Frequently Asked Questions:

[The Vorlon]

I am setting up my own website.

I have coded all the nasty bits (Data base for the polls, etc) and am starting to put some actual content on it.

One section I want to get populated soon is a "frequently asked questions" section.

If anybody has any suggestions as to what would be good topics to cover, I would really appreciate your suggestions!

Thank Crew!

[ATFFL]

General definistions, what is push polling, a tracking poll etc.

A precise definition of margin of error, confidence and ho they change based on sample size.

The ways you can determine who is a likely voter would be good.

Zogby deserves a section on his declien into madness.

[Monty]

Have you been saving this for your 1000th post?

[??????????]

Have you been saving this for your 1000th post?

Vorlon I have a question. Why in the WORLD do you have Utah a lighter shade of blue? It can't be THAT bad!

[The Vorlon]

Have you been saving this for your 1000th post?

Vorlon I have a question. Why in the WORLD do you have Utah a lighter shade of blue? It can't be THAT bad!

that would be... an error... ! LOL !

[??????????]

Have you been saving this for your 1000th post?

Vorlon I have a question. Why in the WORLD do you have Utah a lighter shade of blue? It can't be THAT bad!

that would be... an error... ! LOL !

I was about to go in the corner and cry if that was true! lol

[??????????]

Also I'd like to mention that DCPoliticalReport.com has Florida Bush solid and the poll that they use says.

Bush - 50%
Kerry - 41%

I love being right.

[The Vorlon]

Also I'd like to mention that DCPoliticalReport.com has Florida Bush solid and the poll that they use says.

Bush - 50%
Kerry - 41%

I love being right.

I have always had Florida a shade of blue deeper than most as well.

Bush has dropped back a few % in Ohio relative to 2000 (In 2000 Ohio was 4% better for Bush than his National average) But he has gained, relative to his national average about 3-4% in Florida.

The two states have almost traded places as it were.

[??????????]

Also I'd like to mention that DCPoliticalReport.com has Florida Bush solid and the poll that they use says.

Bush - 50%
Kerry - 41%

I love being right.

I have always had Florida a shade of blue deeper than most as well.

Bush has dropped back a few % in Ohio relative to 2000 (In 2000 Ohio was 4% better for Bush than his National average) But he has gained, relative to his national average about 3-4% in Florida.

The two states have almost traded places as it were.

Any recent polls in VA or NC?

[© tweed]

Why do you have PA as light blue?  I have it as light red.  The last polls (Other than FOX, which has bush up everywhere) has Kerry +1.

[CARLHAYDEN]

I am setting up my own website.

I have coded all the nasty bits (Data base for the polls, etc) and am starting to put some actual content on it.

One section I want to get populated soon is a "frequently asked questions" section.

If anybody has any suggestions as to what would be good topics to cover, I would really appreciate your suggestions!

Thank Crew!

Vorlon,

The can you please either confirm or deny that the weighting of many public polls is higher Democrat in 2004 than it was in 2000, even thought extensive analysis (PEW for one) indicates that the party identification is more favorable to the Republicans than in 2000.

If you find this to be true, what adjustment would you recommend to get a more realistic take on certain polls that have increase the Democrat componenent?

[The Vorlon]

Why do you have PA as light blue?  I have it as light red.  The last polls (Other than FOX, which has bush up everywhere) has Kerry +1.

My map is driven by my national model, not entirely state polls.  

I have had Pennsylvania very very light blue (current model says Bush up by 0.2% - so I actually have it essentially tied)

[© tweed]

So I'm not insane if I think Kerry is up by about 1% there right now?

[??????????]

So I'm not insane if I think Kerry is up by about 1% there right now?

Well.... you are insane but that prediction isn't.

[The Vorlon]

So I'm not insane if I think Kerry is up by about 1% there right now?

Hey, my model says Bush + 0.2%, you say Kerry + 1.0%

The state is a tossup, 4 months out I am not calling anaybody crazy over 1.2%

You are standing on the 49 yard line of the football field, I have crossed the midfield line and am standing on the 49.8 yard line.

In reality, we are in damn near the same place.

[??????????]

With the way polls are trending could this be our actual result? How weird would this be?

R - 312
D - 226

[WalterMitty]

state's rights, that is just delusional.  the only people who think virginia and nc are in play are either 1. card carrying members of the mark warner/jihn edwards fan club, or 2. silly protectionists.

[??????????]

state's rights, that is just delusional.  the only people who think virginia and nc are in play are either 1. card carrying members of the mark warner/jihn edwards fan club, or 2. silly protectionists.

I know, but it's merely speculation.

[Reds4]

How bout a section about several of the top polling organizations and their history along with a bit of commentary on how they conduct their polling and how good they are at it?

[The Vorlon]

I am setting up my own website.

I have coded all the nasty bits (Data base for the polls, etc) and am starting to put some actual content on it.

One section I want to get populated soon is a "frequently asked questions" section.

If anybody has any suggestions as to what would be good topics to cover, I would really appreciate your suggestions!

Thank Crew!

Vorlon,

The can you please either confirm or deny that the weighting of many public polls is higher Democrat in 2004 than it was in 2000, even thought extensive analysis (PEW for one) indicates that the party identification is more favorable to the Republicans than in 2000.

If you find this to be true, what adjustment would you recommend to get a more realistic take on certain polls that have increase the Democrat componenent?

Firstly, not all polls do a hard weight by party ID.

Zogby does, Rasmussen does, Battleground does, CBS News does (it's not a "fixed" number but they adjust to it)  Washington Post/ABC news does not "officially" adjust by party ID, but the way they impose stratification on their samples more or less does the same thing.

The other polls do NOT weight to a fixed party ID.

Zogby has so many weights, adjustments, guesses, fudge factors, tarrot cards, oiuji boards, dice, random number generators, circus clowns, faith healers and exiles from the psychic hotline built into his polls that being out a point or two on party ID is the least of his problems

The NYTimes/CBSNews poll is so structurally ed that their weight to +6 or 7 or so on the Dem side, while clearly crazy, is also a fairly mild in nature compared to the rest of the show.

Rasmussen is at +4 to the Dem side (39/35/26) So yes, If I was running his poll, and using a hard weight (which I disagree with) I would probably go +2 or so.

ABCNews/Washngton Post is an interesting issue technically.

Response rates are indeed dropping - you have to do a lot more calls in total to get your sample now than you did a few years ago.  The wide spread availability of caller ID, call screening, etc have dramatically dropped reponse/completion rates when you do telephone surveys.

Survey Sample International is the firm that supplies the actual calling lists to a lot of polling firms (Gallup, Rasmussen, Survey USA, Mason-Dixon(sometimes anyway), TIPP, Fox, Harris) has compensated for this by enriching the frequency of numbers in various phone exchanges to compensate for the difference in response rates within different exchanges.

Other firms, such as LA Times, CBS, ABC, GreenBurg-Quinlan-Rosner, Zogby generate their own sample lists.

GQR I know does compensate for response rates and I know Zogby also has a weighting factor as well.  

ABC News stratifies their sample and essentially has a quota system by age, race, income, gender, geography, etc - ie they keep calling till they have their "quota" in the various groups they break out (They use the standard 48 "cells" that the US Census uses) - This brings us to what is really the unknown and unknowable question.  Because ABC stratifies so hard, they always have what "looks" like a random sample, but if that means it actually is a random sample is less than clear statistically.

The last two WashingtonPost/ABC news samples have both been +8 to the Dem side on party ID.  This is clearly high, but not "insane" - and it is summer, you typically get a minor fade to the Dem side in party ID.  If I was at ABC news I don't think I would be pushing the panic button yet, but I would be taking a very hard look at my response rates just to make sure something structural had not slipped into their modeling.

The BIG question on this years polling is indirectly related to party ID, which I "think" reading between the lines is what you are really asking (correct me if I am wrong please) and that is the response rate issue.

Response rates are dropping, and the "tools" that let folks not pick up the phone (Caller ID, call screen etc) are substantially more common in higher income => more GOP inclined homes.

Every reputable pollster knows this and is scrambling the best they can to compensate.

Some firms are adjusting their actual calling patterns (Fox, TIPP, Gallup, Survey USA, Quinipiac, Harris, TerranceGroup, POS, Research2000, Mason Dixon, Zogby etc) to compensate for this.

Some, such as ABC and CBS are hoping their sample stratification will catch it.

Some (Pew, LATimes, Newsweek) are doing nothing.

Declining response rates IS a HUGE issue, everybody who is actually legitimately trying to generate accurate polls knows this is a looming and massive problem in telephone polling.

The honest and reputable pollsters are doing the best they can to adapt.

We will see who has made the right "guesses" on election day.

[The Vorlon]

How bout a section about several of the top polling organizations and their history along with a bit of commentary on how they conduct their polling and how good they are at it?

OK - will do

[The Vorlon]

General definistions, what is push polling, a tracking poll etc.

A precise definition of margin of error, confidence and ho they change based on sample size.

The ways you can determine who is a likely voter would be good.

Zogby deserves a section on his declien into madness.

Already planned & Thank-you!

"Zogby deserves a section on his decline into madness."

LOL !!! - I like that - I think I have a title for the article!

[Reds4]

Vorlon, is your map a picture of what you predict for election day, or how you see it if the vote happened today? What made you change it toward Bush some more? Was it something you saw in the national polls? Your map seems pretty reasonable to me, but I see Kerry winning West Virginia.

[The Vorlon]

Vorlon, is your map a picture of what you predict for election day, or how you see it if the vote happened today? What made you change it toward Bush some more? Was it something you saw in the national polls? Your map seems pretty reasonable to me, but I see Kerry winning West Virginia.

This map is a "snapshot" if the election were held today.

My preliminary Nov 2nd prediction actually has Kerry winnng 305/233.



There are three major demographic groups right now that are very very unsettled right now:

Married white women
Lower income males
College educated single men

The Nov 2 is based on all three breaking for kerry.

 I freely admit my Nov 2 prediction is utterly unreliable.  

When married white women break one way or the other, I'll tell you for sure who will win

There is a long, long, long way to go till Nov 2nd

[CO-OWL]

You must be logged in to read this quote.

There are three major demographic groups right now that are very very unsettled right now:

Married white women
Lower income males
College educated single men

The Nov 2 is based on all three breaking for kerry.

 I freely admit my Nov 2 prediction is utterly unreliable.  

When married white women break one way or the other, I'll tell you for sure who will win

There is a long, long, long way to go till Nov 2nd

You must be logged in to read this quote.

Any reasons why you think all three will break for Kerry?
(expected VP choice / trend in the polls / special campaign issues...)

Or are you just making an "utterly unreliable" guess?