuster news events that dominate the news cycle for a week or more. (There generally are only one or two of these per campaign cycle, if that many.) ...

... The sixth category, the end of the campaign, is less well-known as a source of polling movement, but the final days of the campaign can produce sharp shifts in the polls as undecided voters finally settle upon a candidate and as supporters of candidates who look like they canít win (say, a Libertarian who is polling at 4 percent) hold their noses and pick one of the major contenders. Often, especially in primaries, this movement occurs fairly late ó within the final week of the campaign or even the final 24 to 48 hours ...
Yeah again, the final week of the campaign is very special, recency matters more, and outliers are more to be taken seriously.

Back to how to handle sparse polling out of Iowa. If that Warren up by 11 had been a Selzer report, I'd have believed it was at least in the ballpark, that there had been a major shift in the Iowa race, and that future polling would be not horribly inconsistent with it.
  #126  
Old 09-08-2019, 09:10 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Quote:
Originally Posted by Lance Turbo View Post
Post 105, as a whole, was not polite.

Whatever. Here's the 3 week period at the start of the dip.

Code:
	Rasmussen	non-Rasmussen	Combined
7/22/19	-1.636364	-10.086567	-8.894872
8/11/19	-6.4		-10.168142	-9.601504
Net net	-4.763636	-0.081575	-0.706632
Is there please anyway to have this line placed on the same graph for the whole 40 day cycle? If it is not too much work for someone who understands how.

Thank you very much in advance!
  #127  
Old 09-08-2019, 09:12 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by CarnalK View Post
You didn't discuss it. You pretended it didn't matter. You said basically "sure, he discounts them but not completely" and then continued on.
Correct.

I can see how much they are discounted, and also artifacts of the special handling that daily trackers receive in the 'weight' column on the page. This information is already baked in to my position. If you remind me of it a third time it still won't change anything.

Last edited by Lance Turbo; 09-08-2019 at 09:16 PM.
  #128  
Old 09-08-2019, 09:15 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Quote:
Originally Posted by DSeid View Post
BTW Lance Turbo,

Do you believe that one poll (of seven that report each week, and others also impacting the output) "could have a noticeable effect with a large swing. Even a C+ polster."

Or do you believe that 538 would deal with such "a poll that is a large outlier from all other polling ... by adjusting the weight of such a poll downward algorithmically"?

Or are we back at quantum states in which you hold both those opposing belief states at the same time until you see what the measurement is?
I do believe both these things, but these things are not inherently contradictory. These things are both obviously true.
  #129  
Old 09-09-2019, 09:37 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Quote:
Originally Posted by Lance Turbo View Post

Code:
	Rasmussen	non-Rasmussen	Combined
7/22/19	-1.636364	-10.086567	-8.894872
8/11/19	-6.4		-10.168142	-9.601504
Net net	-4.763636	-0.081575	-0.706632
While I would still love to see the whole 40 day cycle with the Rasmussen inclusive aggregate added, I recognize that running the script may be too big of an ask.

That's okay. We can illustrate with this 3 week selected steepest part of Rasmussen's curve, without seeing it graphically.

Your supplied numbers:

Adding Rasmussen (and keeping it weighted as high as any contemporaneously reporting A+ house, the absurd case), which had dropped 4.76 points over that time, altered this version of an aggregate's results by 0.62 points from its non-Rasmussen baseline. Using this method it would take a move of Rasmussen by 10 to move the aggregate by 1.3 points. 538 moved 1.6 during that time.

Over the 40 day period Rasmussen appeared to move less than 4 points. Again, not discounting their weight and length of impact to any less than the highest rated houses, to the fractional weight it actually is, if all else was static, Rasmussen would have moved your version of an aggregate by about half a point.

So the result of accepting your version of an aggregate as "everything else" is that Rasmussen's impact on the total is some significant amount less than half a point (how much less depending on how much less of a weight a C+ house gets relative to A+ to B+ ones).

Per your numbers.

I'd call somewhere significantly less than half a point difference in a curve over 40 days "virtually (but not exactly) the same" curves.

That sub half fraction of a point did, in that specific time period, play a part in 538's output, but the 538 tracker would have dropped its not very significant about 3 points with or without it included, and those who felt that about 3 points of a change was maybe significant would have still been asking if it was and if so what caused it.
  #130  
Old 09-09-2019, 10:29 PM
Lance Turbo is offline
Guest
 
Join Date: Aug 1999
Location: Asheville, NC
Posts: 4,262
Jesus fuck.

I was going tp let this thread die, but now I'm done being nice.

A while back you posted this from 538...

Quote:
The model uses three second-degree (quadratic) smoothers, with bandwidths of 10, 20 and 30 days, and averages these three estimates together. The bandwidth reflects the number of days used to calculate the polynomial: For instance, a 20-day bandwith applied on Feb. 27 would mean that polls from Feb. 7 to Feb. 27 are used in the calculation. The model also requires that a minimum of five polls be used in the calculation, so it will lengthen the time intervals under consideration in the event of sparse data.
And I mused, "I'm not sure how you are reading this and still coming to the conclusion that month old data could still be weighted as heavily as fresh data." I was giving you the benefit of the doubt that you understood what you posted.

Then you posted...

Quote:
Originally Posted by DSeid View Post
Lance Turbo, I did request very politely in post 105 with a thanks in advance explaining that I donít program (or do Excel). Again please and thank you.
And I realized that I should not have given you the benefit of the doubt.

You didn't see that excerpt from 538 that you posted contradicted your position because you didn't understand it. You didn't understand it because you haven't done the work. And you didn't do the work because you aren't capable.

On the 538 page where they go into the approval model details they link to the Wikipedia page for local regression. Do you know how local regression works? Of course you don't. Have you read Cleveland's paper describing the method? No fucking way.

Here's an important quote from the Wikipedia page...

Quote:
At each point in the range of the data set a low-degree polynomial is fitted to a subset of the data, with explanatory variable values near the point whose response is being estimated. The polynomial is fitted using weighted least squares, giving more weight to points near the point whose response is being estimated and less weight to points further away.
Or one from Cleveland...

Quote:
Robust locally weighted regression is a method for smoothing a scatterplot, (x_i, y_i), i = 1, .. ., n, in which the fitted value at x_k is the value of a polynomial fit to the data using weighted least squares, where the weight for (x_i, y_i) is large if x_i is close to x_k and small if it is not.
This means that when we are trying to estimate the response near d over the interval with bandwidth b from start date d - b to end date d points near d - b have almost no weight. The weighting of the polls by quality ameliorates this to a tiny degree, but nowhere near enough to make a d - b have more weight than a d poll. At best this might make two points near each other swap places when ranked by the weight function, but there is no way even a poll with ten times the weight of another could overcome the effect of distance imposed by the tri-cube (or sufficiently similar) weight function they are using.

Even if that were not the case, and it definitely is, a poll nearly a month old is only in one of the three local regressions that they combine to determine the response at d, and the local regression that the month old poll is part of contains the most polls so dilutes each poll in it the most.

For these two reasons, a month old poll has less (probably much less) than 1% of the weight of even the freshest poll according to the cite you provided but didn't understand. Even in the 10 day bandwidth a 10 day old poll probably can't be higher than 3% of a fresh poll.

You have posted numerous things like:

"The current approval rating is not the circumstance that overweights recency over house reliability. Quite the opposite circumstance."
"Yeah again, the final week of the campaign is very special, recency matters more, and outliers are more to be taken seriously."
"Do you really not understand why election eve is "special"?"
"The limit would be that 538 actually weights the C+ one exactly as much as the B ones, right?"

That is equivalent to you waving a giant ignorance flag declaring to the world that you don't understand your own cites.

Please stop posting as if you understand the 538 model better than everyone else. You do not.

Please stop condescending to me and then asking me to do your work. Do your own work and maybe shit like this won't happen again.
  #131  
Old 09-09-2019, 10:51 PM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
Does any of that rambling incoherent wall of text have anything to do with the post above it or the documentation that even large swings of a Rasmussen weighted as high as anything else would, if everything else is static, have relatively little impact on your own Rasmussen inclusive aggregation?

No.

And given that you won’t engage in good faith after multiple attempts ... have a nice day.
  #132  
Old 09-09-2019, 11:04 PM
digs's Avatar
digs is offline
Guest
 
Join Date: Mar 2002
Location: West of Wauwatosa
Posts: 10,007
Quote:
Originally Posted by Lance Turbo View Post
Jesus fuck.

I was going tp let this thread die, but now I'm done being nice.
And apparently, your definition of "being nice" is shutting up and going away (so to punish people you... talk to them? Does that work in real life, too?).


I for one am interested in Trump's approval. And where it's going. So I keep checking this thread in hopes that the arguing over one poll is over. Anyone want to get back to the "Reason why Trump's approval ratings are falling?"

Last edited by digs; 09-09-2019 at 11:09 PM.
  #133  
Old 09-09-2019, 11:07 PM
kirkrapine is offline
BANNED
 
Join Date: Sep 2019
Posts: 381
Quote:
Originally Posted by elucidator View Post
A more interesting question would be about those who disapprove: how many disapprove with a shrug, and how many totally hate his guts. Trouble being, how you could ask the question with a reasonable expectation of accurate results?
Easy. Ask for the strength of the respondent's feelings on a 1 to 5 scale.
  #134  
Old 09-09-2019, 11:22 PM
kirkrapine is offline
BANNED
 
Join Date: Sep 2019
Posts: 381
Why has there been so much discussion in this thread about the accuracy and significance of various polls and analyses of their applications? Let us simply assume arguendo that Trump's approval rates are falling. The important question, raised by the OP, and what we ought to be debating here, is why? What has he done recently that would make inroads in his hitherto-reliable base?
  #135  
Old 09-10-2019, 06:50 AM
DSeid's Avatar
DSeid is offline
Guest
 
Join Date: Sep 2001
Location: Chicago, IL
Posts: 22,661
If we accept that the 3 point drop means anything then we must note that it was drop from the top of his long term narrow range back to its middle section and really the question is why was he near the top of that narrow range at the start, not why he went back to baseline.

So far nothing has made inroads with that reliable base.
  #136  
Old 09-10-2019, 09:15 AM
Hermitian's Avatar
Hermitian is online now
Guest
 
Join Date: Jan 2004
Posts: 2,586
Quote:
Originally Posted by kirkrapine View Post
Let us simply assume arguendo that Trump's approval rates are falling. The important question, raised by the OP, and what we ought to be debating here, is why? What has he done recently that would make inroads in his hitherto-reliable base?
As the OP, I was a little surprised that this thread turned into Statistics 203: The Intricacies and Controversies of Political Polling and You. I didn't really read through all that because it seemed to be getting nowhere fast.

It still could just be noise, but according to 538 he is down a half of percent since the thread start date.
  #137  
Old 09-10-2019, 09:33 AM
Jonathan Chance is offline
Domo Arigato Mister Moderato
Moderator
 
Join Date: Apr 2000
Location: On the run with Kilroy
Posts: 22,927

The Moderator Speaks


Quote:
Originally Posted by Hermitian View Post
As the OP, I was a little surprised that this thread turned into Statistics 203: The Intricacies and Controversies of Political Polling and You. I didn't really read through all that because it seemed to be getting nowhere fast.

It still could just be noise, but according to 538 he is down a half of percent since the thread start date.
Hermitian, I appreciate you starting the thread. Such threads are worthy and capable of bringing light into the darkness.

But this one seems poisoned. Feel free to start another one. Lance? DSeid? Take your feud elsewhere.
Closed Thread

Bookmarks

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off

Forum Jump


All times are GMT -5. The time now is 02:18 PM.

Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2019, vBulletin Solutions, Inc.

Send questions for Cecil Adams to: cecil@straightdope.com

Send comments about this website to: webmaster@straightdope.com

Terms of Use / Privacy Policy

Advertise on the Straight Dope!
(Your direct line to thousands of the smartest, hippest people on the planet, plus a few total dipsticks.)

Copyright © 2019 STM Reader, LLC.

 
Copyright © 2017