Polls with 38% Republicans and 30% Democrat for example....

[kentuck]

How real is a poll with these rations of partisans? How likely is it that Democrats will be outvoted by 8% by the Republicans? Why is it when we hear of all the new registration, it is mostly more Democrats by huge margins? Are we to believe these folks that are registering Democrat are really going to vote Republican? Or that most likely won't vote?

I gave the numbers above as an example because I have read of such polls. If they are wrong this time, they are "Dewey" wrong.

[leftofthedial]

1. newly registered voters are not considered "likely" voters

though they should start showing up in "registered voter" polls.

Any statewide or national poll with a 38% repuke-30% democrat split is not valid because that is not a representative sample. It may be accurate in smaller populations.

[Erika]

2. I doubt if Newly Registered Voters Will Show Up in Polls....

until after the elections. Especially with their cell phones.

[Hello_Kitty]

3. That's funny to me

I should think someone newly registered would be MORE likely to vote, since they felt strongly enough to register. Especially in this election.

[Zuni]

4. seems logical to me

If someone took the time to register then I think they would at least go vote

[RafterMan]

6. Someone quoted a stat

saying just over 40% of newly-registered voters were likely to go to vote. I have no idea if it's accurate or where it came from, but it seems plausible.

[kentuck]

7. Just over 40% of old-registered voters go to vote...

????

[RafterMan]

8. Just passing on the stat, that's all.

"Steve Rosenthal, the Democratic chief executive of America Coming Together, said 44 percent of the 85,000 voters his organization registered last year turned out, compared with 49 percent over all."

http://www.nytimes.com/2004/10/04/politics/campaign/04vote.html?pagewanted=2

[SoCalDemocrat]

11. Turned out for what

Turned out for what? The the election is not until Nov 2nd. Are you talking about the primaries?

http://www.usatoday.com/news/politicselections/nation/president/2004-03-09-voter-turnout_x.htm

"WASHINGTON — Only 7.2% of American voters participated in the presidential primaries through March 2, according to a report released Tuesday.
Democratic turnout — at 11.4% — was the third lowest on record,
"

Jesus man you're comparing turnout for a vote for mayor! This race is going to have record turnout. Who turns out for a mayors race?

"Ms. Maxwell, of the League of Women Voters, noted that surges in registration have sometimes dissolved in disappointing turnout. But last year in the Philadelphia mayor's race, independent groups that registered thousands of new voters claimed their turnout was nearly as high as that in the rest of the electorate. Steve Rosenthal, the Democratic chief executive of America Coming Together, said 44 percent of the 85,000 voters his organization registered last year turned out, compared with 49 percent over all. "

[RafterMan]

17. "Jesus man"? Relax already.

It's pegged against registered voter turnout, what're you complaining about?

[mahina]

15. I once heard that over 98.75% of statistics are made up on the spot.

:)

[leftofthedial]

21. I'm sorry to correct you,

but it's 72.4% of statistics that are made up on the spot.

:-)

[mahina]

24. I stand corrected

:)
I think it should be at least a hundredths of a decimal tolerance just for accuracy's sake though. We should be careful when tossing bogus numbers aroundhowever- we might be tapped to work for the * admin, and I can only handle so much bullshit.
aloha noooooo!

[kentuck]

5. That seems logical to me also, ccbombs...

? Why would anyone bother to register unless they intended on voting? And why would they register Democrat unless they intended to vote Democratic?

[RevCheesehead]

10. um, maybe they're dying to serve on jury duty

[Tansy_Gold]

13. A lot of people register because someone hands them the form.

Not all these newly registered voters are those who made the effort to get the form and take it or mail it to the proper official office. Many were "helped" by interested party members, who set up a table at the county fair or in front of the new supermarket or go knocking on doors. The campaign volunteers have done a lot of the "work" getting these people registered, so in some cases there may be little to no motivation to actually go to the polling place and vote.

The real work will come on November 2, when the REAL "Get Out The Vote" effort has to be made.

[KharmaTrain]

9. Lots Of New Voters Won't Make It Until Next Time

These pollsters use call sheets from local voter roles...and some, only some, also cross check with drivers license registrations. If you're not on either list, you're not going to get called this time. So, unless you get cold-called, unregistered voters are passed over by most pollsters.

Also...the lists these places work off of are at least one, if not more elections old. If you've just registered, chances are you won't appear on their radars for maybe 2 years, more like 4 or more. Also, young people move alot, thus harder to track as are minorities (Hispanics are notoriously undercounted).

I'm glad to see Kerry still behind in these polls as it's getting our side more determined to prove them wrong and work harder. Let the Repugnicans think things are going well, doing just fine...it'll make what's coming their way so much more enjoyable to experience.

BTW...I just got my Zogby online poll about the debates last night...I'm sure I'm not the only one who hasn't been counted in the post-debate reactions...let's see where the number are right before Friday's debate...if there's not a Kerry lead or bounce showing, then something's amiss.

[DemocratSinceBirth]

12. Some Like CBS Use Random Dialing...

That seems the fairest...

[Machiavelli05]

14. thats not at all

because then you cannot control the sample at all...

[DemocratSinceBirth]

16. Why not?

You ask people if they are registered and likely to vote...


Of course they can lie...


But they could also lie about who they are voting for if you use registration rolls or phonebooks...

[Machiavelli05]

19. there needs to be more control than someone's voting probability

If you randomly pick numbers you can easily oversample in demographic and economic areas that are comparable. Random means you have approx. the same chance to pick 10 numbers starting with 555 (only eliminating the one you pick each time, decreasing the odds just slightly) as you do picking 5 555 numbers and 5 554 numbers. Now what if 555 numbers were overwhelmingly owned by a certain demographic or economic group because the first 3 numbers are frequently the same by area.

See?

[Chomskyite]

18. Their argument

. . . has been that they don't do anything to stack the proportion of Republicans to Democrats to independents, that they just call their sample number, usually 700-1,000, and that they just naturally fall into 38 Republican and 30 Democrat.

My problem is that they're doing this EVERY time. And that there's been NO reason for any sudden surge in the number of self-identifying Republicans. Their party is weak. Their candidate is weak. The demographic trends are against them.

I just think SOMETHING is wrong. I don't know what.

[depakid]

20. It's called bias (the statistical kind)

their sample systematically excludes some kinds of people- and those kinds of people are different than those who the "results" are drawn from. The most common you'll see in cheap media polls is non-response bias, but there are others.

[footinmouth]

22. Polls just seem bogus to me

I live in NY state. My state & its sought after electoral votes will all go to Kerry as will the state of California.

I'll admit, I know nothing about polling, but if they are sampling 1,000 registered or likely voters, is this polling data weighted demographically? Does my Kerry vote count more than a Wyoming Kerry voter?

I don't know why I'm even asking. Nobody ever calls me. I consider myself blessed that I am asked to participate in the online Zogby poll.

[TexasSissy]

23. I read that more Dems than Repubs voted in 2000. Is that true?

If so, then I guess they voted for Nader (or - shudder - for *).

[HFishbine]

25. "outvoted by 8% by the Republicans?"

Actually a poll that has Rs at 38% and Ds at 30% has Rs out-voting D's by 25% -- 25% more Rs than Ds.

[atre]

26. The last times Republicans outvoted Democrats in a presidential election

before World War II.