does anyone think polls are trustworthy???? check this....

[spanone]

Another Day, Another Poll, and a Very Different Result

Just one day after a poll showed his Republican opponent, Carl P. Paladino, with surprising strength in a one-to-one matchup, a new survey from a different polling organization showed Mr. Cuomo with a far wider margin. The new

poll (pdf), issued by the Siena Research Institute, showed Mr. Cuomo leading Mr. Paladino by 57 percent to 24 percent, with Rick A. Lazio, who lost the Republican primary but remains in the race on the Conservative Party line,

earning 8 percent.

Mr. Cuomo’s favorable rating was 60 percent and Mr. Paladino’s 32 percent — while Mr. Paladino’s unfavorable rating was 36 percent. The new poll had a margin of sampling error of plus or minus four percentage points.

The two polls were different in significant ways. While Wednesday’s poll, released by the Quinnipiac University Polling Institute, screened out voters that Quinnipiac judged as unlikely to vote in November, Siena surveyed a sample of

all registered voters. And the Quinnipiac survey did not include Mr. Lazio in its general election matchup.

[Kurt_and_Hunter]

1. Those sound like reasonable results

If a person doesn't care to understand a poll that's what it is.

Their incomprehension doesn't make the poll worthless.

Comparing a registered voter poll and a likely voter poll you should get different results. Comparing a poll with two-candidates to one with three candidates you'd definitely expect different results. What would be suspicious would be if the two polls were the same.

[Gormy Cuss]

4. +1

A comparison of apples to oranges will yield much greater difference than a comparison of navel oranges to seeded oranges.

[DemocratSinceBirth]

7. Such Large Differences Can Not Be Explained By Different Voter Models

There are such things as confidence levels and crappy pollsters.

[Gormy Cuss]

9. Political preference polls are usually limited to likely voters.

Assuming the methodology used has no inherent flaw the Quinnipiac poll is probably a better gauge of voter sentiment even if Siena used just as rigorous a methodology.

[DemocratSinceBirth]

10. One Of Thlose Polls Is Defintely Crap

You also have a confidence level of .95 which means there is a .05 chance you are not measuring what you think you are measuring.

[Gormy Cuss]

11. Well, yeah, but both could be crap on that basis.

It's hard to tell if either is any good based on the excerpt in the OP. If I had the interest I'd find the methodological info from the pollsters' own publications.

[stray cat]

2. It depends on how the poll is done - more reliable than DU opinion most of the time

[AtomicKitten]

3. Polling is often political, not analytical. nt

[ejpoeta]

5. I often wonder about cell only voters. I am cell only and know many of people who are also

cell only. I heard about that big poll that my local station touted showing how close the governor's race is. but i honestly don't think paladino has as big a chance as the tv folks seemed to think. paladino supposedly has people calling HIM to donate instead of having to go and ask. That may be. People around where i live seem to LOVE him. i think there are more of us then there are of them though. and i will be voting. i think many will just to ensure that nutjob doesn't get anywhere near the governor's mansion. i am pissed at albany, yes. but i think some people aren't really thinking but just reacting.

[SoCalDem]

6. By commissioning many polls, with widely varying results, a candidate or his/her staff

or a media outlet can "explain" ANY eventual outcome..and by cherry-picking polls they like, they can portray a candidate's support as vastly complimentary if they like. This drives fund raising in many cases.. No one wants to donate to a loser, but if polls show them as 'rising: or "gaining" or "just moved ahead", it creates a bandwagon effect, and especially if there is election-day chicanery in the works, these polls can legitimize a suspicious election.

[ChairmanAgnostic]

8. you should poll this question.